PALMER, J.

Timothy J. Sugrue, assistant state's attorney, with whom, on the brief, were Kevin T. Kane, chief state's attorney, Jo Anne Sulik, supervisory assistant state's attorney, and Michael E. O'Hare, former senior assistant state's attorney, for the appellant (respondent).

Paul Casteleiro, pro hac vice, with whom was W. James Cousins, for the appellee (petitioner).

ROGERS, C.J., and PALMER, ZARELLA, EVELEIGH, McDONALD and ESPINOSA, Js.

Opinion

PALMER, J. This certified appeal by the respondent, the Commissioner of Correction, requires us to decide whether the Appellate Court correctly concluded, contrary to the determination of the habeas court, that the petitioner, Richard Lapointe, is entitled to a new trial on the charges underlying his 1992 conviction of capital felony and other offenses because prior habeas counsel (first habeas counsel) rendered ineffective assistance in failing to demonstrate that the state withheld certain exculpatory evidence prior to trial in violation of Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), and its progeny, the disclosure of which would have supported an alibi defense. We agree with the petitioner both that he was deprived of a fair trial because his rights under Brady were violated and that his first habeas counsel's representation was constitutionally deficient in that counsel failed to establish that violation. We therefore affirm the judgment of the Appellate Court.

The respondent's appeal arises out of events that have their origin in the early evening hours of Sunday, March 8, 1987, when the victim, eighty-eight year old Bernice Martin, was raped, bound and murdered in her Manchester apartment, which her killer thereafter set ablaze in an apparent effort to destroy all evidence of the crime. The case remained unsolved until early 1989, when police focused their suspicions on the petitioner, the then forty-two year old mentally impaired husband of the victim's granddaughter, Karen Martin (Martin), with whom the petitioner resided along with their eight year old son. Until then, the petitioner was not a suspect: he had no criminal record or history of violence of any kind, and he seemed physically, mentally and temperamentally incapable of the brutal crime. Nevertheless, on July 4, 1989, over the course of a nine hour stationhouse interrogation by the Manchester police that lasted until the early morning hours of July 5, the petitioner gave three written statements in which he purported to take responsibility for the victim's murder. The petitioner repeatedly told the police, however, that he had no recollection of killing the victim and that he was confessing only because they wanted him to do so.

On the basis of these statements, the police obtained a warrant for the petitioner's arrest, and he ultimately was charged with capital felony and arson murder, among other offenses. Following a jury trial, he was convicted as charged and sentenced to life imprisonment without the possibility of release. After this court affirmed his conviction; State v. Lapointe, 237 Conn. 694, 739, 678 A.2d 942, cert. denied, 519 U.S. 994, 117 S.Ct. 484, 136 L.Ed.2d 378 (1996) ; the petitioner sought a writ of habeas corpus, claiming, inter alia, that his due process rights were violated because the state had failed to disclose a note, authored by Detective Michael Ludlow of the Manchester Police Department (Ludlow note), containing details concerning the length of time that the fire burned inside the victim's apartment prior to being discovered. The petitioner claimed that the note was both exculpatory and material under Brady because it purported to identify the time frame within which the fire was set, and Martin would testify that the petitioner was home, with her and their son, during that entire period, thereby providing the petitioner with a complete alibi. First habeas counsel, however, failed to pursue the claim, and, consequently, the first habeas court, Freed, J., rejected that claim as abandoned, as well as the petitioner's other claims on the merits. On appeal, the Appellate Court affirmed the judgment of the first habeas court. Lapointe v. Commissioner of Correction, 67 Conn.App. 674, 681, 789 A.2d 491, cert. denied, 259 Conn. 932, 793 A.2d 1084 (2002).

The petitioner subsequently filed the habeas petition that is the subject of this appeal, alleging, inter alia, that the state's failure to disclose the Ludlow note deprived him of due process of law and that his first habeas counsel had rendered ineffective assistance under Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), by failing to pursue and prove that claim. The second habeas court, Fuger, J., dismissed the claim, concluding, inter alia, that the petitioner had not established a prima facie basis for the exculpatory nature of the Ludlow note because, despite its existence, he could not account for his whereabouts for the entire window of time within which the victim was murdered. The petitioner appealed to the Appellate Court, which reversed in part the second habeas court's judgment and remanded the case for further proceedings. Lapointe v. Commissioner of Correction, 113 Conn.App. 378, 404, 966 A.2d 780 (2009). The Appellate Court's decision was predicated on its determination that the Ludlow note, when viewed in the light most favorable to the petitioner and considered together with certain statements from Martin as to when the petitioner was home with her, was exculpatory because it tended to support a finding that he could not have committed the crime in the requisite time frame. See id., at 392, 966 A.2d 780.

At the proceeding following the Appellate Court's remand of the case to the habeas court, the sole issue with respect to the Ludlow note was whether it was material. In support of his claim that the note was material, the petitioner presented expert testimony concerning the length of time the fire burned in the victim's apartment. Based on the burn time estimates of the petitioner's two experts, which were consistent with the notation that had been made in the Ludlow note, the fire was set in a relatively narrow window of time. The petitioner also presented evidence establishing that, if the state had disclosed the Ludlow note as required, his trial counsel would have called Martin as a witness, and Martin would have testified that the petitioner was home with her during the time frame within which, according to the petitioner's burn time experts, the fire was set. The respondent also presented expert testimony concerning the likely burn time of the fire. Under the far longer burn time estimate proffered by the respondent's expert, the petitioner could not establish, even with Martin's testimony, that he was home during that entire period. At the conclusion of the trial, the third habeas court, Nazzaro, J., rejected the petitioner's claim that his first habeas counsel was ineffective for failing to pursue a Brady claim on the basis of the state's nondisclosure of the Ludlow note. In particular, the third habeas court found that the testimony of the respondent's expert was far more persuasive than the testimony of the petitioner's experts and that it was not reasonably probable that, if the jury at the petitioner's criminal trial had heard the testimony of the petitioner's experts, it would have credited that testimony and reached a different result.

On appeal to the Appellate Court from the judgment of the third habeas court, the petitioner argued, inter alia, that, contrary to the finding of the third habeas court, he is entitled to a new criminal trial at which the jury would decide how much weight to assign to the testimony of the petitioner's experts. The Appellate Court agreed, concluding that the determination of which expert or experts were most persuasive was an issue to be decided by the jury at a new trial. See Lapointe v. Commissioner of Correction, 138 Conn.App. 454, 476–77 and n. 17, 53 A.3d 257 (2012). Accordingly, the Appellate Court reversed in part the judgment of the third habeas court and remanded the case with direction to grant the petition for a writ of habeas corpus and for a new trial. Id., at 480, 53 A.3d 257.

We then granted the respondent's petition for certification to appeal, limited to the following issue: “Did the Appellate Court properly determine that the [petitioner's] first habeas counsel was ineffective for failing to pursue a claim that the state had suppressed evidence in violation of Brady v. Maryland, [supra, 373 U.S. 83, 83 S.Ct. 1194 ]?” Lapointe v. Commissioner of Correction, 307 Conn. 940, 941, 56 A.3d 948 (2012). We answer the certified question in the affirmative because the testimony of the petitioner's experts was more than sufficient to call into question the reliability of the petitioner's conviction. Indeed, even if that expert testimony only tended to support the petitioner's claim that he could not have murdered the victim, in view of the tenuous nature of the state's case against the petitioner—based as it was on his suspect admissions—the state's Brady violation would warrant a new trial because, as the United States Supreme Court has recognized, exculpatory evidence of even “minor importance” may well be “sufficient to create a reasonable doubt” when, as in the present case, “the [guilty] verdict is already of questionable validity ....” United States v. Agurs, 427 U.S. 97, 113, 96 S.Ct. 2392, 49 L.Ed.2d 342 (1976). Accordingly, we affirm the judgment of the Appellate Court reversing in part the judgment of the third habeas court and ordering a new trial.

I

FACTS AND PROCEDURAL HISTORY

Our resolution of the issue presented by this appeal requires an extended discussion of the long and vexing history of the petitioner's case. At all times relevant to this appeal, the petitioner, his wife, Martin, and their son resided in the town of Manchester, within walking distance of the apartment of the victim, whom they visited every Sunday. On Sunday, March 8, 1987, the family followed their usual routine, attending church in the morning and then stopping for breakfast at My Brother's Place, a nearby restaurant where the petitioner was employed. When the restaurant closed at 2 p.m., they walked across the street to the victim's apartment, where they stayed for approximately two hours, watching television with the victim. At about 4 p.m., the petitioner and Martin, neither of whom had a driver's license, and their son took the ten minute walk home. After they arrived home, Martin, who has cerebral palsy and is slightly paralyzed, prepared dinner, and the petitioner, who has certain physical and mental impairments as the result of Dandy–Walker syndrome, took the family dog out for a walk. The petitioner was gone for around twenty minutes. Upon his return, at approximately 5 or 5:15 p.m., the family sat down for dinner.

At or about 5:45 p.m., the victim's daughter, Nathalie Howard, and her husband drove past the victim's apartment and saw the victim outside emptying her trash. At the time, the victim appeared fine, and the couple continued on their way. At approximately 7:55 p.m., Howard called the victim to check on her, as she did every evening, but the victim did not answer her telephone. Howard tried calling again at 8:05 p.m. When the victim still did not answer, Howard called her niece, Martin, to inquire whether Martin's father, who resided in New York, had come to visit the victim and had taken her out to dinner. When Martin told Howard that her father was not in town that day, Howard asked if the petitioner would walk over to the victim's apartment to make sure that nothing was wrong. Howard also requested that the petitioner call her as soon as he got to the victim's apartment to let Howard know that everything was alright.

According to Howard, the petitioner called her right back from the home of Jeannette King, the victim's neighbor. He told Howard that the doors to the victim's apartment were locked and that the victim must be sleeping because there were no lights on inside the apartment. Howard told the petitioner that the victim never went to bed at that hour and that she was heading right over to the victim's apartment. The petitioner also called Martin, who told him to go back to the victim's apartment immediately and to try to get inside because the victim might have fallen and injured herself. A few minutes later, the petitioner returned, out of breath, to King's apartment and told her that there was smoke coming from the victim's apartment. The petitioner dialed 911 from King's apartment at 8:27 p.m.

The first firefighter on the scene, Michael Tomkunas, who was off duty and arrived within moments of the 911 call, saw the petitioner standing in front of the building, motioning him in the direction of the victim's apartment. Tomkunas could see smoke coming from the apartment, and the front door was hot to the touch. Tomkunas immediately kicked in the front door and tried to enter the apartment, but the smoke and heat were too intense, and he was forced out. By then, other firefighters were arriving on the scene. One of them, Douglas Boland, ran to the back of the building and opened a set of sliding glass doors. The cross-ventilation dissipated enough of the smoke and heat to allow the firefighters to enter. When they did, they discovered the victim lying on the living room floor, six to eight feet from the burning couch. She was naked except for two pieces of fabric, one of which was tied so tightly around her neck that Tomkunas had trouble removing it; the other piece bound her hands and midsection. The firefighters carried the victim to the front lawn and performed cardiopulmonary resuscitation until paramedics arrived. Howard arrived just as rescue personnel were carrying her mother from the building. The drive from her home had taken approximately ten minutes. The victim subsequently was transported to Manchester Memorial Hospital, where she was pronounced dead.

The associate medical examiner, Arkady Katsnelson, performed an autopsy on the victim and determined that her cause of death was a combination of asphyxia by strangulation and smoke inhalation. He concluded that the victim was not manually strangled but was asphyxiated by pressure to the right side of her neck from a blunt object. The victim also suffered a three inch stab wound to her abdomen, ten less severe stab wounds to her back, extensive hemorrhaging and contusions and lacerations to the vaginal area, as well as first and second degree burns to various parts of her body. Katsnelson opined that the contusions and lacerations to the vaginal area were caused by a blunt object rather than by sexual intercourse.

The police recovered several items of potential evidentiary value from the victim's apartment, including a pair of men's gloves containing strands of the victim's hair and a semen stain from the bedspread belonging to a person with type A blood who also was a secretor. The petitioner is a secretor with type A blood, as is approximately one third of the male population. The police also recovered a pubic hair from the victim's clothing, but DNA tests were not then available. Buttons from the victim's blouse were found strewn about the bedroom floor, indicating that the perpetrator had forcibly removed the victim's clothing. The police also found a pool of the victim's blood on top of the bed, indicating that she was stabbed in that location.

The police investigation into the victim's homicide remained open for more than two years. In March, 1989, the case was reassigned to Detective Paul Lombardo of the Manchester Police Department. At that time, Lombardo decided to reinterview individuals who previously had been questioned by the police. On June 8, 1989, Lombardo interviewed the petitioner and took a saliva sample from him.

On July 4, 1989, Lombardo asked the petitioner to come to the police station for questioning. By that time, Lombardo had become convinced of the petitioner's guilt because of his blood type, his peculiar nature and mannerisms, and his repeated questions to the police about whether he was a suspect in the victim's murder. According to Joseph J. Brooks, Lombardo's commanding officer, the purpose of the interview was to elicit a confession from the petitioner. When the petitioner arrived at the police station, Lombardo informed him that there was incontrovertible evidence of his guilt. Although the petitioner initially denied any involvement in the victim's murder, after Lombardo told him that a person could commit a crime and not remember doing so, the petitioner signed two statements in which he acknowledged his involvement in the murder. In the first, one sentence statement, the petitioner purported to accept responsibility for the victim's death, but stated only that “it was an accident, my mind went blank.” In the second statement, the petitioner stated that he had no memory of killing the victim but that, if the evidence showed that he was there and that he killed her, then he must be guilty. Because the first two statements were devoid of any detail corroborating the petitioner's equivocal admission that he had killed the victim, Lombardo asked another Manchester police officer involved in the investigation, Detective Michael Morrissey, to take over the interrogation and to try to obtain more specific information from the petitioner. Several hours later, under questioning by Morrissey, the petitioner signed a third statement that was largely inconsistent with the crime scene evidence but that also contained three details about the crime that, presumably, would have been known only to the police and the killer. After signing the third statement, the petitioner was told to go home. Before leaving the station at 1:30 a.m. on July 5, however, the petitioner spoke to Brooks and again expressed uncertainty about his role in the victim's murder. Brooks asked him why he had he signed the three statements confessing to the murder. The petitioner responded that he simply had repeated what the officers had told him to say.

At the petitioner's criminal trial, the petitioner's counsel presented evidence to demonstrate that the petitioner was physically, mentally and emotionally incapable of committing, much less concealing, such a brutal and cold-blooded murder, as evidenced by the testimony of numerous psychologists and other witnesses who had known the petitioner at various stages in his life, including childhood friends, employers, fellow parishioners, and other members of the community. According to these witnesses, the petitioner's cognitive and motor skills are impaired, apparently the result of Dandy–Walker syndrome, and, as a consequence, he is slow-witted, easily confused, child-like and gullible—his mother-in-law described him as having the mentality of an eight year old—as well as physically awkward and uncoordinated. In addition, the petitioner was extremely close to the victim and her family, and he had no conceivable reason or motive to harm her, let alone to sexually assault and kill her. With respect to the petitioner's admissions, counsel argued that they were the product of a highly manipulative interrogation of an extremely vulnerable and impaired man, who had spent his entire life accommodating and agreeing with others in an effort to gain favor and to avoid conflict. Focusing principally on the petitioner's third incriminating statement, the state argued that the statement was freely given, and that it was powerful evidence of guilt because it contained information about the murder that only the killer would know. The state also argued that the petitioner was not nearly as compromised intellectually as defense witnesses had made him out to be because he had an intelligence quotient (IQ) of 92, which, the state asserted, was “nowhere close to being [even] slightly retarded.” The state also emphasized that the petitioner's blood type and secretor status were consistent with the seminal stain on the victim's bedspread. Finally, the state referred to certain conduct by the petitioner in the days and weeks following the murder that, in the state's view, suggested that the petitioner was the murderer: on several occasions, the petitioner had asked the police whether he was a suspect in the crime; he had told a neighbor that the victim was sexually assaulted before that information was known publically; and, on the night of the murder, he had gone to King's front door to ask to use her telephone, even though the back door was closer to the victim's apartment. On June 30, 1992, a jury found the petitioner guilty of all charges, and, after a penalty phase hearing, the trial court sentenced him to a term of life imprisonment without the possibility of release.

On direct appeal to this court, the petitioner claimed, inter alia, that the trial court improperly had denied his motion to suppress the oral and written statements that he had given to the police on July 4 and 5, 1989, and improperly concluded that the state constitution did not require the police to record his confession electronically. See State v. Lapointe, supra, 237 Conn. at 696, 678 A.2d 942. We rejected those claims and affirmed the petitioner's conviction. Id., at 703, 735, 739, 678 A.2d 942. Thereafter, on May 30, 1997, the petitioner filed his first writ of habeas corpus, alleging, inter alia, prosecutorial impropriety predicated on the state's failure to disclose certain exculpatory evidence, including the Ludlow note. At the first habeas trial, however, first habeas counsel presented no evidence to support a claim that the state had suppressed the Ludlow note, and, therefore, the first habeas court deemed the claim abandoned. The first habeas court rejected the petitioner's other claims, and the petitioner appealed to the Appellate Court, which affirmed the judgment of the first habeas court. Lapointe v. Commissioner of Correction, supra, 67 Conn.App. at 681, 789 A.2d 491.

On August 2, 2002, the petitioner filed a second habeas petition, alleging, inter alia, that his first habeas counsel rendered ineffective assistance by failing to pursue a claim that the state improperly had suppressed the Ludlow note. The following relevant facts relating to the Ludlow note are set forth in the decision of the Appellate Court in Lapointe v. Commissioner of Correction, supra, 138 Conn.App. 454, 53 A.3d 257. “At the time of the victim's homicide, Ludlow was ... assigned as the evidence officer for the crime scene. He subsequently assumed the position of case officer, which meant that he was responsible for the entire criminal investigation. A few days after the homicide, Ludlow had conversations with two state fire marshals who were assisting with the investigation. Ludlow took notes and wrote ‘CSP,’ which stood for Connecticut State Police, and ‘Steve Igoe’ and ‘Joe Roy,’ the names of the state fire marshals. Underneath those notations [in] the Ludlow note, the words ‘30–40 mins. Poss.’ were written. The numbers ‘30’ and ‘40’ were underscored twice. Ludlow testified that the notation ‘[P]oss.’ meant ‘possible’ and that the times represented the minimum amount of time that the fire could have been burning before the first responding firefighters arrived at the victim's apartment. At the time of the second habeas trial, Ludlow stated that he could not remember who gave him the burn time information. He admitted, however, that he was not an expert on fires and that he would not have made that estimate on his own. He testified that he would have asked one of the experts for the burn time if he was trying to determine a window of time within which the fire could have been started. Ludlow also acknowledged that, at the time of the first habeas proceeding, he testified that he had obtained the information as to a possible burn time from either Igoe or Roy. “The Ludlow note was first disclosed to [the petitioner's] counsel in 1999, after the petition for [a writ of] habeas corpus had been filed in the first habeas action. [The petitioner's trial counsel, Patrick] Culligan and [Christopher] Cosgrove both testified that they had not seen the Ludlow note prior to the petitioner's criminal trial in 1992.... [A]fter receiving the Ludlow note and other materials, [first habeas counsel] amended the petition to allege the state's failure to disclose or to produce the Ludlow note as Brady material, as required by the [federal and state] constitutions.... He did not, however, pursue that issue during the first habeas trial. The first habeas court, in its memorandum of decision, did not address the claim regarding the Ludlow note because it deemed that claim to [be] abandoned.

“[The petitioner alleged in a second habeas proceeding that his first habeas counsel's] failure to pursue the claim that the state suppressed the Ludlow note was ... ineffective assistance of counsel.... During the second habeas trial, [the petitioner's first habeas counsel] testified that he had not pursued that claim because he did not believe [that] the notation as to burn time in the Ludlow note [was] exculpatory. Culligan and Cosgrove, in their testimony before the second habeas court, opined that the Ludlow note was exculpatory and that the information could have been used by [them] at the criminal trial to buttress the petitioner's alibi defense. Culligan and Cosgrove further testified that if the Ludlow note had been disclosed to [them] prior to the [petitioner's criminal] trial, their strategy would have changed.

“[Martin] was not called as a witness in the petitioner's criminal trial. Culligan stated that Martin and the petitioner were divorced by that time and that the working relationship between her and the defense was no longer a good one. Although [Culligan] discovered prior to trial that [Detective] Morrissey had interviewed [Martin] on July 4, 1989, and that she had expressed her support for the petitioner at that time, and although [Culligan] knew that [Martin] had testified at the suppression hearing that the only time the petitioner was out of her sight on the night of the homicide was when she bathed their son between 6:15 [or 6:30] p.m. and 7 p.m., Culligan decided not to compel her testimony. He testified that he was concerned about her attitude toward the petitioner. Further, he was unaware of the existence of the Ludlow note and the importance of the start time of the fire to support the petitioner's alibi defense. Culligan testified that, if he had known that Igoe, the state's fire expert, gave a burn time of thirty to forty minutes, he would have called [Martin] as a witness because her testimony would have established that the petitioner was home when the fire was set. He also [testified] that [if] he ... had ... known of the information in the Ludlow note [he would have been prompted to hire an expert to testify about the fire's burn time for the purpose of demonstrating that the petitioner could not have started the fire because he was home when it was set].” (Footnotes omitted.) Id., at 468–71, 53 A.3d 257.

At the close of the petitioner's evidence at the second habeas trial, the respondent moved to dismiss the petition on the ground that the petitioner had failed to make out a prima facie case with respect to any of his claims. On August 2, 2007, the second habeas court granted the motion, concluding, with respect to the petitioner's claim predicated on the nondisclosure of the Ludlow note, that the note was not exculpatory. The court reasoned that, although the note arguably supported a finding that the minimum burn time was thirty to forty minutes—a period of time for which the petitioner could account for his whereabouts through Martin's testimony—the note said nothing about the maximum possible burn time, and the petitioner could not establish his whereabouts for a longer time period.

The petitioner appealed from the judgment of the second habeas court to the Appellate Court, which reversed in part the second habeas court's judgment as it related to the Ludlow note and remanded the case for further proceedings. Lapointe v. Commissioner of Correction, supra, 113 Conn.App. at 404, 966 A.2d 780. In reversing that portion of the second habeas court's judgment pertaining to the Ludlow note, the Appellate Court observed that “the record reveals ... that the victim was last seen outside of her apartment by Howard at about 5:45 p.m. The petitioner's former wife, [Martin], testified at a suppression hearing that she prepared dinner at the petitioner's home and that they ate dinner at about 5:15 p.m. or 5:30 p.m. Prior to sitting down for dinner, the petitioner ... walked the family dog for approximately twenty minutes. Therefore, according to her testimony, the petitioner had returned from his walk before the time that the victim was last seen outside of her apartment.... Martin ... further testified that [the petitioner] did not leave their house again until she received a telephone call from Howard, who requested that the petitioner walk over to the victim's house to check on [the victim]. According to Howard, this telephone call was placed a little after 8 p.m.” Id., at 391, 966 A.2d 780. The Appellate Court further noted that, although the state had adduced testimony from Martin acknowledging that the petitioner was out of her sight from approximately 6:15 or 6:30 p.m. until 7 p.m., when she was upstairs bathing their son, Martin also testified that, “while she was upstairs, it was possible to hear someone downstairs”; id.; and, in addition, that, from 7 p.m. until the time the petitioner left to check on the victim, she and the petitioner were watching television with their son. See id., at 391–92, 966 A.2d 780.

Viewing this evidence in the light most favorable to the petitioner, the Appellate Court concluded that “the thirty to forty minute minimum burn time, if credited as an accurate estimation, establishes that the fire was set at or before 7:50 p.m. The petitioner submitted evidence that, if credited, can account for his whereabouts, albeit tenuously, for the full window of time encompassing the last time the victim was seen alive outside her apartment to the time her body was discovered. Evidence that tends to prove his temporal inability to have committed the crime satisfies the definition of exculpatory and, therefore, is sufficient to establish the first prima facie element of a Brady claim.” Id., at 392, 966 A.2d 780.

On remand, the case was assigned to the third habeas court, Nazzaro, J., and the case proceeded to trial. At trial, the petitioner sought to demonstrate the materiality of the expert burn time testimony that the petitioner's trial counsel, Culligan and Cosgrove, would have presented, along with the testimony of Martin, to establish an alibi defense, if the state had timely disclosed the Ludlow note. In support of this claim, the petitioner presented the testimony of two fire experts, Gerard Kelder, Jr., and John DeHaan; the respondent presented expert fire testimony from Robert Corry for the purpose of rebutting the testimony of DeHaan and Kelder. All three experts had been provided with the same materials to review in connection with their investigations, including photographs of the damage to the victim's apartment caused by the fire, blueprints of the victim's apartment, the trial testimony of firefighters who responded to the fire, the state fire marshal's report on the fire, a video-recorded narration depicting the aftermath of the fire, and the other experts' reports on the fire. All three experts were highly experienced fire investigators. A review of their testimony reveals, moreover, that they all agreed that the fire began on the living room couch, that it burned intensely for a short period of time, and that it was quickly extinguished or reduced to a smoldering fire due to decreasing levels of oxygen in the victim's small, relatively airtight apartment. The principal distinction between the testimony of DeHaan and Kelder, on the one hand, and Corry, on the other, concerned the fire's maximum burn time.

Primarily on the basis of the smoke and fire damage to the apartment, Kelder concluded that the fire burned for approximately forty-five minutes to one hour. Under this estimate, the fire was set no earlier than 7:30 p.m. Because Martin could account for the petitioner's whereabouts from at least 7 p.m. until he left his house to check on the victim, Kelder's testimony, together with Martin's testimony, established that the petitioner was not at the victim's home when she was sexually assaulted and killed. DeHaan's testimony also supported the petitioner's alibi. In his view, in light of the smoke and heat conditions that Tomkunas had encountered when he arrived on the scene, the fire could have been set as late as 8 p.m. or 8:05 p.m. He further concluded, however, on the basis of the damage to the apartment and the dynamics of the fire, that it was set no earlier than 7:30 p.m. Like Kelder's estimate, DeHaan's burn time estimate was favorable to the petitioner because Martin could account for the petitioner's whereabouts from 7:30 p.m. until 8:05 p.m.

Corry thereafter testified that the fire could have been set anytime from 5:45 p.m., when the victim was last seen alive, until she failed to answer her telephone at around 8 p.m. Under this estimate of the fire's possible burn time, Martin did not provide the petitioner with a complete alibi because she could not positively account for the petitioner's whereabouts from 6:15 or 6:30 p.m. until 7 p.m.—when she was upstairs getting her son ready for bed—a time period within which, according to Corry, the fire could have been set.

On direct examination, Corry also was asked whether he had been present in the courtroom for DeHaan's testimony, two months earlier, and, if so, whether he agreed with DeHaan's conclusions regarding the fire. Corry responded that he had been present for DeHaan's testimony and that he disagreed with several of DeHaan's findings. In particular, Corry characterized DeHaan as having testified that the fire was a “high energy” fire. Corry then explained that, if the fire actually had been a high energy fire, there would have been more damage to the apartment. Specifically, Corry stated that a high energy fire “would have destroyed the wall behind [the couch], it would have destroyed the ceiling over it, and it would have probably brought this room to flash over.” When asked what may have caused DeHaan to overestimate the fire's energy level, Corry opined that DeHaan likely did not consider the probability that there was some kind of object, possibly a blanket, on the couch, which is where the fire began, and that the blanket or other object had interfered with and retarded the spread and intensity of the fire. Corry also characterized DeHaan as having testified that the apartment was 400 degrees at floor level when Tomkunas and the other firefighters entered the apartment. Corry disagreed with this testimony, as well, explaining that, if the temperature had been 400 degrees, as DeHaan stated, all of the firefighters would have been burned. According to Corry, the fact that they were not burned indicated that the entry level temperature was considerably lower than DeHaan had believed, which, in turn, indicated that the fire must have been started earlier than DeHaan had indicated because, by the time the firefighters entered the apartment, it had cooled to the point that it was safe for them to do so.

Following the trial, the third habeas court issued a memorandum of decision in which it rejected the petitioner's claim that the petitioner's first habeas counsel provided ineffective assistance by failing to pursue, as a claim in the first habeas proceeding, the state's suppression of the Ludlow note. For purposes of its analysis, the third habeas court treated the Ludlow note as “potentially exculpatory” and assumed that the state had inadvertently failed to disclose the note. The third habeas court concluded, however, that the note did not support the petitioner's alibi defense and, therefore, that the note was not material. In reaching its determination, the third habeas court concluded that Corry's burn time estimate, which did not support the petitioner's alibi defense, was more persuasive than the estimates that DeHaan and Kelder had provided. In support of this finding, the third habeas court made several subordinate findings. With respect to DeHaan, the third habeas court found that Corry had convincingly refuted DeHaan's testimony that (1) the fire was a high energy fire, (2) the fire's “peak temperature” reached 400 degrees, and (3) the temperature inside the apartment was 400 degrees when firefighters entered.

With respect to the latter finding, the third habeas court stated: “The fact that Tomkunas was not burned contradicts any opinion that the temperature was 400 degrees. At 400 degrees, Corry cogently explained, one would expect Tomkunas to suffer injury to the back of his neck and his hands at a minimum. In addition, the photographs of the fire damage, or lack thereof, [to] materials in the apartment ... do not support ... DeHaan's testimony that temperatures reached 400 degrees.” The third habeas court further noted that Corry had “also criticized the validity of ... DeHaan's opinions regarding temperatures as not taking into account the material in the room [that] helped to slow the fire's burning, such as a blanket and poly foam cushions on the couch, [which was] thought to be one of three ... points of origin [of the fire].” Finally, the third habeas court found that Corry also had convincingly refuted Kelder's testimony that (1) the fire was a high energy fire, and (2) the fire's peak temperature reached 1800 degrees at ceiling level.

In light of these findings, the third habeas court stated: “Considerable and extensive testimony by arson experts/investigators was presented by both the petitioner and the respondent in their efforts to determine the burn time.” This “testimony amount[ed] to a contest among experts.... [Nevertheless] [t]he court, as the finder of fact in this proceeding, assigns far more credit or weight to the testimony of ... Corry ... than [that of] ... Kelder or ... DeHaan regarding estimation of the burn time.” The third habeas court also observed that “[w]hat is clear from all the evidence in the record, the original trial testimony, crime scene photographs, reports, and the expert testimony presented to [the] court on the fire, is that the precise time the fire was set cannot be determined. At best, a range is established that includes that time period of [6:15 to 7 p.m.] ... when ... Martin cannot account for the petitioner's whereabouts, and [this] does not provide an alibi for him.” With respect to this point, the court further explained that, because Martin was upstairs giving her son a bath from 6:15 to 7 p.m., she could not have known whether the petitioner was at home, downstairs, during that time frame, and, consequently, her testimony, standing alone, did not provide “the petitioner with anything that remotely amounts to an alibi” for that forty-five minute period.

The third habeas court concluded that, “[g]iven all of the foregoing, the court simply cannot conclude that the nondisclosure of the Ludlow note, assuming it is potentially exculpatory and was inadvertently not disclosed by the state, supported the petitioner's ‘alibi’ defense. The inability to precisely determine the start time of the fire, coupled with ... Martin's testimony that she [could not] account for the petitioner's whereabouts from [6:15 to 7] p.m., leads [the] court to conclude that the petitioner has failed to [satisfy] both the third prong of Brady and the second Strickland

prong. That is, the petitioner has failed to show how the Ludlow note and the ‘30–40 min [P]oss' reference, if disclosed to [the petitioner's trial] counsel, reasonably would have led the jury to conclude [that] there was reasonable doubt in light of all the evidence presented to the jury. “Accordingly ... the court concludes that the petitioner has failed to show that [his first habeas counsel] rendered ineffective assistance of counsel.... The court is unable to conclude that [first habeas counsel's] failure to have an arson/fire expert testify in the first habeas [proceeding] would have resulted in anything different [from the present] habeas [case]: a prototypical battle of the experts resulting in diverging opinions.... [T]he court finds more persuasive the testimony presented by the [respondent's expert, Corry], rather than either of the petitioner's two experts, [Kelder] and ... DeHaan.” (Footnote added.)

The petitioner appealed from the judgment of the third habeas court to the Appellate Court, claiming, inter alia, that the third habeas court incorrectly concluded that the petitioner's evidence regarding the fire's burn time was not material and, consequently, that the petitioner's first habeas counsel did not render ineffective assistance in failing to pursue the claim that the state's suppression of the Ludlow note deprived the petitioner of a fair trial. Lapointe v. Commissioner of Correction, supra, 138 Conn.App. at 468, 53 A.3d 257. The Appellate Court agreed with the petitioner. See id., at 476–80, 53 A.3d 257. After setting forth the legal principles and standard of review that govern claims under Strickland and Brady; id., at 474–76, 53 A.3d 257 ; see also footnotes 3 and 4 of this opinion; the Appellate Court stated in relevant part: “[T]he state's suppression of the Ludlow note, and [first habeas counsel's] failure to pursue that claim, warrants a new trial for the petitioner. [The court] reach[es] that conclusion for the following reasons. Culligan and Cosgrove testified that had the burn time information in the Ludlow note been disclosed prior to the petitioner's criminal trial, their trial strategy would have changed. They stated that they would have used the thirty to forty minute estimate to buttress the petitioner's alibi defense, particularly because the estimate came from one of the state's fire marshals assigned to the investigation. As Culligan testified, [he and Cosgrove] would have retained the services of an arson expert. At the ... habeas trial, the two experts called by the petitioner testified that the fire could not have been set any earlier than 7:30 p.m. If that testimony had been presented at the [petitioner's] criminal trial, and credited by the jury, the petitioner's whereabouts at and after 7:30 p.m. would have been critical to his defense.

“For that reason, as both trial counsel testified, they would have called ... Martin as a witness at the criminal trial. She consistently had maintained that the petitioner was in their home with her and their son the entire evening of the victim's homicide. During her testimony at the suppression hearing, [Martin] stated that the only time that the petitioner was not in her sight was between 6:15 p.m. and 7 p.m., when she was bathing their son. If the jury credited ... Martin's testimony, it could have concluded that the petitioner was at home watching television with [Martin] and their son when the fire had been set.

“[In addition], if ... Martin had testified and the jury believed her testimony, the jury could have concluded that the petitioner had, at most, a forty-five minute window of time within which to commit the crimes. This would mean that between 6:15 p.m. and 7 p.m., on the night of the homicide, the petitioner: (1) walked the distance between his home and the victim's apartment; (2) had a cup of coffee with the victim while they were chatting on the couch; (3) used the victim's bathroom, located close to the victim's bedroom; (4) emerged from the bathroom, saw the victim combing her hair and decided to sexually assault her; (5) undressed himself, then tore the clothes off the victim; (6) sexually assaulted the victim; (7) retrieved a knife from the kitchen; (8) stabbed the victim ten times in the back and once in the abdomen; (9) used strips of cloth to tie them as a ligature so tightly around the victim's neck that the responding firefighters had difficulty removing the cloth; (10) loosely tied bindings around the victim's wrists and stomach area; (11) removed the victim from the bed and placed her on or near the couch; (12) washed any blood from his body and dressed himself; (13) set fires in three separate locations in the victim's apartment; and (14) walked the distance from the victim's apartment back to his home. According to ... Martin's recorded statements to Morrissey on July 4, 1989, and her testimony at the suppression hearing, which the jury did not hear, the petitioner was sitting in the living room when she came downstairs from bathing their son, and there were no signs of exertion or excitement. She noticed nothing out of the ordinary in his behavior that evening.” (Footnotes altered.) Lapointe v. Commissioner of Correction, supra, 138 Conn.App. at 476–78, 53 A.3d 257.

“[T]here is a reasonable probability that the result of [the petitioner's] criminal trial would have been different had the Ludlow note been disclosed to Culligan and Cosgrove prior to trial. Nondisclosure prior to trial of the portion of the Ludlow note describing the possible burn time affected the overall fairness of the trial and was so unfair as to undermine [the court's] confidence in the jury's verdict. With the burn time estimate provided by one of the state's fire marshals, trial counsel testified that they would have retained the services of an arson expert and that ... Martin would have testified as to the petitioner's whereabouts during the critical times of that evening. That evidence, if believed by the jury, could have resulted in the jury's finding that it was temporally impossible for the petitioner to have committed the crimes [of] which he was convicted. The Ludlow note was exculpatory and material in these circumstances. [First habeas counsel's] performance was deficient [in that] he failed to pursue that issue at the first habeas proceeding, and the petitioner was prejudiced by his failure to do so. The petitioner has demonstrated that had there been effective representation by [first habeas counsel], there is a reasonable probability that the first habeas court would have found that the petitioner was entitled to ... a new trial.” (Footnote omitted.) Id., at 478–80, 53 A.3d 257. Accordingly, the Appellate Court reversed in part the judgment of the third habeas court and remanded the case with direction to grant the petition for a writ of habeas as to the petitioner's claim alleging ineffective assistance of counsel with respect to the Brady violation and to order a new criminal trial. Id., at 480, 53 A.3d 257. Although the Appellate Court did not expressly say so, it appears, as the parties agree, that the Appellate Court accorded no deference to the determination of the third habeas court that there was no reasonable probability that a jury would credit the testimony of the petitioner's experts. Instead, the Appellate Court exercised its own judgment with respect to whether a jury reasonably could credit the testimony of the petitioner's experts.

We granted the respondent's petition for certification to appeal, limited to the issue of whether the Appellate Court properly reversed the judgment of the third habeas court and concluded that the petitioner is entitled to a new trial on the ground that first habeas counsel's representation of the petitioner during the first habeas trial was ineffective due to his failure to pursue a claim that the state had suppressed the Ludlow note. See Lapointe v. Commissioner of Correction, supra, 307 Conn. at 941, 56 A.3d 948. On appeal, the respondent asserts that the Appellate Court applied an incorrect standard of review in evaluating the petitioner's Brady claim, and, as a result, it improperly concluded that a new trial is required. In particular, the respondent contends that the Appellate Court improperly failed to afford due deference to the third habeas court's determination that the petitioner's expert burn time testimony would have had no effect on the outcome of his criminal trial, and that, if the Appellate Court had accorded such deference to that determination, it would have concluded, like the third habeas court, that first habeas counsel's failure to pursue a claim predicated on the state's suppression of the Ludlow note did not constitute ineffective assistance of counsel under Strickland. The petitioner, on the other hand, maintains that the Appellate Court correctly determined that the testimony of his burn time experts reasonably could have been credited by the original jury, and, if the jury did credit that testimony, it undoubtedly would have concluded that the petitioner could not possibly have committed the crimes with which he was charged. As a result, the petitioner contends that first habeas counsel's failure to raise a Brady claim predicated on the state's suppression of the Ludlow note rendered counsel's representation of the petitioner constitutionally deficient.

We agree with the petitioner that, notwithstanding the findings of the third habeas court, there is, at the very least, a reasonable likelihood that a jury, upon hearing the testimony of the petitioner's burn time experts and Martin in light of the other trial evidence, would harbor a reasonable doubt as to whether the petitioner sexually assaulted and murdered the victim, and then set her apartment on fire. As we explain more fully hereinafter, because the materiality issue ultimately presents a question of law for this court, and because we are not bound by the third habeas court's appraisal of the scientific underpinnings of the parties' expert testimony, we exercise plenary review over those issues. On the basis of that review, it is clear that the question presented by this appeal must be resolved in favor of the petitioner: because the state deprived the petitioner of the opportunity to avail himself of highly relevant alibi evidence—evidence that, as the petitioner contends, the original jury readily could have credited—fundamental fairness requires that the petitioner be afforded the opportunity to have a second jury consider that exonerating testimony. Our conclusion takes due account of the fact that the state's case against the petitioner was relatively weak, founded as it was on highly questionable admissions. Accordingly, the Appellate Court properly concluded that the judgment of the third habeas court must be reversed in part and that the petitioner was entitled to a new trial. II

GOVERNING LEGAL PRINCIPLES

A

Brady and Strickland

“In Brady, the United States Supreme Court held that the suppression by the prosecution of evidence favorable to an accused upon request violates due process [when] the evidence is material either [as] to guilt or to punishment, irrespective of the good faith or bad faith of the prosecution.... In Strickler v. Greene, 527 U.S. 263, [281–82] 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999), the United States Supreme Court identified the three essential components of a Brady claim, all of which must be established to warrant a new trial: The evidence at issue must be favorable to the accused, either because it is exculpatory, or because it is impeaching; that evidence must have been suppressed by the [s]tate, either [wilfully] or inadvertently; and prejudice must have ensued.... Under the last Brady prong, the prejudice that the defendant suffered as a result of the impropriety must have been material to the case....” (Footnote added; internal quotation marks omitted.) State v. Ortiz, 280 Conn. 686, 717, 911 A.2d 1055 (2006). “[T]he evidence will be deemed material only if there would be a reasonable probability of a different result if the evidence had been disclosed.” (Internal quotation marks omitted.) State v. Jordan, 314 Conn. 354, 370, 102 A.3d 1 (2014). This standard is met if “the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict.” (Internal quotation marks omitted.) State v. Ortiz, supra, at 717, 911 A.2d 1055.

Furthermore, with “respect to Brady 's third prong, a showing of materiality does not require demonstration by a preponderance [of the evidence] that disclosure of the suppressed evidence would have resulted ultimately in the defendant's acquittal.... The question is not whether the defendant would more likely than not have received a different verdict with the evidence, but whether in its absence he received a fair trial, understood as a trial resulting in a verdict worthy of confidence.... The United States Supreme Court [has] emphasized that the [relevant test under United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985) ] is not a sufficiency of the evidence test.... A defendant need not demonstrate that after discounting the inculpatory evidence in light of the undisclosed evidence, there would not have been enough left to convict.... One does not show a Brady violation by demonstrating that some of the inculpatory evidence should have been excluded, but by showing that the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict.... Accordingly, the focus is not whether, based [on] a threshold standard, the result of the trial would have been different if the evidence had been admitted. We instead concentrate on the overall fairness of the trial and whether nondisclosure of the evidence was so unfair as to undermine our confidence in the jury's verdict.” (Internal quotation marks omitted.) State v. Ortiz, supra, 280 Conn. at 717–18, 911 A.2d 1055. Put differently, materiality is established if the withheld evidence is of sufficient import or significance in relation to the original trial evidence that it reasonably might give rise to a reasonable doubt about the petitioner's guilt. See United States v. Agurs, supra, 427 U.S. at 112, 96 S.Ct. 2392. Additionally, “a trial court's determination as to materiality under Brady presents a mixed question of law and fact subject to plenary review, with the underlying historical facts subject to review for clear error.” (Internal quotation marks omitted.) State v. Ortiz, supra, at 720, 911 A.2d 1055. Finally, in the present case, we conduct the required “independent review” of the record; id., at 721, 911 A.2d 1055 ; mindful of the fact that this court's “duty to search for constitutional error with painstaking care is never more exacting than it is in a capital case....” (Internal quotation marks omitted.) Kyles v. Whitley, 514 U.S. 419, 422, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995).

Under Strickland, “[a] claim of ineffective assistance of counsel consists of two components: a performance prong and a prejudice prong. To satisfy the performance prong ... the petitioner must demonstrate that his attorney's representation was not reasonably competent or within the range of competence displayed by lawyers with ordinary training and skill in the criminal law.... To satisfy the prejudice prong, [the petitioner] must demonstrate that there is a reasonable probability that, but for counsel's unprofessional errors, the result of the proceeding would have been different.” (Internal quotation marks omitted.) Janulawicz v. Commissioner of Correction, 310 Conn. 265, 268 n. 1, 77 A.3d 113 (2013). “A reasonable probability is a probability sufficient to undermine confidence in the outcome.” Strickland v. Washington, supra, 466 U.S. at 694, 104 S.Ct. 2052. As in the case of an alleged Brady violation, “[i]n order to demonstrate such a fundamental unfairness or miscarriage of justice, the petitioner should be required to show that he is burdened by an unreliable conviction.” (Internal quotation marks omitted.) Michael T. v. Commissioner of Correction, 307 Conn. 84, 102, 52 A.3d 655 (2012). Finally, the respective roles of the habeas court and the reviewing court are also the same under Strickland as they are under Brady . As a general matter, “the underlying historical facts found by the habeas court may not be disturbed unless they were clearly erroneous.... [W]hether those facts constituted a violation of the petitioner's rights under the sixth amendment [however] is a mixed determination of law and fact that requires the application of legal principles to the historical facts of [the] case.... As such, that question requires plenary review by this court unfettered by the clearly erroneous standard.” (Internal quotation marks omitted.) Gonzalez v. Commissioner of Correction, 308 Conn. 463, 469–70, 68 A.3d 624, cert. denied sub nom. Dzurenda v. Gonzalez, ––– U.S. ––––, 134 S.Ct. 639, 187 L.Ed.2d 445 (2013).

In the present case, the petitioner's claim of a constitutional violation under Strickland and Brady implicates yet another constitutional protection, namely, the right to present a defense. “[T]he federal constitution require[s] that criminal defendants be afforded a meaningful opportunity to present a complete defense.” (Internal quotation marks omitted.) State v. Santana, 313 Conn. 461, 470, 97 A.3d 963 (2014). This right affords a defendant the opportunity to present his or her version of the facts, along with that of the prosecution, “so that [the jury] may decide where the truth lies”; (internal quotation marks omitted) id.; and includes the right to present an alibi defense. E.g., State v. Bryant, 202 Conn. 676, 704, 523 A.2d 451 (1987). Finally, it bears emphasis that, if the petitioner had been able to raise an alibi defense at his criminal trial based on the testimony of DeHaan, Kelder and Martin, the state would have been required to disprove that defense beyond a reasonable doubt. See, e.g., State v. Milardo, 224 Conn. 397, 407, 618 A.2d 1347 (1993).

B

The Principles of Brady and Strickland in Relation to the Facts of the Present Case

For purposes of the present habeas petition, the petitioner was required to establish, under Strickland, that first habeas counsel's failure to pursue the Brady claim concerning the state's suppression of the Ludlow note fell below the performance standard expected of competent habeas counsel and that the petitioner was prejudiced thereby. To make the necessary showing of prejudice, the burden was on the petitioner to demonstrate that the Ludlow note had been withheld by the state and was both exculpatory and material. In view of the fact that the third habeas court treated the note as exculpatory and assumed that it had been suppressed by the state, the petitioner was required to establish only that the note was material, a standard that is met upon a showing that, if the note had been disclosed before the petitioner's criminal trial, there is a probability of a different outcome sufficient to undermine confidence in the outcome of that trial. Because the test for materiality under Brady and the test for prejudice under Strickland are the same—with respect to both, the petitioner must demonstrate that the alleged constitutional impropriety gives rise to a loss of confidence in the original outcome—the petitioner is entitled to a new trial if he can demonstrate that first habeas counsel's failure to pursue a Brady claim predicated on the state's suppression of the Ludlow note was sufficiently harmful to satisfy that standard. Because it has been determined that the state's failure to disclose the Ludlow note deprived the petitioner of information that would have led him to obtain expert burn time testimony demonstrating that the fire was set at a time when, according to Martin, the petitioner was home with her, the success of the petitioner's Brady claim depends on whether the original jury reasonably could have credited that expert testimony.

As we previously noted, the third habeas court found that the testimony of both parties' burn time experts was thorough and extensive. Indeed, the third habeas court characterized that testimony as “a prototypical battle of ... experts” who are highly qualified in their field. Nevertheless, the third habeas court concluded that the testimony of the petitioner's experts, when viewed in the light of Corry's testimony, was not sufficiently credible to give rise to a reasonable probability of a different result at the original trial, that is, it was not persuasive enough such that the jury reasonably might have credited it. The Appellate Court disagreed, however, concluding that the petitioner's expert testimony was of sufficient import and credibility that the petitioner is entitled to a new trial at which the jury will evaluate that testimony.

We therefore must decide whether the Appellate Court properly determined that the third habeas court was incorrect in concluding that the jury reasonably could not have credited the testimony of the petitioner's burn time experts. As a reviewing court, we ordinarily accord deference to credibility determinations that are “made on the basis of [the] firsthand observation of [a witness'] conduct, demeanor and attitude.” (Internal quotation marks omitted.) State v. Kendrick, 314 Conn. 212, 223, 100 A.3d 821 (2014). In the present case, however, the third habeas court's assessment of the testimony of DeHaan and Kelder was not predicated on their demeanor or conduct on the witness stand, nor was it related to anything else that would reflect adversely on their credibility, such as untruthfulness, bias, poor memory or substandard powers of observation. That assessment also was not dependent on any underlying factual findings requiring the trial court's firsthand observation and determination of the credibility or reliability of other witnesses. Rather, the third habeas court rejected the opinions of DeHaan and Kelder solely because, in its view, those opinions lacked an adequate foundation, first, because they were premised on facts that were contrary to the record in the case, as reported by Corry, and, second, because the court did not credit the scientific underpinnings of those opinions. In such circumstances, when the habeas court's assessment of the expert testimony has nothing to do with the personal credibility of the expert witness but instead is based entirely on the court's evaluation of the foundational soundness of the witness' professional opinion, this court is as well situated as the habeas court to assess that testimony for Brady purposes.

The Indiana Court of Appeals recently addressed this precise issue in a case that is factually and procedurally indistinguishable in any material respect from this one. In Bunch v. State, 964 N.E.2d 274 (Ind.App.), trans. denied, 971 N.E.2d 1215 (Ind.2012), the defendant, Kristine Bunch, who had been convicted of murder in the arson-related death of her son, sought a new trial based on newly discovered evidence concerning the place of origin of the fire that she was found to have set. Id., at 279–80. As in the present case, one of the issues in Bunch was the materiality of expert testimony concerning the fire; see id., at 290 ; and the Court of Appeals was called on to review the trial court's finding that the testimony of Bunch's expert was not sufficiently worthy of credit to warrant a new trial. See id., at 283–85, 293. In reversing the judgment of the trial court on that issue; id., at 297, 304 ; the Court of Appeals reviewed that expert testimony without affording any deference to the trial court's finding that the testimony lacked persuasive force and, therefore, did not support Bunch's new trial claim. See id., at 292–93. Because we agree fully with the well reasoned analysis of the Court of Appeals, we repeat it in relevant part here: “In general, [w]hether a [witness'] testimony at a [postconviction] hearing is worthy of credit is a factual determination to be made by the trial judge who has the opportunity to see and hear the witness testify.... It is not within the province of [an] appellate court to replace a trial judge's assessment of witness credibility with its own.... Thus, if the [trial] court expressly finds that the testimony of a fact witness is or is not worthy of credit, we must accept that determination....

“However ... the [trial] court did not find [that the] testimony [of Bunch's fire expert] was not worthy of credit because it doubted her credibility or veracity based [on] a firsthand evaluation of her demeanor; the [trial] court found her testimony not worthy of credit because it was in conflict with trial evidence.... In other words, the [trial] court did not find [the expert] unworthy of credit on the basis of her demeanor; it found her expert opinion unworthy of credit on the basis of its foundation. Under these circumstances, we do not think it [is] necessary or appropriate to impute a personal credibility determination to which we must defer to the [trial] court. To do so would virtually eviscerate appellate review of [postconviction] denials because we would have to speculate in every instance that the [trial] court could have concluded [that] the witness was not credible based on his or her demeanor.

“Thus, although we would defer to the [trial] court's assessment of fact witnesses—for instance, a trial witness now recanting trial testimony or a new witness offering never-before-heard exculpatory testimony—we will not defer in this case to the [trial] court's assessment of an expert's scientific evidence. We have the ability to assess [the expert's] testimony ourselves because her credentials and the basis for her opinion are part of the record. The [trial] court found [that the expert's] testimony was not reliable because she did not establish the scientific principles for her conclusion and because [her] conclusions contradicted [the] undisputed evidence and eyewitness testimony from the trial. In making such a determination, the [trial] court did not rely on her demeanor ... but rather on the stated bases for her opinion and review of the trial record. We have the same information before us, and therefore are able to independently assess whether [the expert's] testimony is worthy of credit without invading the province of the [trial] court.” (Citations omitted; emphasis altered; internal quotation marks omitted.) Id.; see also State v. Behn, 375 N.J.Super. 409, 431–33, 868 A.2d 329 (App.Div.), cert. denied, 183 N.J. 591, 874 A.2d 1108 (2005). This case presents the same exceptional factual and procedural scenario as Bunch presents. Therefore, we, like the Appellate Court, see no reason to defer to the third habeas court's predictive or probabilistic judgment as to whether the original jury reasonably might have credited the testimony of the petitioner's experts. This is particularly true because, as we explain in part III A of this opinion, the record does not support any of the reasons that the third habeas court provided in substantiation of its conclusion that the jury reasonably could not have credited the testimony of the petitioner's experts. In fact, our review of the record reveals nothing to support the third habeas court's conclusion. On the contrary, there is strong reason to believe that the jury might well have found the testimony of the petitioner's experts persuasive, especially in light of their unquestioned qualifications and experience.

In light of this determination, we finally must decide whether the testimony of the petitioner's experts nevertheless is insufficient, when considered together with the original trial evidence, to call into question the outcome of the petitioner's criminal trial. We conclude that there is a very high probability that that testimony, if credited, along with Martin's testimony, would give rise to a reasonable doubt as to whether the petitioner committed the crimes of which he was convicted. Having demonstrated, therefore, that there is a real and substantial probability that the testimony of DeHaan, Kelder and Martin will result in the petitioner's acquittal,the petitioner is entitled to a new criminal trial at which he will have the benefit of that highly exculpatory evidence.

III

THE ISSUES

A

Whether a Jury Reasonably Could Credit the Testimony of DeHaan and Kelder

We first address the respondent's claim that the Appellate Court improperly rejected the third habeas court's conclusion that DeHaan's and Kelder's testimony is not sufficiently credible to give rise to a reasonable doubt about the petitioner's guilt. We disagree with the respondent's contention because the record belies the third habeas court's determination that, in all probability, the original jury would have reached the same result even if it had heard that testimony.

We begin with the credentials of the petitioner's experts. As the third habeas court observed, both DeHaan and Kelder are highly experienced fire investigators, and they have worked in that field for a combined total of more than sixty years. DeHaan's resume is particularly impressive. In addition to his years of service as a fire investigator for the Federal Bureau of Alcohol, Tobacco and Firearms, and the California Department of Justice, he has authored or coauthored close to seventy-five scholarly articles or professional papers and a college level treatise entitled, “Kirk's Fire Investigation,” which is presently in its seventh edition. DeHaan's treatise is utilized not only by students but also by investigators in the field. In fact, the record reflects that Corry himself relied on that treatise in preparing his report on the fire at issue, as counsel for the respondent did in connection with his cross-examination of Kelder. The record also reveals that DeHaan has taught fire investigation for approximately thirty years, and, at the time of the third habeas trial, in the graduate forensic sciences program at the University of California, Davis. Because these factors alone support the conclusion that a jury reasonably could credit DeHaan's and Kelder's testimony, we turn to the third habeas court's reasons for reaching the opposite conclusion.

As we noted previously, the third habeas court, in concluding that Corry was so much more persuasive than DeHaan and Kelder that no jury reasonably could credit DeHaan's and Kelder's testimony, relied on three threshold or subordinate findings: (1) DeHaan and Kelder mischaracterized the fire's energy level; (2) DeHaan and Kelder overestimated the fire's “peak” temperature; and (3) DeHaan overestimated the temperature of the fire when Tomkunas entered the victim's apartment. With respect to the first two findings, the third habeas court stated: “In contrast to ... DeHaan and Kelder, Corry opined credibly that ‘this was not a high energy fire.’ He disputed DeHaan's and Kelder's estimate[s] of peak temperature.... Corry testified credibly that ... DeHaan overstated the amount of heat generated by [the] fire.... In addition, the photographs of the fire damage, or lack thereof, [to] materials in the apartment ... do not support ... DeHaan's testimony that temperatures reached 400 degrees.... When compared with Corry's reasonable and measured analysis, Kelder's estimate of temperature at 1800 degrees appears wildly exaggerated.” With respect to the third finding, the third habeas court stated: “Corry noted that Tomkunas first responded to the call and found [that] the door to the apartment was ‘hot,’ but he did not burn himself. He sustained no injury to his hands, his exposed ears or any other body part. These facts, Corry pointed out, contradict ... DeHaan's estimate that the temperature upon entry was 400 degrees.... Corry noted that Tomkunas had dropped to his knees upon entering the apartment. Corry opined [that] the temperature at that time was more likely 150 to 190 degrees. The fact that Tomkunas was not burned contradicts any opinion that the temperature was 400 degrees. At 400 degrees, Corry cogently explained, one would expect Tomkunas to [have] suffer[ed] injur[ies] to the back of his neck and his hands at a minimum.”

A review of the record reveals that the third habeas court's critique of the testimony of DeHaan and Kelder is comprised of factually unfounded assertions. First, contrary to the court's finding, DeHaan and Kelder never testified that the fire was a high energy fire. To the contrary, DeHaan emphasized throughout his testimony that the limited damage to the apartment was evidence that the fire was not a high energy fire. He described it as a “very limited scale fire,” no different from “an average fireplace fire.” When Kelder was asked whether he would characterize the fire as high or low energy, he similarly responded, “low energy,” which he also believed was evidenced by the fact that the apartment was not seriously damaged. Thus, although the third habeas court found that Corry had persuasively rebutted DeHaan's and Kelder's testimony that the fire was a high energy fire, the record is clear that DeHaan and Kelder never provided the testimony that Corry attributed to them. Indeed, when questioned about the fire's energy level, DeHaan and Kelder both expressly characterized it as a low energy fire.

With respect to the third habeas court's statement that Corry had convincingly “disputed DeHaan's and Kelder's estimate[s] of peak temperature,” the petitioner correctly observes that, as between Corry and DeHaan, it was Corry who provided the higher estimate of the fire's peak temperature. On cross-examination, the petitioner's counsel asked Corry to comment on several photographs that were taken inside the victim's apartment after the fire, one of which depicts a stack of charred newspapers. Corry testified that all of the papers in the photograph reveal “evidence of edge burning” and “a heating to the ... preignition point.” When asked whether the temperature must have been 600 degrees for the papers to begin to char, as depicted in the photographs, Corry agreed and stated “[o]r more.” In contrast, DeHaan opined that temperatures inside the apartment never exceeded 400 to 450 degrees. Thus, the court's reliance on Corry's testimony to reject DeHaan's opinion that temperatures reached a peak of 450 degrees is misplaced because Corry himself testified that temperatures reached at least 600 degrees. Indeed, on cross-examination, Corry agreed that the temperature inside the victim's apartment could have been as high as 400 degrees even when firefighters entered the building. His only point, he explained, was that it could not have been 400 degrees at floor level because none of the firefighters, all of whom entered on their hands and knees within seconds of one another, was burned. Similarly unfounded is the third habeas court's assertion that, “[w]hen compared with Corry's reasonable and measured analysis, Kelder's estimate of temperature at 1800 degrees appears wildly exaggerated.” On cross-examination, counsel for the respondent asked Kelder his opinion as to “the maximum heat ... that was reached in the apartment....” Kelder responded: “The maximum, I would say, would be ... about 1800 degrees at ceiling level. ” (Emphasis added.) When asked what he had based this estimate on, Kelder responded, “scientific proof.... [A]ny of the ... fire books that I've read and the schools that I [have] attended [say] it's 1800 to 2000 degrees at ceiling level [for] any fire.... That's the normal ceiling [temperature].” Kelder then explained that temperatures are always much hotter at ceiling level. Corry never disputed this testimony. To the contrary, although he was never questioned about ceiling level temperatures, Corry stated that they necessarily would have been “hotter because the temperature rises as you go up, for sure.” He also stated that “heat stratifies. The highest temperatures are always found at the ceiling or close to the ceiling, and, as you go down, they get less, which is why little children are taught in schools to crawl in fires, because you're generally underneath a smoke layer, and there's more oxygen there and you can survive and actually escape a fire in many cases by crawling, [whereas] if you stood up, you might be killed or seriously burned.” Finally, as we previously indicated, Corry testified that temperatures would have to have reached at least 600 degrees to bring the newspapers to their preignition point, as depicted in some of the crime scene photographs. In view of this testimony, there simply is no factual basis for the third habeas court's finding that Kelder's estimate of ceiling temperatures was “wildly exaggerated” and, therefore, that his burn time estimate was unreliable. In light of the foregoing, it is readily apparent that the third habeas court's first two assertions—that Corry disputed DeHaan's and Kelder's testimony regarding peak temperature, and that, “[i]n contrast to ... DeHaan and Kelder, Corry opined credibly that ‘this was not a high energy fire’ ”—are not supported by the record. In fact, we can discern no disagreement among the experts with respect to the fire's peak temperature or energy level. It is not surprising, therefore, that the respondent makes no attempt to defend the third habeas court's assertion that there was a material difference in the testimony of the parties' experts with respect to the energy level of the fire and its peak temperature.

The third reason why the third habeas court found Corry's burn time estimate to be more persuasive than DeHaan's estimate was DeHaan's purported testimony that the floor level temperature inside the victim's apartment was 400 degrees when Tomkunas entered the apartment. Relying on Corry's testimony, the third habeas court reasoned that the temperature could not have been 400 degrees because, if the temperature had been that high, Tomkunas would have been injured, and he was not. Relying on Corry's characterization of DeHaan's testimony about the temperature at the time of entry, the respondent contends that, because DeHaan's burn time estimate rested in large measure on his assessment of entry level temperatures, “Tomkunas' lack of injury ... knocked [the] principal underpinnings of DeHaan's opinion out from under it.” As the petitioner notes, however, DeHaan never stated that the floor level temperature inside the apartment was 400 degrees when Tomkunas entered. He testified, rather, that the temperature inside the hot gas layer was likely between 300 or 400 degrees when Tomkunas entered the apartment. DeHaan explained that this estimate was based, in part, on Tomkunas' testimony at the petitioner's criminal trial that, when he arrived at the victim's apartment, the outside of the front door was hot to the touch, and “the temperatures and ... hot gases he encountered at near floor level were untenable, and he couldn't go in, even though he suspected there was a victim inside. ” (Emphasis added.) When asked whether he was able to determine what the temperature was at that time, DeHaan responded: “Only very approximately, [on the basis of Tomkunas' testimony] that, when he opened the door, there was smoke down very low, and he got on his hands and knees and he had no protective equipment on, and ... when he tried to crawl in, it was unpleasantly hot to his skin ... as well as smoky, and, at that point, he backed out knowing he couldn't go any further.... [S]hort exposure like that, that's probably going to be temperatures in the range of say [300] or 400 degrees.... [T]hat's enough to make most people not ... go in any further [with] ... bare skin.... [That level of temperature would cause second degree burns upon] prolonged contact....”

Later, during cross-examination, DeHaan clarified that his estimate of entry level temperatures referred to the temperature inside “the hot gas layer” of the fire. DeHaan explained that, in his experience, if the temperature had been much lower than 300 to 400 degrees, Tomkunas would not have been deterred from entering the victim's apartment, and, if the temperature had been much higher, there would have been more damage to the apartment. DeHaan also testified that studies have shown that the average person will not attempt to enter a room if the temperature is 250 degrees or higher, unless he has no option but to do so. Finally, DeHaan opined that, although an experienced firefighter might be willing to brave slightly higher temperatures, the fact that Tomkunas was hindered in his efforts to enter the victim's apartment suggests that the temperature in the hot gas layer of the fire was in the range of 300 to 400 degrees.

Notably, at the petitioner's criminal trial, Tomkunas testified that, when he first arrived on the scene, he tried to enter the victim's apartment on his own but was forced back by the substantial heat and “[h]eavy smoke condition.” Specifically, Tomkunas stated that, after he kicked in the front door, there “was a lot of heat, so [he] dropped to [his] knees ... to stay below [the] smoke and heat.” Tomkunas testified that he crawled several feet into the room but immediately had to retreat because “[i]t was too hot and smoky.” According to Tomkunas, after his initial, unsuccessful attempt to enter the apartment, he told the firefighters who arrived after him that they had to “vent” the building “to remove the heat and smoke and hot gases.” One of those firefighters, Boland, ran to the back of the building and opened a set of sliding glass doors. According to Boland, “once [he] opened the [sliding glass] door[s] ... [the apartment] cross vented, because the front door was open and the [sliding glass] door[s] [were] open, so it started to clear.” Both Tomkunas and Boland testified that, after the apartment was vented, they were able to enter and locate the victim. Consistent with Tomkunas' and Boland's trial testimony, DeHaan also testified at the petitioner's habeas trial that, once “conditions [inside the apartment] were changed by the responding firefighters ... Tomkunas' opening the front door and [Boland's] ... opening the slid[ing] [glass doors] at the rear, that allowed ... some of the accumulated smoke to leave and allowed entry for the firefighters. ” (Emphasis added.) Although it is perfectly clear that both DeHaan and Kelder were aware that firefighters vented the victim's apartment prior to entry, and that Tomkunas had tried unsuccessfully to gain entry into the apartment by himself prior to the arrival of the other firefighters, Corry either was unaware of these facts or refused to accept them. Corry maintained, throughout his testimony, that there was only one entry into the apartment by firefighters, which was not preceded by a venting. On direct examination, Corry was asked on several occasions whether he agreed with DeHaan's estimate of entry level temperatures. The first time he was asked this question, he responded: “[T]he fact that, [on] the night of the fire, all ... of the firefighters: [Tomkunas], William Boland ... William Parker and Douglas Boland, who were [all] in [the apartment] at basically the same time from different angles—none of them described high temperature on entry. They described heavy smoke. None of them talked about temperature. And, more importantly, none of them [was] injured in any way. And they would have been if the temperature had been anywhere close [to] 400 degrees, as ... DeHaan said.... They would have been burned.” Counsel for the respondent then asked Corry whether he had “reach[ed] a conclusion regarding the temperatures that Tomkunas and the other firefighters encountered when they entered the apartment....” Corry responded: “Safe to say, it was elevated 120, 130 [degrees], something like that.” Thereafter, the following exchange between counsel for the respondent and Corry ensued:

“Q. Do you recall reading ... DeHaan's report?

“A. Yes.

“Q. And you heard his testimony? “A. I did.

“Q. Do you recall what temperature he opined ... Tomkunas encountered when he entered the apartment?

“A. 400 degrees.

“Q. Based on your research, is that possible?

“A. It is not possible.

“Q. Would an individual who was not wearing protective clothing and entered [an apartment] with temperatures of 400 degrees be injured?

“A. Yes. Couldn't help it.

“Q. And would they be able to tolerate remaining in that environment?

“A. No. Well ... it all depends [on] how much they wanted to be burned, but they would be burned, and they'd be in substantial pain. And these were men who were, number one, on their hands and knees, which would expose the nape of their neck [and] the back of their ears, which are the most sensitive parts of our bodies, to temperature, and none of them described any burns or even mentioned heat in their original statements taken [on] the night of the fire.”

Thus, it is evident that Corry attributed to DeHaan testimony that he never gave, namely, that the floor level temperature inside the victim's apartment when the firefighters entered was 400 degrees. It also is evident that the third habeas court accepted Corry's characterization of DeHaan's testimony without verifying whether DeHaan actually had opined in the manner attributed to him by Corry. In fact, DeHaan never was questioned about the temperature inside the apartment after it was vented and all of the firefighters were able to enter. Nor did he express an opinion with respect to the floor level temperature when Tomkunas first entered the apartment on his own and was turned back. DeHaan simply stated that the temperature inside the hot gas layer was likely in the range of 300 to 400 degrees when Tomkunas made his first attempt to enter, before the arrival of the other firefighters. DeHaan also opined that the hot gas layer did not radiate much heat downward toward the floor because there was no damage to the apartment's synthetic carpet, which would have melted even at much lower temperatures.

Corry's testimony regarding the effects of heat on human skin, in particular, his statements that the “thermal tolerance data for unprotected skin of humans at rest would suggest a limit of about 248 degrees” and that the “upper level limit of survivability in terms of breathing [is] 300 degrees,” supports rather than contradicts DeHaan's testimony that the temperature inside the apartment was approximately 300 to 400 degrees when Tomkunas arrived. If the temperature had been much lower, an experienced firefighter like Tomkunas would not have been deterred from entering. As we have indicated, DeHaan testified that Tomkunas would have sustained second degree burns upon prolonged exposure to such temperatures. As DeHaan also testified, however, this explained why Tomkunas was unable to proceed farther into the apartment when he first arrived, even though he knew that there might be someone inside. As DeHaan stated, “it was hot enough to dissuade [Tomkunas from] entering ... the room.”

Moreover, it is difficult to fathom Corry's response when, on cross-examination, the petitioner's counsel pointed out to him that Tomkunas had testified at the petitioner's criminal trial that the heat and smoke prevented him from entering the apartment when, unaccompanied by other firefighters, he first attempted to do so: Corry's response was simply to refuse to accept Tomkunas' testimony. Specifically, when the petitioner's counsel asked Corry why Tomkunas' testimony that the smoke and heat forced him out of the building was inconsistent with DeHaan's testimony that the temperature was likely 300 to 400 degrees in the hot gas layer at that stage of the fire, Corry responded: “That's what [Tomkunas] said at the trial, not what he said in his initial statement.” In support of this response, Corry explained that all of the firefighters gave statements to the police on the night of the fire, and, although those statements are contained in a police report that never was entered into evidence, Corry read them in preparing to testify. According to Corry, the report does not mention that any of the firefighters either were injured or forced out of the apartment by the heat. Corry stated that, as far as he was concerned, the police report was a more reliable gauge of the conditions that Tomkunas had encountered on the night in question, more so than “what [Tomkunas] said under oath in [a] capital [felony] trial,” because, “in general, witnesses closer to the time of the incident will give a better recollection.”

Without the benefit of the report to which Corry was referring, we can only surmise its contents. It is unlikely, however, that the investigating officer who prepared the report questioned Tomkunas and his fellow firefighters in minute detail about their entry into the apartment. But even if we assume that the officer had thought to ask the firefighters about entry level temperatures, no doubt none of them reported extremely high temperatures inside the apartment because, as both Tomkunas and Douglas Boland testified, they vented the apartment before entering together as a group to extinguish the fire and to search for the victim. The fact that Corry felt free to disregard Tomkunas' sworn testimony on this issue, on the basis of what he himself conceded was nothing more than “a report written by a police officer [on the night of the fire] as to what somebody told him”—or did not tell him, as the case may be—is perplexing, to say the least. In any event, insofar as the third habeas court's determination that a jury reasonably could not credit the expert opinions of DeHaan and Kelder was predicated on Corry's mischaracterization of their testimony regarding the energy level and peak temperature of the fire, and entry level temperatures, that determination lacks support in the record.

It bears noting that the third habeas court also found that DeHaan's and Kelder's burn time estimates were “totally contradicted by the historical and physical evidence marshaled by Corry.” Because the respondent elected not to place Corry's written report in evidence, the only basis for this finding was Corry's trial testimony. When asked to explain why DeHaan's burn time estimate was unreliable, however, Corry could point only to DeHaan's testimony concerning the 400 degree temperature that the firefighters confronted upon entry—testimony that, as we have explained, DeHaan never gave. Corry explained that, if DeHaan was wrong that the floor level temperature was 400 degrees, “then he's wrong ... his assumptions are ... incorrect on the longevity of the fire....” Corry also stated that it was “clear that the fire occurred at an earlier time than what [DeHaan was] saying [because the fire] dropp[ed] in temperature enough ... so that Tomkunas and the other firefighters were not injured.” Because Corry's assessment of DeHaan's opinion concerning the burn time of the fire was predicated on a misconception of DeHaan's testimony about the entry level temperature, Corry's criticism of DeHaan's burn time estimate is similarly baseless.

Significantly, Corry's description of the fire and the variables that affected it was otherwise identical to DeHaan's and Kelder's analysis of the fire. Indeed, on cross-examination, Corry even conceded, consistent with DeHaan's testimony, that the temperature “might very well have been 400 degrees somewhere in that room” when Tomkunas entered the apartment. His point, he explained, was simply that it could not have been 400 degrees at floor level because none of the firefighters was burned when the firefighters entered and made their way into the interior of the apartment. Like DeHaan and Kelder, Corry testified that the fire burned out quickly due to the apartment's heavy insulation and lack of ventilation. Corry likened the apartment to a “thermos” and stated that the fire was “substantial” but burned only “for a brief period of time” due to “decreasing levels of oxygen....” On the basis of the photographs, Corry also opined that there might have been a blanket or some other object on the couch that prevented the fire from exhausting its main fuel source, which, according to Corry, was the couch cushions, and that that the blanket or object prevented the fire from becoming much larger. Apart from expressing the view that the firefighters were not injured when they entered the apartment, however, Corry offered no scientific reason for his conclusion that the fire could have been set considerably earlier than DeHaan and Kelder hypothesized. Rather, when asked whether he had reached a conclusion regarding the earliest time that the fire could have been set, Corry merely cited witness accounts of the victim's whereabouts on the day of the murder. Above and beyond our determination that the third habeas court's stated reasons for discrediting the burn time estimates of DeHaan and Kelder are baseless, our examination of the entire record reveals no other apparent reason why a jury would be apt to discredit their testimony. As we have indicated, their credentials certainly provide no such reason. Furthermore, although there is nothing in the record that would lead to the conclusion that a jury necessarily would credit Corry's burn time estimate over those of DeHaan and Kelder, there are several reasons why a jury might well be inclined to discredit that estimate. First, as we previously discussed, Corry's opinion is based on a police report that purportedly contains certain statements that firefighters gave to the police on the night of the fire but that never was introduced into evidence. In particular, Corry based his conclusions on inferences drawn from what Tomkunas did not say in those statements and elected to ignore Tomkunas' sworn trial testimony that directly contradicted the strained inferences Corry gleaned from the report not in evidence. Second, although counsel for the respondent opted not to introduce Corry's written report into evidence, the petitioner's counsel read extensively from it during his cross-examination of Corry. At the time, the petitioner's counsel accused Corry of having altered his conclusions about the fire after listening to DeHaan's testimony in an effort to create an opportunity for the respondent's counsel to argue that the fire could have been started earlier than DeHaan had determined. Although Corry denied the accusation, insisting that his testimony was wholly consistent with the findings in his report, he admitted that, after listening to DeHaan's testimony, he felt the need to reexamine the available evidence for the purpose of discrediting DeHaan's findings. He also acknowledged that of the more than ninety-five criminal cases he had worked on, he never had testified on behalf of a criminal defendant and would be “hesitant” ever to do so. Although there is nothing inherently untrustworthy about Corry's preference for working for the prosecution, certain portions of his testimony reasonably might be viewed as tending to support the petitioner's counsel's portrayal of him as someone more concerned with assisting the respondent than with seeking to establish the truth.

For example, on direct examination, Corry was asked a series of questions designed to elicit criticism of DeHaan's and Kelder's findings. With respect to Kelder, the questions focused on the information that he had relied on in formulating his burn time estimate, in particular, the fact that he had assumed, for purposes of his analysis, that the cushions on the victim's couch were made of latex foam. DeHaan had concluded that the cushions were made of polyurethane foam, a different material, and Corry had agreed with that conclusion at trial. In an effort to discredit Kelder's burn time estimate, counsel for the respondent asked Corry: “Would you be able to properly analyze this fire if you didn't know the difference between latex foam rubber and polyurethane foam?” Corry responded, “[n]o.” The next day, however, on cross-examination, it was revealed that Corry also had assumed that the cushions were made of latex foam for purposes of his analysis. Following that revelation, Corry provided a very different answer to the question, posed by the petitioner's counsel, of whether one could properly analyze the fire if he did not know that the couch cushions were made of polyurethane foam:

“Q. But, in your report, you have that ... the material was latex rubber.... Correct?

“A. Right.

“Q. Which is not what you're saying now?

“A. Well, there's a reason for that. This couch closely resembles couches that I've seen manufactured in the 1970s ... that did contain foam rubber. That was the principal foam material up [until] the time that polyurethane began to replace foam rubber, but it's really a matter of six of one, half dozen of another, since both of them perform in a similar manner in fires.... [Foam] rubber and foam plastics have similar heat release rates, fire propagation rates, ignition temperatures, and so forth.” These contradictory answers by Corry hardly instill confidence in his testimony.

Corry's testimony is replete with similar inconsistencies. Sometimes, he contradicted himself for no apparent reason, such as when he denied ever calling the fire a low energy fire. As we previously noted, at trial, the respondent's counsel focused on the fire's energy level and how it may have affected the fire's development, so much so that the third habeas court, in its memorandum of decision, included a finding that, “[i]n contrast to ... DeHaan and Kelder, Corry opined credibly that ‘this was not a high energy fire.’ ” As we explained, however, this finding lacks support in the record because both DeHaan and Kelder opined that the fire was a low energy fire. Nevertheless, on direct examination, Corry repeatedly characterized the fire as a “low energy” fire, on one occasion describing it as “a very low energy fire ” and asserting, at another point, that there was “abundant evidence that clearly shows ... that this fire was low energy....” (Emphasis added.)

On cross-examination, however, Corry denied ever characterizing the fire in terms of low energy. During one heated exchange with the petitioner's counsel, Corry was asked: “Now, as I understand your testimony ... [you said] that there's a low energy fire set in this couch, and this low energy fire kind of really becomes a smoldering fire, right?” Corry responded: “I didn't say it was a low energy fire. It's lower than it could have been.” The petitioner's counsel then asked Corry: “Well, I thought that's exactly what you said ... in your testimony on direct examination. That's exactly what you said; it was a low energy fire.” Corry responded, in part: “I don't believe I said anything like that.”

Needless to say, it is not the role of this court to make credibility determinations, and we therefore express no opinion as to whether Corry is a more persuasive or convincing witness than DeHaan or Kelder. Our sole purpose in drawing attention to the inconsistences in Corry's testimony is merely to underscore why, contrary to the conclusion of the third habeas court, the original jury might well have credited the burn time opinions of DeHaan and Kelder, and why it might well have rejected Corry's estimate. Of course, on retrial, it is possible that the jury will not find the testimony of DeHaan and Kelder persuasive; like all witnesses, they will be subject to impeachment, and, at that new trial, the state may be able to present an expert or experts who, in contrast to Corry, provide testimony explaining why the burn time opinions of the petitioner's experts lack persuasive force. Or the jury might conclude, like the third habeas court did, that it simply is not possible to determine the burn time of a fire, even within relatively broad parameters, with any degree of confidence or exactitude. But, for purposes of the present case, which involves the suppression of exculpatory evidence by the state, our task is not to determine whether the jury more likely than not would have credited their testimony, such that the petitioner would have prevailed at a new trial. Cf. Shabazz v. State, 259 Conn. 811, 827–28, 792 A.2d 797 (2002) (in resolving merits of petition for new trial based on newly discovered evidence, trial court must determine whether petitioner has established that new trial will produce different result). The question, rather, is whether the jury reasonably could have credited the testimony of the petitioner's witnesses. On the record before us, there simply is no legitimate basis for concluding that the jury reasonably could not have credited the testimony of DeHaan and Kelder, both of whom are well credentialed and highly experienced experts in their field, and whose testimony, when coupled with that of Martin's, supports the petitioner's contention that he could not possibly have sexually assaulted and killed the victim.

This observation, however, does not end our inquiry. We still must determine whether, if the original jury had considered the original trial evidence together with DeHaan's and Kelder's testimony that the fire was set no earlier than 7:30 p.m., as well as Martin's unwavering testimony that the petitioner was home with her and their son watching television at that time, there is a probability of a different result sufficient to undermine confidence in the jury's verdict. For the reasons set forth in part III C of this opinion, this standard has been met. Indeed, the state's less than compelling case against the petitioner was such that any new evidence tending to cast doubt on the petitioner's responsibility for the charged crimes could well have lead to an acquittal, and the petitioner's expert testimony concerning the burn time of the fire, coupled with Martin's testimony, certainly raises doubts about the reliability of the petitioner's conviction. Before turning to a discussion of the original trial evidence, however, we first address the three contentions raised by Justice Zarella in his dissenting opinion: (1) we have treated the respondent unfairly by deciding a claim that the parties have not raised; (2) we incorrectly have concluded that the Appellate Court properly exercised de novo review of the third habeas court's determination as to whether a jury reasonably could credit the testimony of the petitioner's experts; and (3) the record establishes, as the third habeas court determined, that the petitioner's expert testimony is not worthy of credit. We reject each of these claims.

B

Justice Zarella's Arguments

1

Justice Zarella contends that, in concluding that the third habeas court's materiality determination is not entitled to deference, we have “summon[ed] down [our] deus ex machina” and decided an issue that the parties never raised, thereby “silencing” the respondent, inflicting “acute” harm on the state and “undermin[ing] the fairness of our judicial process.” Justice Zarella's rhetoric may make for entertaining reading, but the facts categorically refute his accusations.

Some brief background is necessary in order to fully understand why Justice Zarella's argument is both unfaithful to the record and baseless. It is undisputed that the Appellate Court accorded no deference to the third habeas court's findings with respect to the credibility of the parties' burn time experts. See Lapointe v. Commissioner of Correction, supra, 138 Conn. at 476–79, 53 A.3d 257. According to Justice Zarella, however, the Appellate Court did not evaluate the testimony of the petitioner's experts, as Brady requires, to ascertain the materiality of the Ludlow note in light of the record as a whole but, rather, concluded that neither the Appellate Court nor the third habeas court had any role whatsoever in assessing the expert testimony. In other words, as Justice Zarella indicates, the Appellate Court never undertook an analysis of the testimony and, instead, merely “hypothesiz[ed]” what a jury could find if it credited the testimony and, on the basis of that “speculation,” concluded that “a new jury, and not the habeas court, ” should evaluate it. (Emphasis in original.) In Justice Zarella's view, because this approach obviously is improper, and because, according to Justice Zarella, the petitioner's only claim is that we should approve the Appellate Court's use of such a methodology, our analysis begins and ends there: the respondent prevails on appeal, the judgment of the Appellate Court is reversed, and the judgment of the third habeas court denying the habeas petition is reinstated. Proceeding any farther, Justice Zarella asserts, would take us beyond the scope of the claims raised in this appeal.

Justice Zarella's assertion to the contrary notwithstanding, there is absolutely no basis for presuming that the Appellate Court concluded that neither it nor the third habeas court had a role in evaluating the credibility of the parties' expert testimony and that those courts' sole responsibility, instead, was to speculate as to what the original jury might have concluded if it had credited the petitioner's expert testimony. Under this approach, any defendant who makes a claim under Brady —no matter how preposterous or outlandish the testimony that provides the basis for that claim—automatically would be entitled to a new trial if, assuming that a jury believed it, that jury might find the defendant not guilty. No one, let alone a court of law, ever would advocate for such a ludicrous standard, as it is perfectly obvious that a defendant cannot be entitled to a new trial under Brady when an objective assessment of the testimony on which the defendant's claim is based reveals that it is incredible or unworthy of belief. Because there is no support in the law of this or any other jurisdiction for the bizarre proposition that our courts play no part in evaluating the credibility of evidence presented to establish a Brady claim—and because common sense dictates that that approach cannot possibly be correct—it verges on insulting to presume that the Appellate Court employed such a patently improper and utterly unworkable approach.

Moreover, a review of the Appellate Court's decision positively belies any such presumption. First, the Appellate Court discussed the facts of the case at length; Lapointe v. Commissioner of Correction, supra, 138 Conn.App. at 456–58, 462–63, 468–73, 53 A.3d 257 ; and summarized the parties' expert burn time testimony in the context of the relevant facts. Id., at 472–73, 53 A.3d 257. The Appellate Court also explained the decision and factual findings of the third habeas court; id., at 472–74, 53 A.3d 257 ; and observed that the third habeas court had found that the petitioner's experts provided thorough and extensive testimony concerning the fire's burn time. Id., at 473, 53 A.3d 257. The Appellate Court then properly set forth the standard of review and governing legal principles under Brady and Strickland; id., at 474–76, 53 A.3d 257 ; and, in so doing, quoted from this court's decision in State v. Ortiz, supra, 280 Conn. at 720, 911 A.2d 1055, in which we explained that, on appeal, the issue of materiality presents a mixed question of law and fact, with the trial court serving as the fact finder. Lapointe v. Commissioner of Correction, supra, 138 Conn.App. at 475, 53 A.3d 257. Ultimately, on the basis of the testimony of the parties' experts, the Appellate Court concluded that the petitioner is entitled to a new trial, and, as Justice Zarella acknowledges, it did so without affording any deference to the third habeas court's contrary findings. See id., at 476–79, 53 A.3d 257. Because, under our law—and so far as we know, the law of every other jurisdiction—the only alternative to deferential appellate review is de novo appellate review, it is apparent that the Appellate Court engaged in that latter form of review. As we have explained, that is the correct standard of review in the present case.

Finally, in defending his interpretation of the Appellate Court's decision, Justice Zarella asserts that there is no indication in that decision that the Appellate Court evaluated the petitioner's expert testimony to determine whether a jury reasonably could credit it. Justice Zarella argues, rather, that the Appellate Court simply concluded “that assessing the credibility of the expert witnesses was a task best left to a jury,” without any assessment of their testimony either by the trial court or by the Appellate Court. In support of his assertion that the Appellate Court concluded that no judicial review of the parties' expert testimony was the proper approach, Justice Zarella quotes the following language from a footnote in the Appellate Court's decision: “If the Ludlow note had been disclosed to trial counsel, however, it would have been the responsibility of the jury and not the court to weigh the credibility of the arson experts. Whether the burn time evidence, which was so critical in buttressing [the petitioner's] alibi defense, raised a reasonable doubt as to the petitioner's guilt would best be a determination left to the jury and not [to] a habeas court.” (Emphasis omitted; internal quotation marks omitted.) Part II of Justice Zarella's dissenting opinion, quoting Lapointe v. Commissioner of Correction, supra, 138 Conn.App. at 476–77 n. 17, 53 A.3d 257. The portion of the Appellate Court's decision that Justice Zarella fails to mention—the portion that places the language on which Justice Zarella relies in its proper context—leaves no doubt that the Appellate Court reached its conclusion that the petitioner is entitled to a new trial under Brady only after undertaking an independent review of the parties' expert testimony.

More specifically, after stating that the testimony of the petitioner's two experts would have been critical to establishing an alibi, the Appellate Court “conclude [d ] that there is a reasonable probability that the result of [the petitioner's ] criminal trial would have been different had the Ludlow note been disclosed to [the petitioner's trial counsel ] prior to [his criminal ] trial. Nondisclosure prior to trial of the portion of the Ludlow note describing the possible burn time affected the overall fairness of the trial and was so unfair as to undermine our confidence in the jury's verdict. With the burn time estimate provided by one of the state's fire marshals, trial counsel testified that they would have retained the services of an arson expert and that ... Martin would have testified as to the petitioner's whereabouts during the critical times of that evening. That evidence, if believed by the jury, could have resulted in the jury's finding that it was temporally impossible for the petitioner to have committed the crimes [of] which he was convicted. The Ludlow note was exculpatory and material in these circumstances. [First habeas counsel's ] performance was deficient when he failed to pursue that issue at the first habeas proceeding, and the petitioner was prejudiced by his failure to do so. The petitioner has demonstrated that had there been effective representation by [first habeas counsel ], there is a reasonable probability that the first habeas court would have found that the petitioner was entitled to [the ] reversal of [his ] conviction and a new trial. ” (Emphasis added; footnote omitted.) Id., at 478–80, 53 A.3d 257.

It is simply impossible to read this passage and conclude that the Appellate Court did not undertake a substantive and independent review of the petitioner's expert testimony. Only by evaluating the substance of the testimony and concluding that it was sufficiently credible could the Appellate Court have concluded, as it did, that the petitioner's inability to use it at trial adversely “affected the overall fairness of [the petitioner's criminal] trial and was so unfair as to undermine [the Appellate Court's] confidence in the jury's verdict.” Id., at 478–79, 53 A.3d 257. In other words, unless the Appellate Court actually examined and assessed the evidence, there is no way that the Appellate Court could have determined that, if the expert testimony had been available to the petitioner at his criminal trial, there is a reasonable probability of a different verdict, as Brady requires. Not surprisingly, the respondent does not claim that the Appellate Court concluded that neither it nor the third habeas court was obligated to review the expert testimony but, rather, was to merely assume that such testimony was credible. In fact, the respondent expressly recognizes that the Appellate Court did review the parties' expert testimony. For example, the respondent claims that, “[a]s for the testimony of Martin, Kelder and DeHaan, the Appellate Court's conclusion that this evidence would have supported the [petitioner's ] alibi defense is not sustainable.” (Emphasis added.) The respondent also claims that the Appellate Court “failed to consider the [state's] compelling evidence of guilt” in conducting its review of the testimony of DeHaan, Kelder and Martin. At no time has the respondent ever suggested that the Appellate Court viewed itself and the third habeas court as having no role in evaluating the credibility of the parties' expert testimony. Nor has the respondent ever claimed that the Appellate Court did not conduct a substantive, de novo review of that testimony. He argues, rather, that the Appellate Court improperly failed to defer to the determination of the third habeas court in concluding that a jury reasonably might credit the petitioner's experts. Consistent with this contention, the respondent's brief contains a thoroughgoing discussion of why the Appellate Court's review of the expert testimony should have been tempered by deference to the findings of the third habeas court. Because the respondent himself claims that the Appellate Court should have deferred to the conclusion of the third habeas court, rather than conducting a de novo review, it defies credulity to assert that the respondent was not on notice that we would decide the claim, as Justice Zarella asserts. Ironically, then, it is Justice Zarella who would decide this appeal on the basis of a claim that has not been raised or briefed by the parties, that is, the propriety of the Appellate Court's purported determination that a Brady violation automatically entitles a defendant to a new trial, without any judicial evaluation of the credibility of the testimony that supports a finding of that violation.

Having asserted, wrongly, that the respondent lacked notice that we would decide whether the third habeas court's determination was subject to deferential or de novo review, Justice Zarella compounds his error by alleging that the respondent also could not have known that we might decide that issue against him and, if we did, that we then would proceed to the issue of whether the Appellate Court properly concluded that the third habeas court had resolved the issue incorrectly. This argument fails, of course, because it is based on a false premise. Because the respondent asked us to decide the proper standard of review, it hardly could be clearer that he was on notice that, if we disagreed with his contention that the Appellate Court improperly employed a de novo standard of review, we also would decide whether the Appellate Court, having applied that standard, reached the correct conclusion in reversing the judgment of the third habeas court. Otherwise, we would be unable to determine whether the judgment of the Appellate Court should be affirmed or reversed. See State v. Fausel, 295 Conn. 785, 793, 993 A.2d 455 (2010) (“in a certified appeal, the focus of [this court's] review is ... [on] the actions of the Appellate Court” [internal quotation marks omitted] ). We therefore reject Justice Zarella's manufactured allegation that we have “silenc[ed]” the respondent by “denying [him] notice and a chance to brief the issue....” In light of the foregoing, it is apparent that we have not decided an issue that the parties have not raised. On the contrary, the respondent himself raised the issue about which Justice Zarella complains and then fully briefed it. Clearly, Justice Zarella's assertion that it is somehow unfair of us to decide the respondent's claim against him fails because it is against all logic and common sense. As we have explained, in arguing that the judgment of the Appellate Court should be reversed because that court merely assumed the credibility of the petitioner's expert testimony without evaluating that testimony, it is Justice Zarella who would resolve this appeal on the basis of a claim that has not been raised. That would be unfair.

2

Justice Zarella also challenges our determination that the Appellate Court properly exercised de novo review over the third habeas court's predictive judgment as to whether a jury reasonably could have credited the petitioner's experts. He contends that our conclusion is foreclosed by our case law, that we have improperly usurped the function of the third habeas court in reviewing its findings de novo because that court necessarily was in a better position to assess credibility than we are, and that employing that standard of review in the present case violates the state constitutional prohibition against fact-finding by an appellate tribunal. We disagree. With respect to Justice Zarella's first claim, we have explained, in part II of this opinion, why, like the courts in Bunch v. State, supra, 964 N.E.2d at 292–93, and State v. Behn, supra, 375 N.J.Super. at 431–33, 868 A.2d 329, we are convinced that the Appellate Court properly exercised de novo review over the third habeas court's conclusion concerning the likelihood that a jury would credit the petitioner's experts, and we need not belabor our reasoning here. Suffice it to say that this case presents a highly unusual scenario: the third habeas court's probabilistic assessment was predicated solely on the scientific opinions of highly qualified experts whose character, reliability and veracity never have been questioned, and the third habeas court's ability to evaluate their credibility firsthand, on the basis of their performance on the witness stand, had no bearing on that court's judgment. The court's decision, rather, was based solely on its view that the foundation of the opinion expressed by the respondent's expert was sounder than the opinions of the petitioner's experts. In such circumstances, we are in as good a position as the third habeas court to gauge whether a jury reasonably could credit the opinions of the petitioner's experts. Although we agree fully with Justice Zarella that the general rule is one of deference, even in cases involving claims under Brady, this case represents a limited exception to that rule because, as we previously discussed; see footnotes 42 and 43 of this opinion and accompanying text; there is no justification for deferring to the third habeas court's findings and compelling reason not to do so. Justice Zarella's observation that we cannot cite an appellate case from this state in which we have applied such an exception proves only one thing: we previously have not had a case on all fours with this one. The fact that such a case is so uncommon, moreover, is a complete answer to Justice Zarella's “the sky is falling” contention that we have opened the floodgates to future litigants, for only in the rare case that a litigant can establish that his case is materially similar to this one will he be entitled to de novo review of the lower court's materiality determination.

Justice Zarella accuses us of improperly usurping the fact-finding role of the third habeas court because, in his view, “[e]valuating the credibility of a witness' oral testimony is a complex process that always entails consideration of subjective factors like the attitude, candor and demeanor of the witness, and this assessment cannot be based solely on objective factors reflected in the printed record.” (Emphasis in original.) We disagree that the assessment of testimony, in particular, the evaluation of expert testimony by the court, necessarily and invariably involves consideration of the subjective factors that Justice Zarella identifies. This is a case in point. As we have explained, there is nothing in the record to suggest that the decision of the third habeas court was predicated on the court's subjective view of the credibility of the parties' experts, and the court's explanation of its decision demonstrates convincingly that subjective considerations did not cause the court to credit the respondent's expert, Corry, over the petitioner's experts, DeHaan and Kelder. Rather, as the third habeas court expressly stated, it was persuaded by the soundness of Corry's scientific opinion and by Corry's critique of the testimony of DeHaan and Kelder. In view of the objective reasons that the third habeas court gave for its conclusion, and in the absence of any indication, expressed or otherwise, that the court relied on purely subjective factors in discrediting the petitioner's experts, we will not presume that it did so. This conclusion is reinforced by the fact that, at oral argument before this court, the respondent's appellate counsel acknowledged that the third habeas court had “to fully articulate its reasons for making a credibility determination” and, further, that “we have a habeas court here making a credibility determination and assigning very specific reasons for why [it] did that.” (Emphasis added.)

With respect to Bunch v. State, supra, 964 N.E.2d 274, Justice Zarella makes three arguments in an attempt to discredit our reliance on the reasoning of that case. None of those arguments undermines the validity of that reasoning to even the slightest degree. First, he suggests that Bunch is distinguishable because it involves a petition for a new trial based on newly discovered evidence, whereas the present case involves a claim founded on Brady and Strickland. Justice Zarella further asserts that Indiana applies a different standard for determining the merits of new trial petitions than we do in this state. These are classic distinctions without any difference. The fact that the court in Bunch addressed a claim based on newly discovered evidence is wholly irrelevant because that claim requires exactly the same analysis as claims under Brady and Strickland, as they entail the same considerations. With respect to the argument that the elements of a new trial claim based on newly discovered evidence differ in Indiana and Connecticut, this, too, is irrelevant because the precise nature of those elements has absolutely nothing to do with the separate and distinct issue of whether the findings of either state's trial courts are entitled to deference in those cases.

Justice Zarella next argues that Bunch is inconsistent with this state's jurisprudence because our cases reveal that we have erected an absolute bar to de novo review of findings by a trial court, no matter what the circumstances. This is simply not true. When there is sound reason not to defer to the trial court—that is, when we are in as good a position as the trial court to decide the issue—we need not, and will not, do so. Like Bunch, this is such a case, and Justice Zarella has failed to identify any flaw in the reasoning of Bunch that would warrant a contrary conclusion.

Justice Zarella's third and final contention is that engaging in de novo review of the third habeas court's materiality determination runs afoul of the state constitutional ban on fact-finding by an appellate tribunal. See, e.g., Styles v. Tyler, 64 Conn. 432, 461, 30 A. 165 (1894). The respondent has not raised a constitutional objection to the Appellate Court's allegedly improper failure to defer to the findings of the third habeas court, and it therefore is not proper for Justice Zarella to do so. Because Justice Zarella addresses it, however, we also will do so. Review of the claim makes clear why the state did not raise it: it is devoid of merit.

In Styles, this court discussed the division of authority between trial courts and appellate courts as contemplated by the state constitution, and explained that “pure issues of fact” fall outside the authority of the latter. Id. In other words, appellate courts do not try cases or retry cases on appeal. As the court in Styles made clear, however, the fact-finding that is constitutionally prohibited does not include appellate review of facts found by the trial court. See id., at 459, 30 A. 165. Furthermore, “[a]mong the questions of law belonging to the jurisdiction of this court ... are ... questions of legal conclusion when law and fact are so intermingled that the main fact is not a pure question of fact but a question of the legal conclusion to be drawn from subordinate facts and also questions whether particular subordinate facts constitute the basis for a conclusion of fact or a conclusion of law....” Id. As we have explained, materiality under Brady and Strickland is a mixed question of law and fact.

In the present case, de novo appellate review of the third habeas court's materiality decision does not impermissibly intrude on that court's fact-finding function because it does not require appellate fact-finding at all. Our review of the testimony of the parties' experts does not require us to disturb the third habeas court's relative credibility judgment, that is, its determination that the respondent's expert was more credible than the petitioner's experts. Nor does it deprive the third habeas court of its responsibility to evaluate that expert testimony on the basis of its firsthand assessment of the parties' experts, because the court's ultimate conclusion with respect to the credibility or persuasiveness of the experts' opinions was not based on its superior position to evaluate them. This is so because the third habeas court made no factual findings that derived from its ability to assess the credibility of the experts in light of their conduct and demeanor on the witness stand. Indeed, the third habeas court's assessment did not require any findings at all with respect to disputed issues of historical fact. Rather, the third habeas court's probabilistic judgment was predicated on its belief, based solely on the foundational underpinnings of the various expert opinions, that the opinion of the respondent's expert was sounder, and therefore more deserving of credit, than those of the petitioner's experts. Because we are in as good a position as the third habeas court to evaluate the substance of those opinions, we are also as well positioned to decide whether a jury reasonably might credit them. For that reason, our de novo review of the third habeas court's predictive judgment does not usurp the role of that court in any way, and, consequently, there is no reason for us to defer to it. In such circumstances, the issue gives rise to a question of law, and there is no constitutional violation.

Although the foregoing discussion resolves Justice Zarella's constitutional claim, it bears noting that Justice Zarella has a very broad view of the scope of the constitutional ban on appellate fact-finding. This view, however, ignores the fact that our appellate authority extends to actions that implicate the fact-finding function to a far greater degree than the de novo review that we exercise in the present case. For example, under the second prong of the clearly erroneous standard, we may substitute our view for that of the fact finder when, although there is sufficient evidence in the record to support a particular finding, the reviewing court nevertheless “is left with the definite and firm conviction that a mistake has been committed.” (Internal quotation marks omitted.) State v. Maurice M., 303 Conn. 18, 27, 31 A.3d 1063 (2011) ; see also id., at 37, 44, 31 A.3d 1063 (reversing conviction of defendant, despite evidentiary support for jury's finding of guilt, because this court had definite and firm conviction that mistake had been made).

Another such example is State v. Kitchens, 299 Conn. 447, 10 A.3d 942 (2011), in which we concluded that defense counsel's failure to raise a constitutional challenge to the trial court's jury instructions may, depending on the circumstances, constitute a waiver of the defendant's right to raise such a claim on appeal. See id., at 482–83, 10 A.3d 942. As we explained in Kitchens, because such a waiver is implied rather than express, it “arises from an inference that the defendant knowingly and voluntarily relinquished the right in question.” (Emphasis omitted.) Id., at 483, 10 A.3d 942. As this court also has observed, the decision whether to draw an inference in any given set of circumstances is a quintessential “question of fact, to be answered by a fact finder.” State v. Diaz, 226 Conn. 514, 526 n. 8, 628 A.2d 567 (1993) ; see also id. (“The process of making a legal determination of whether a particular inference drawn by a fact finder is reasonable, as opposed to the process of determining whether to draw such an inference, is particularly familiar to a reviewing court. That is the essence of the difference between a question of law, to be answered by a reviewing court, and a question of fact, to be answered by a fact finder.”). Despite the fact intensive nature of this inquiry—as we explained in Kitchens, whether to draw an inference of waiver “must be based on a close examination of the record and the particular facts and circumstances of each case”; State v. Kitchens, supra, at 483, 10 A.3d 942 —this court, and not the trial court in which the alleged waiver occurred, is responsible for deciding, as a matter of fact, whether counsel waived the claim. See id. Nevertheless, this decision is considered a question of law over which we exercise plenary review. E.g., State v. Davis, 311 Conn. 468, 477, 88 A.3d 445 (2014). It simply cannot be that the constitution permits the kind of fact bound determinations that are required under the second prong of the clearly erroneous test and under our Kitchens waiver methodology but bars plenary review in the present case, as Justice Zarella maintains.

3

Justice Zarella argues that, notwithstanding the myriad factual errors contained in the third habeas court's memorandum of decision, the record nevertheless supports the conclusion that no jury reasonably could credit the petitioner's experts. To this end, Justice Zarella makes three primary arguments in an effort to demonstrate why the testimony of the petitioner's experts “simply is not reliable....” They are: (1) the fact that Tomkunas was not burned upon entering the victim's apartment renders DeHaan's burn time estimate unreliable; (2) DeHaan mistakenly believed that the temperature inside the victim's apartment prevented Tomkunas from entering the apartment when he first arrived; and (3) DeHaan's testimony was internally inconsistent and therefore unreliable. None of these arguments, however, is supported by the record.

First, Justice Zarella argues that, because Tomkunas crawled several feet into the victim's apartment when he first arrived, and remained inside for as long as fifteen to twenty seconds before being forced to retreat, the fact that Tomkunas was not burned “completely contradicts DeHaan's claim that the [victim's] apartment was close to 400 degrees when Tomkunas first kicked in the door.” Justice Zarella's argument is predicated on the same fundamental misunderstanding of DeHaan's testimony that led the third habeas court so far astray, namely, that DeHaan testified that floor level temperatures were 400 degrees when Tomkunas entered the building. As we have explained, DeHaan never was asked about floor level temperatures. More importantly, he made it clear that, when he spoke of entry level temperatures in the 300 to 400 degree range, he was referring to temperatures in the hot gas layer of the fire, not floor level temperatures. See footnote 47 of this opinion and accompanying text. Indeed, DeHaan was of the view that not even the hot gas layer temperature reached higher than 450 degrees at the fire's peak, and that it radiated very little heat downward toward the floor because the victim's synthetic rug was not damaged in any way. See footnote 48 of this opinion. DeHaan explained that, “even under fairly low radiant heat intensities, synthetic fabrics like carpets tend to start to melt, and, when they do, they get kind of crispy to the touch,” which did not happen to the victim's carpet.

Despite the clarity of DeHaan's testimony with respect to the meaning of “entry level temperatures,” Justice Zarella insists that “DeHaan estimated that Tomkunas would have experienced temperatures near 400 degrees ‘to the bare skin,’ ” and, therefore, the third habeas court properly discredited his burn time estimate. Only by taking a handful of DeHaan's words out of context and ignoring the remainder of his testimony is it possible for Justice Zarella to attribute to DeHaan the view that floor level temperatures were 400 degrees when Tomkunas entered the building, or at any other time. The testimony that Justice Zarella relies on to support this argument is the last four words of DeHaan's response to the questions of whether, in the course of his investigation, he was able to determine entry level temperatures, and what bearing those temperatures had on his estimate of the fire's burn time. With respect to whether he was able to determine entry level temperatures, DeHaan responded: “Only very approximately, [on the basis of Tomkunas' testimony] that when he opened the door, there was smoke down very low, and he got on his hands and knees and he had no protective equipment on, and ... when he tried to crawl in, it was unpleasantly hot to his skin ... as well as smoky, and, at that point, he backed out knowing he couldn't go any further.... [S]hort exposure like that, that's probably going to be temperatures in the range of say [300] or 400 degrees.... [T]hat's enough to make most people not ... go in any further [with] ... bare skin.” We do not understand DeHaan to be saying that Tomkunas' bare skin was exposed to 400 degree temperatures. To the extent that there may be ambiguity in this portion of DeHaan's testimony, however, it is entirely dispelled by the remainder of his testimony. As we previously indicated; see footnote 47 of this opinion; on cross-examination, the respondent's counsel asked DeHaan exactly what he meant when he “estimated that the temperature in the apartment when ... Tomkunas entered was 300 [to 400] degrees....” DeHaan responded, “I'm talking about the hot gas layer temperature....” DeHaan then explained that, in his view, the hot gas layer never was much hotter than 400 degrees because there was so little damage to the rug, which would have melted at much lower temperatures. Moreover, it does not take an expert in fire forensics to realize that a person would be severely burned from any kind of prolonged exposure to 400 degree temperatures. To suggest that a person of DeHaan's professional standing and expertise was unaware of this fact is patently unreasonable.

Accordingly, contrary to Justice Zarella's assertions, it is not in the least surprising that Tomkunas was not burned when he first crawled into the apartment, particularly in light of his testimony that he was on his hands and knees the entire time, intentionally staying below the hot gas layer. As Corry himself stated, temperatures are always much hotter the higher up you go in a room, which is why children are told to crawl on their hands and knees in a fire. In this way, Corry explained, they may avoid injury, even in situations where, if they were to stand up, they could be severely burned or killed.

Indeed, the entire issue of whether it was 400 degrees at floor level when Tomkunas entered the victim's apartment seems little more than a construct, inserted into the proceedings by Corry two months after DeHaan had testified. As we previously indicated, Corry stated that, after listening to DeHaan's testimony, he went back to the drawing board in an effort to refute it. What he came up with, it appears, was the notion that DeHaan had stated that it was 400 degrees at floor level when the firefighters entered the apartment, and that the temperature obviously could not have been that high because none of the firefighters was burned. Specifically, Corry stated that “DeHaan said [it was] 400 degrees. That's what we were investigating.” Corry further stated that, “if it was 400 degrees at the floor, [the fire] would have burned [the firefighters], and it didn't.” Notably, on cross-examination, Corry admitted that it could very well have been 400 degrees in the hot gas layer when Tomkunas entered the apartment. Corry explained that his only point was that it could not have been 400 degrees at floor level because none of the firefighters was burned. Indeed, it bears emphasis that Corry's acknowledgment that the temperature in the hot gas layer could have been 400 degrees when Tomkunas arrived fully supports DeHaan's estimate of the fire's burn time, which, as we previously discussed, was predicated on hot gas layer temperatures being in the range of 300 to 400 degrees.

In a similar vein, Justice Zarella asserts that DeHaan's burn time estimate is unworthy of credit because it is predicated on DeHaan's mistaken belief that the heat prevented Tomkunas from crossing the threshold of the victim's apartment when he first arrived. Specifically, Justice Zarella argues that DeHaan appears not to have been aware that Tomkunas was able to crawl several feet into the apartment before being forced out by the heat, which likely affected DeHaan's calculation of the entry level temperature. We disagree with Justice Zarella's view of the record, which clearly reflects the fact that DeHaan read all of Tomkunas' trial testimony and was fully aware that Tomkunas entered the victim's apartment briefly, on his hands and knees, before being forced out by the heat. In fact, not only did DeHaan testify that Tomkunas entered the apartment on his hands and knees before being forced out, his written report on the fire, which was entered into evidence as a full exhibit, provides in relevant part: “[Tomkunas] kicked the door open to find heavy smoke and considerable heat within. Without protective equipment, he crawled on hands and knees into the front entry far enough to see flaming fire on the couch against the left wall but was forced out by the conditions. ” (Emphasis added.) In light of this evidence, Justice Zarella's assertion that DeHaan was unaware that Tomkunas actually proceeded a short distance into the apartment when he first arrived is simply wrong.

We are also perplexed by Justice Zarella's insistence that, “[a]t no time during his [trial] testimony did [Tomkunas] say that he was prevented from even entering the apartment as a result of the heat.” Justice Zarella contends that, after his initial entry, Tomkunas left the apartment only because “the heat and smoke conditions convinced him that rescue efforts would be ‘a lot easier’ if the apartment was vented,” not because it was too hot to remain inside. As with Justice Zarella's other factual arguments, this contention is contradicted by the record. When initially asked whether he was able to remain in the apartment after kicking in the front door, Tomkunas responded: “No, I was not.... It was too hot and smoky. I had to go ... out.” He also stated that the heat and the smoke, which he described as “very, very dangerous,” “forced [him] to retreat....” Tomkunas further stated that, before attempting a second entry, it was necessary to “smash a window ... to remove the heat and smoke and hot gasses from the building.” In arguing that the heat was not so great as to prevent Tomkunas from remaining in the building, Justice Zarella relies on a single statement by Tomkunas: “If I could get somebody to vent, it was a lot easier.” The broader context of that statement, however, as reflected in the following colloquy between the state and Tomkunas, which Justice Zarella has omitted from his dissenting opinion, leaves no doubt that Tomkunas, upon his arrival at the scene, was forced from the apartment due to the high heat and dangerous smoke conditions inside the apartment:

“Q. Now, with respect to the volume of smoke, how would you characterize that when you first entered—what [was] the volume of smoke ... in the apartment?

“A. If I could use a technical term, ‘well charged.’

“Q. And is that a term of significance as far as firefighting is concerned?

“A. Yes it is.

“Q. And how would you explain it to us?

“A. A building that's well charged with smoke is so full, there's no—there's actually little to no oxygen left in the building, and it's a very, very dangerous situation.

“Q. Dangerous to whom ...?

“A. It would be dangerous to anybody in there and anybody who attempted to get in without proper precautions.

“Q. And you've indicated that the heat and the smoke forced you to retreat after some fifteen to twenty seconds. Is that correct?

“A. That's correct.

* * * “Q. Can you tell us why you retreated?

“A. It was just too hot and too smoky. If I could get somebody to vent, it was a lot easier.”

Justice Zarella also argues that DeHaan's testimony “that entry level temperatures were too high for Tomkunas even to enter the apartment conflicts with DeHaan's own testimony about the fire's energy level” because “DeHaan testified that the fire would have produced a relatively small amount of heat and would have been approachable, even at its maximum intensity. DeHaan explained that the fire's intensity ‘would probably be about the same as an average fireplace fire. It would be pumping a lot of heat into this room, but not so much that you couldn't—you couldn't approach it, for instance, to try to extinguish it....’ ” (Emphasis omitted.) For reasons that we previously discussed, we reject Justice Zarella's assertion that DeHaan testified that entry level temperatures prevented Tomkunas from entering the apartment at all when he first arrived. We also do not perceive any conflict between DeHaan's testimony regarding the fire's energy level and his estimate of entry level temperature. When DeHaan likened the fire's energy level to that of an average fireplace fire, it was solely in an attempt to explain why there was so little damage to the victim's apartment. DeHaan attributed the lack of damage in the apartment to the heat release rate of the fire, which he estimated never exceeded 250 to 350 kilowatts. DeHaan explained that a 350 kilowatt fire will produce temperatures in the range of 400 degrees. DeHaan further explained that, if there had been more ventilation inside the apartment, “a love seat like [the one in the victim's apartment] is capable of generating [a] maximum heat release rate ... in the order of 2000 kilowatts. In other words, it would have been six times [as] big [of a] fire; the fire would have basically reached the ceiling, actually spread across the ceiling, and it might have actually brought this whole room to involvement.” If this had occurred, DeHaan testified, it would not have been possible for a firefighter simply to walk up to the fire and extinguish it. In contrast, a firefighter easily could have approached the fire on the victim's couch, assuming, of course, that he or she was wearing appropriate attire, because the fire's heat release rate never exceeded 350 kilowatts, and the fire never spread beyond the couch to other materials in the room. Indeed, it was undisputed that, with proper clothing and breathing equipment, which Tomkunas did not have with him when he arrived, he would have had no difficulty approaching and extinguishing the fire upon his initial entry. In addition to his primary arguments, Justice Zarella also asserts that the testimony of Igoe, Roy and Christopher Marvin, three fire investigators who were at the crime scene on the night of the murder, supports the third habeas court's determination that it was not possible to ascertain the fire's approximate burn time. We disagree with Justice Zarella's interpretation of the testimony of these witnesses, all of whom stated unequivocally that they did not perform the type of investigation that would have enabled them to estimate when the fire started. For example, one of the witnesses, Roy, testified that he had no recollection of the events of that evening and that his role at the time was limited to moving furniture, sweeping the floor, and the like. Although we also dispute the remainder of Justice Zarella's factual contentions, we need not belabor the matter further because, even if we were to assume that a jury reasonably could find Corry's testimony more persuasive than that of DeHaan and Kelder, that would not alter our conclusion that the third habeas court's materiality determination was incorrect as a matter of law. This is so because, as we have indicated, there is absolutely nothing in DeHaan's or Kelder's testimony, or anywhere else in the record, that would permit the conclusion that a jury reasonably could not credit their testimony.

Having addressed Justice Zarella's claims, we now turn our attention to the second component of the test for materiality under Brady and Strickland, that is, consideration of the petitioner's expert testimony in light of the original trial evidence, for the purpose of determining whether our confidence in the verdict is undermined by that expert testimony. As we explain, the answer to that question is most certainly “yes.” C

Original Trial Evidence

It is beyond dispute that the state's case against the petitioner rested almost entirely on his incriminating statements, without which the state would have been unable to obtain a warrant based on probable cause, let alone a conviction of capital felony. There was no physical evidence connecting the petitioner to the crime, except for the fact that, along with approximately one third of the male population, including another suspect in the present case, the petitioner is a secretor with type A blood, the same blood type as the person who left a semen stain on the victim's bedspread. The state also sought to link the petitioner to the victim's murder because he previously had undergone a vasectomy, and no sperm were found in the semen stain. At trial, however, Beryl Novitch, the lead criminalist at the state forensic laboratory who examined the stain at issue, testified that it was not at all unusual to obtain a semen sample containing no sperm even though the donor's semen contains sperm.

With respect to the petitioner's statements, the petitioner alleged in his first habeas petition that his confession was false, and, in furtherance of this claim, he maintained that he was unable to knowingly, intelligently and voluntarily participate in the police interrogation because of his mental impairment. The first habeas court dismissed this claim upon concluding, inter alia, that, because “[t]he petitioner's real argument ... is that his statements were involuntary [and therefore inadmissible] under the federal constitution”; Lapointe v. Warden, Superior Court, judicial district of Hartford, Docket No. CV–97–0571161, 2000 WL 1409721 (September 6, 2000) (Freed, J. ); the claim was barred by this court's determination, in connection with the petitioner's direct appeal from the judgment of conviction, that the trial court reasonably had concluded that the petitioner's admissions were voluntary. State v. Lapointe, supra, 237 Conn. at 703, 678 A.2d 942 ; see footnote 16 of this opinion. For purposes of evaluating the strength of the state's case against the petitioner, however, the issue is not the voluntariness of the petitioner's admissions—a threshold issue that concerns the admissibility of his statements—but, rather, whether a jury would find those statements trustworthy or reliable because they constitute a true and accurate reflection of the petitioner's involvement in the victim's murder.

Relatively little was known about false confessions at the time of the petitioner's criminal trial. See, e.g., State v. Perea, 322 P.3d 624, 641 (Utah 2013) (observing that, “[i]n the 1990s, little research had been conducted on the phenomenon of false confessions”). Significantly, since then, an abundance of social science research about the phenomenon has been conducted, and, due largely to advances in DNA testing, scores of false confessions have been identified. These developments prompted the United States Supreme Court recently to observe: “By its very nature, custodial police interrogation entails inherently compelling pressures. ... Indeed, the pressure of custodial interrogation is so immense that it can induce a frighteningly high percentage of people to confess to crimes they never committed. Corley v. United States, 556 U.S. 303, 321, 129 S.Ct. 1558, 173 L.Ed.2d 443 (2009)....” (Citations omitted; footnotes added; internal quotation marks omitted.) J.D.B. v. North Carolina, ––– U.S. ––––, 131 S.Ct. 2394, 2401, 180 L.Ed.2d 310 (2011). Moreover, it is now “well established that people with mental illness and mental deficiencies are more prone than others to confess falsely, either because of an inordinate desire to accommodate and agree with authority figures or because they are unable to cope with the psychological intensity of the police interrogation, which frequently includes the use of sophisticated ploys and techniques designed to weaken the suspect's resolve.” State v. Edwards, 299 Conn. 419, 446, 11 A.3d 116 (2011) (Palmer, J., concurring); see also, e.g., B. Garrett, Convicting the Innocent: Where Criminal Prosecutions Go Wrong (2011) p. 38 (noting that of forty DNA exonerations discussed, 76 percent of exonerees who falsely confessed were either juveniles—33 percent—or mentally disabled—43 percent).

One cannot evaluate the strength of the state's evidence against the petitioner in the present case incognizant of the fact that our awareness of the phenomenon of false confessions has increased vastly in the nearly twenty-five years since the petitioner's conviction. One also cannot read the petitioner's statements to the police, particularly in light of the testimony of the officers who elicited them, and not be left with serious concerns about their reliability. First, as we previously indicated, the record reveals that the petitioner suffers from a mental impairment. Although the state argued at his criminal trial that he was not “retarded” in a technical sense, it did not dispute that Dandy–Walker syndrome, a congenital malformation of the skull, had left the petitioner with such serious cognitive deficits that many people who knew him simply assumed that he was “retarded....” According to his school records, the petitioner was eighteen years of age when he finally completed the eighth grade, which was as far as he progressed in school. Among other deficits, Dandy–Walker syndrome probably caused the petitioner to be slow and unsteady on his feet, and he also was severely visually impaired from a young age, and later became hearing impaired.

In an effort to demonstrate the severity of the petitioner's mental impairment, his trial counsel called numerous witnesses to testify at the petitioner's criminal trial and at a hearing on his motion to suppress. Without exception, they described the petitioner as a quirky, gullible and childlike man whose funny appearance and simplemindedness made him the constant butt of jokes. Following his arrest, the petitioner also was examined by a number of psychiatrists and psychologists, all of whom testified at the petitioner's trial and described him in remarkably similar terms: slow-witted; childlike; compliant; genial; guileless; talkative; and very concrete and inflexible in his reasoning. None of them observed any signs of psychosis or mood disorder. All of them seemed to agree, however, that, because of the nature of his cognitive impairment, the petitioner was quite capable of confessing to a murder that he did not commit. Donald R. Grayson, a psychiatrist, wrote in one report that the petitioner “reportedly signed a confession ... in an attempt to extricate himself from a stressful situation and to be able to go to the bathroom and get water. Hearing this explanation from the average person would not have seemed believable to [me]. However, after spending a few hours with [the petitioner], [I] could not help but find his explanation quite believable, in view of his concrete approach to life and his apparent[ly] limited intellectual capacity.”

According to Anne M. Phillips, a clinical psychologist in private practice, the petitioner presents “considerably younger than his chronological age” and has a “[f]ull [s]cale IQ [of] 92,” which puts him in the “lower end of the average range of intelligence....” Phillips concluded that neurological damage from Dandy–Walker syndrome, however, prevented the petitioner from fully accessing his intelligence. According to Phillips, the petitioner's “expressive vocabulary is ... fairly weak, which may reflect an aphasic difficulty related to his historical neurological problems. His limited vocabulary may also contribute to the interpersonal impression of much lower overall intellectual ability which [he] creates. He is not good at expressing his thoughts or feelings and is apt to be awkward and rather childlike in his expression. [The petitioner] also appears limited in receptive language, partially due to acuity problems, and partially to comprehension limitations. As a result, he is inclined to sometimes respond arbitrarily to conversation or questions [that] he does not fully understand. [The petitioner also] tends to be quite concrete and inflexible in his reasoning, understanding situations in a narrow and set way, and having considerable difficulty adopting alternative interpretations of events, or even making sense of unfamiliar events.” At trial, Phillips testified that “interpersonally, [the petitioner] comes across as a funny little guy, kind of odd, and, when he responds to questions in a very literal fashion—[such as] when you ask him who his mother is, [and] he says, ‘Mrs. Lapointe,’ or when you ask him about his childhood, and he says, ‘I was a boy’—there is a sense of ‘H[uh]?’ ” With respect to the statements to the police, although the nine hour interrogation was not recorded, the testimony of the three officers who conducted it provides significant insight into the tactics that were used and the petitioner's state of mind during questioning. Upon his arrival at the Manchester police station, the petitioner initially was questioned by Lombardo. According to Lombardo, as soon as the petitioner entered the interrogation room, Lombardo told the petitioner that there was no question in his mind that the petitioner was responsible for the victim's murder. Lombardo testified that the petitioner did not react to the accusation as Lombardo had expected, that is, with a strong or loud objection; instead, the petitioner just sat there passively, shaking his head and quietly saying, “[n]o....” Lombardo stated that the petitioner immediately asked him “if it would be possible for someone to commit a crime and then not remember it....” According to Lombardo, approximately one hour into the interrogation, the petitioner slumped in his seat, breathed heavily, and said, “I killed her.” The petitioner then asked Lombardo whether it was possible “that he committed the crime, and then ... blacked out and just [did not] remember it.” Although Lombardo told the petitioner that it was possible, he thought the petitioner's question was “just another ploy” to avoid telling the truth. According to Lombardo, after he told the petitioner that a person could commit a crime and not remember it, the petitioner gave the first of three sworn statements confessing to the victim's murder. In it, he stated: “On March 8, 1987, I was responsible for [the victim's] death and it was an accident. My mind went blank.”

After the first statement was signed and notarized, Lombardo allowed the petitioner to go to the bathroom. At this point, Lombardo was aware that he did not have enough evidence to charge the petitioner with any crime. When the petitioner returned to the interrogation room, Lombardo told him that he needed to provide more details about the crime. According to Lombardo, the petitioner stated, “[s]he wouldn't cooperate with me so I killed her,” but then immediately “denied that he had committed the crime and told [Lombardo] that the only reason ... he had given [the earlier] statement was because he [thought Lombardo] wouldn't let him go to the bathroom unless he [did]....” Lombardo then began to put more pressure on the petitioner. For example, he told the petitioner that a witness had seen him walking his dog near the victim's apartment one hour before the fire was reported and that the police had found his fingerprints on the murder weapon, neither of which was true. According to Lombardo, the petitioner “became quiet and kind of slumped down in his chair when confronted ... with that information.” Immediately thereafter, the petitioner gave a second sworn statement, in which he admitted to the following: “On March 8, 1987, I went to visit [the victim] with my wife and son. We left the apartment in the late afternoon and went home. I left my house sometime after that to take the dog for a walk. I was at [the victim's] apartment with the dog. We were both there together and the time was right. I probably made a pass at her and she said no. So I hit her and I strangled her. If the evidence shows that I was there, and that I killed her, then I killed her, but I don't remember being there. I made a pass at [the victim] because she was a nice person and I though[t] that I could get somewhere with her. She was like a grandmother to me, that I never had.” Following this second statement, the petitioner became visibly despondent. According to Lombardo, when he pressed him for more details about the crime, the petitioner stated, “[i]f I tell you everything, then the whole town's going to find out and know that I am a sex fiend.” He also told Lombardo that Martin and their son were the only family that he had and that, if Martin found out that he had killed her grandmother, she would leave him. According to Lombardo, the petitioner stated that “he couldn't bear the thought of losing his family and that, if his family did leave him, he might as well be dead.” Thereafter, Lombardo left the room to speak to Morrissey, who had just returned from interviewing Martin. At this time, Lombardo decided that it would be best for Morrissey to continue the interrogation, which he did.

According to Morrissey, his job was “to get more information” from the petitioner about the murder. Toward that end, he began the interview by informing the petitioner that he had just come from speaking to Martin, that she had been informed of his role in the victim's death, and that Martin wanted him to cooperate with the police. The petitioner again denied any involvement in the murder, stating repeatedly that he must have “black[ed] out” because he simply had no memory of the crime. In an effort to “refresh his memory of what happened,” Morrissey repeatedly led the petitioner through the entire “sequence of events,” a process that, according to Morrissey, took several hours and considerable coaxing. When they were finished, Morrissey prepared a written statement for the petitioner to sign. In that statement, the third and final statement, the petitioner provided a more detailed account of his involvement in the murder: “[O]n Sunday March something I was at [the victim's] apartment with my son ... and my wife [Martin]. We visited from about p.m. and then walked home. After being home [a while] I left to walk the dog. I then walked back up to [the victim's] apartment and she invited me in. We each had a cup of coffee (I think [the victim] had tea) and I sat on the couch. I recall having my matches and my smoking pipe in my jacket pocket.

“After my coffee I went into the bathroom (which is located off the bedroom). When I came out [the victim] was in the bedroom combing her hair. She was wearing a pink house coat type of outer wear with no bra. (I could see her breasts when she bent over). I grabbed her with my hand around her waist area. When I did that she pushed me. I threw her on the bed and took off her underwear because I wanted to have intercourse with her. I got my penis inside her for a few strokes and then pulled out and masturbated. I did [ejaculate] on the bed spread when I was finished. I had already thrown her underwear on the right side of the bed. After the sex she said she was going to tell my wife.... I then went to the kitchen and got a steak knife with a hard plastic brown handle and stabbed [the victim] in the stomach while she was [lying] on the couch. The rest of the incident I do not recall although I admit to having strangled her.” According to Morrissey, the petitioner indicated that he strangled the victim by placing both of his hands around her neck.

The petitioner signed his third statement at approximately 12:30 a.m. on July 5, 1989. At that time, Morrissey and Lombardo asked Brooks, their commanding officer, to speak to the petitioner because the petitioner continued “to vacillate, and they wanted to see whether ... [Brooks] could get him to [stop] doing that.” According to Brooks, the petitioner “would give a statement to [his officers], sign it, and then claim that he was only ... giving back information that was given to him, only saying what he thought [the officers] wanted to hear....” Brooks testified that the petitioner's “recanting behavior had continued right to [the] end of [the interrogation],” and that Lombardo and Morrissey enlisted Brooks to speak to the petitioner because Brooks and the petitioner had a prior relationship, and the officers thought that Brooks could get the petitioner “to perhaps take a stand one way or the other.” When Brooks went into the interrogation room and spoke with the petitioner, the petitioner continued to vacillate and appeared extremely agitated. At approximately 1:20 a.m., the petitioner was instructed to go home, and he did so. The petitioner would later testify that he does not know why he confessed except that he had to go to the bathroom, and because he wanted to go home. The petitioner also testified that Morrissey had told him that, if he did not cooperate with the police, Martin, his wife, could be charged with a crime, and their son would become a ward of the state, accusations that Morrissey has denied. It is difficult to read the officers' account of the petitioner's statements in light of the other trial evidence without experiencing the sinking discomfort that comes with the realization that an injustice may have occurred. Indeed, no fair-minded person who is familiar with the evidence in the present case can read the petitioner's statements and feel confident that they represent a true and accurate account of the victim's murder by the person responsible for her death. The first statement—“I was responsible for [the victim's] death and it was an accident” and “[m]y mind went blank”—is totally devoid of incriminating detail. The second statement—“I probably made a pass at her and she said no,” and, “[i]f the evidence shows that I was there, and that I killed her, then I killed her, but I don't remember being there”—sounds more like the statement of a man who, having just been told by someone he trusts that there is incontrovertible proof that he sexually assaulted and murdered his wife's grandmother, is trying to understand how he could have no memory of such a horrific event. Although the third statement is more specific, most of the details bear little resemblance to the actual crime scene evidence; see footnote 12 of this opinion; and those that do, in particular, the facts that presumably only the killer would know, have the ring of someone confirming information (“I did [ejaculate] on the bed spread when I was finished” and “I had already thrown her underwear on the right side of the bed”) rather than conveying it.

The respondent raises several arguments why there would be no reasonable probability of a different result at a new trial notwithstanding the testimony of Martin and the petitioner's burn time experts. We do not find the respondent's arguments persuasive. For example, the respondent argues that a jury would be unlikely to credit Martin's testimony because she did not inform the police, until she was questioned by Morrissey two years after the murder, that the petitioner had walked their dog on the day of the murder. The record reveals, however, that the petitioner did not become a suspect in the victim's murder until two years after the murder, at which time Martin was interviewed for the first time about the petitioner's whereabouts on the day in question. Indeed, the audio recording of Martin's interview reveals, contrary to the respondent's contentions, that a jury readily could conclude that, until that moment, it had never crossed Martin's mind that her husband had any role in the victim's murder.

The respondent also refers to certain facts that the state relied on at the petitioner's criminal trial, most of which are set forth in this court's decision in State v. Lapointe, supra, 237 Conn. at 696–702, 678 A.2d 942, in which we rejected the petitioner's claims on direct appeal, to support the contention that there is no reasonable probability of a different result even if a jury were to consider the testimony of DeHaan, Kelder and Martin. For example, the respondent argues that, “before any information regarding a possible sexual assault became known to the police or the public, the [petitioner] stated in a conversation with ... a friend of the Lapointe family ... that ‘it was a shame they killed an old lady, but they didn't have to rape her, too.’ ” Id., at 699, 678 A.2d 942. At trial, the petitioner testified that, on the night of the murder, he overheard someone at the hospital discussing the fact that the victim had been sexually assaulted. This is a perfectly plausible explanation as to how he knew about the sexual assault because Marvin, a Manchester volunteer fire marshal, testified that he also had heard hospital personnel speaking openly about the sexual assault. In addition, Elizabeth Martin, the victim's daughter-in-law, testified that, on the morning after the murder, she called the Manchester Police Department and was informed by Brooks that the victim had been sexually assaulted. Elizabeth Martin stated that she immediately conveyed this information to Karen Martin and to her brother with instructions not to tell their father, the victim's son, because she was afraid the news would make him emotionally distraught. Finally, within days of the victim's murder, there was speculation in the newspaper that the victim had been sexually assaulted and that Frederick Merrill, a recently released convict who had committed a similar home invasion and sexual assault of a fifty-five year old woman just three days after the victim's murder and three miles from the victim's apartment, also had sexually assaulted and killed the victim.

The respondent further argues that DeHaan's and Kelder's testimony would not have made a difference at the petitioner's criminal trial because it is duplicative of the testimony of Marvin, who, according to the respondent, “assisted [the state fire marshal] in the investigation, and offered his own burn time estimate for the defense....” Marvin's estimate, the respondent maintains, “differed in no meaningful way” from the estimates provided by DeHaan and Kelder. In reliance on these assertions, the respondent contends that Marvin's testimony, coupled with the petitioner's testimony at trial that he was at home with his family between the hours of 7 and 8 p.m., compels the conclusion that the jury considered and rejected an alibi defense based on the probable burn time of the fire. This claim is wholly without merit. Marvin was not a professional firefighter much less a forensic fire expert; he was a volunteer firefighter who was called by the petitioner to testify that, on the night of the murder, he overheard hospital personnel discussing the fact that the victim had been sexually assaulted. It is true that, when asked on direct examination how long it might have taken for the fire to reach its maximum temperature, Marvin estimated that the fire started about fifteen to twenty minutes prior to reaching its peak temperature at around 8:10 p.m. Marvin qualified his answer, however, by stating that it was “extremely rough” and “real hypothetical,” undoubtedly because he was not a professional fire investigator. Indeed, on cross-examination, Marvin admitted that he really had “no idea when the fire was set” and that his estimate of the fire's maximum temperature was just “a guess....” More important, however, the petitioner did not rely on Marvin's estimate to support an alibi defense based on the fire's burn time, or for any other purpose. In fact, to the extent that the petitioner sought to establish that he could not have murdered the victim because he was at home when the murder occurred, the petitioner never contended that his claim was supported by any burn time evidence. In light of the foregoing, it strains credulity to believe that the jury—entirely on its own, without any argument by the petitioner's trial counsel or guidance from the trial court—would have considered an alibi defense based on Marvin's rough and inexpert testimony concerning the time that it took for the fire to reach its maximum temperature.

The respondent also relies on the testimony of Jean Strimike, the petitioner's former neighbor, who testified that, after the petitioner became president of their four member condominium association, he memorized all of the association's complex rules and bylaws, and was a stickler about enforcing the rules. Strimike stated that she and the petitioner got into a heated dispute over nonconforming flowers that she had planted in front of their building, and, after she refused to remove them, the petitioner stomped them into the ground. The most we can say about this evidence, which very much typifies the kind and quality of evidence that the state presented at the petitioner's criminal trial, is that it reveals that the petitioner is someone who could behave poorly on occasion. It clearly does not give rise to a reasonable inference that the petitioner was capable of cold blooded murder. Of course, the state will be free to seek to introduce this evidence at a new trial and, to the extent it is admissible, to rely on whatever probative value it may have.

IV

CONCLUSION

The petitioner was forty-two years old when he allegedly committed one of the most brutal crimes in our state's history—the rape, torture and murder of a defenseless eighty-eight year old woman, a person who, by all accounts, was like a grandmother to him. Although there is abundant evidence in the record concerning the petitioner's simplemindedness, his peculiarities and his very rigid way of thinking, one searches the record in vain for evidence that he ever was physically violent, that he suffered from a mood disorder, psychosis, drug addiction or anything else that could explain why, after visiting the victim every Sunday for years, he suddenly went back to her apartment on the Sunday in question and brutally murdered her, without his wife noticing either that he had left their house or any change in his demeanor or appearance upon his return. Furthermore, at the petitioner's criminal trial, the state was not required to commit to any particular time frame for the murder, arguing only that it occurred sometime between 5:45 p.m., when the victim was last seen alive by Howard, her daughter, and 8:05 p.m., when she failed to answer Howard's telephone calls. If, however, the original jury were to have heard and credited DeHaan's and Kelder's testimony that the fire was set between 7:30 and 8:05 p.m., and if that jury also were to have heard and credited Martin's testimony that the petitioner was at home with her watching television at that time, there is not just a reasonable probability of a different result, there is a near certainty of one. And, as we have explained, there simply is no reason why the jury reasonably could not have credited that testimony. The petitioner therefore has established, under Strickland, that his first habeas counsel's representation of him was constitutionally deficient due to counsel's failure to pursue a Brady claim founded on the state's suppression of the Ludlow note because that nondisclosure deprived the petitioner of evidence establishing a complete and potentially compelling alibi, thereby gravely undermining the reliability of the verdict against him. Because the record demonstrates convincingly that the petitioner is burdened by an unreliable conviction, he is entitled to a new criminal trial. The judgment of the Appellate Court is affirmed.

In this opinion ROGERS, C.J., and EVELEIGH and McDONALD, Js., concurred.

ROGERS, C.J., concurring.

I agree with and join in the conclusion reached by the majority that the judgment of the Appellate Court ordering that the petitioner, Richard Lapointe, receive a new criminal trial must be affirmed, because the petitioner proved that he was prejudiced by his first habeas counsel's deficient failure to pursue a Brady claim predicated on the state's suppression of a note written by Detective Michael Ludlow of the Manchester Police Department, who was the initial lead investigator in the crimes underlying the petitioner's conviction (Ludlow note). See Lapointe v. Commissioner of Correction, 138 Conn.App. 454, 53 A.3d 257 (2012). More specifically, the petitioner has shown that, had the Ludlow note been disclosed prior to his criminal trial, the probability of a different outcome is sufficient to undermine confidence in the result of that trial.

Although I embrace, for the most part, the thorough and persuasive analysis employed by the majority to reach this conclusion, I write separately to disavow the majority's repeated reliance on law review and social science articles postdating the petitioner's criminal trial by many years as part of its assessment of the strength of the evidence supporting the petitioner's conviction and, by extension, the likelihood that the addition of the burn time evidence would have altered the outcome. Instead, I conclude that the record of the petitioner's criminal trial standing alone, without the embellishment of subsequently produced research and literature that was not in existence and, therefore, could not have been considered by the jury at that trial, adequately demonstrates that the case against the petitioner was not a particularly compelling one.

Because the majority opinion outlines the trial evidence and its weaknesses in detail, a brief summary here will suffice. The physical evidence against the petitioner was practically nonexistent, such that the case rested almost entirely upon his own incriminating statements. Those statements were characterized by vacillation and equivocation, were expressed in an odd, confirming fashion accompanied by repeated denials of recollection and, on multiple points, were inconsistent with the physical evidence. Significantly, the petitioner inaccurately described the victim's clothing, the number of stab wounds and the location in her apartment at which they were inflicted, and the method of her strangulation. It further was undisputed that the petitioner suffered from a congenital malformation of the skull that left him with cognitive, physical and sensory deficits. Numerous acquaintances and relatives testified as to his simplemindedness and gullibility. Several psychiatrists and psychologists examined the petitioner and also testified about his deficits, describing him similarly as did the acquaintances and agreeing that he was capable of confessing falsely to murder. Finally, various witnesses' testimony demonstrated that the knowledge that the victim had been sexually assaulted was not, as the state had asserted, uniquely in the possession of the petitioner, but rather, was widespread.

Instead of relying solely on the foregoing weaknesses in the petitioner's 1992 criminal trial to conclude that confidence in the reliability of his conviction has been undermined, the majority supplements the trial record with secondary materials produced only recently, within the last five or six years. My objection to the use of these materials is twofold. First, as a temporal matter, they were not in existence at the time of the petitioner's criminal trial and, therefore, could not have been presented to, or considered by, the jury that convicted him. Accordingly, they should not be relied upon by this court to evaluate the likelihood of a different result had the Ludlow note been timely disclosed. Second, as a jurisprudential matter, the majority, in discussing the content of extra-record materials in relation to the specifics of this case, treads dangerously close to employing those materials to make findings on appeal in regard to the reliability of the petitioner's confession. This court repeatedly has drawn the distinction between the proper use of extra-record materials such as social science texts or journal articles as “legislative facts ... which help determine the content of law and policy, and adjudicative facts ... concerning the parties and events of a particular case.... Legislative facts may be judicially noticed [on appeal] without affording the parties an opportunity to be heard, but adjudicative facts, at least if central to the case, may not.” (Citation omitted; internal quotation marks omitted.) State v. Edwards, 314 Conn. 465, 478–79, 102 A.3d 52 (2014) ; see also State v. Rizzo, 303 Conn. 71, 97–98 n. 16, 31 A.3d 1094 (2011) ; see also State v. Rizzo, supra, at 184 n. 81, 31 A.3d 1094.

I also want to emphasize certain points with respect to the majority's conclusion in part II B of its opinion that it need not defer to the habeas court's determination regarding the credibility of the petitioner's expert witnesses. Specifically, I would emphasize that this court is concluding that it need not defer to the habeas court's credibility determination only because: (1) the habeas court is not making an ultimate finding as to the credibility of the expert witnesses, but is determining only whether there is a reasonable probability that the jury could credit their testimony, pursuant to the materiality standard of Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), and Brady v. Maryland, 373 U.S. 83, 87, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963) ; and (2) the habeas court did not base its credibility determination on the expert witness' personal demeanor, but on a factor that this court is in as good a position to assess, namely, the substance of the testimony. The majority is not holding, and I would strongly reject any suggestion, that this court may ever second-guess the factual findings of the ultimate finder of fact, even when those findings were based on the cold record. The majority also is not holding that the habeas court need not be provided with an opportunity to assess a witness' credibility on the stand, even when it is addressing a claim pursuant to Strickland or Brady. Rather, it is concluding that, when the habeas court has had that opportunity, and has concluded only that the substance of the testimony undermines its credibility, this court is in as good a position as the habeas court to make that assessment.

Accordingly, I do not share Justice Zarella's fear, expressed in his dissenting opinion, that, as the result of the majority opinion, the habeas court will no longer have any important role to play when addressing Strickland and Brady claims. The habeas court still must be afforded an opportunity to assess credibility based on the personal demeanor of witnesses and this court will continue to defer to such assessments. The majority concludes only that, when a credibility assessment is not based on the witness' demeanor or on other factors that this court is not in a position to evaluate, but is based solely on the substance of the evidence, this court is not required to defer to the assessment. In addition, I believe that the rule that the majority adopts today should apply primarily to the testimony of expert witnesses, for whom demeanor generally is much less important than the witness' qualifications and the soundness of his opinion when a finder of fact is assessing his or her credibility, and should rarely if ever apply to factual testimony. See State Board of Physicians v. Bernstein, 167 Md.App. 714, 760, 894 A.2d 621 (2006) (“[d]emeanor most often is a factor in deciding the credibility of a fact witness who is testifying about a fact that may be true or false, not of an expert who is offering his opinion based on assumed facts”); id., at 760–61, 894 A.2d 621 (Maryland Board of Physicians was not required to defer to administrative law judge's assessments of credibility of expert witnesses when assessments were not based on demeanor). Accordingly, I have no reason to believe that the rule will be invoked with a high degree of frequency.

In all other respects, I agree with the majority opinion and its ultimate holding that the judgment of the Appellate Court should be affirmed.

ZARELLA, J., with whom ESPINOSA, J., joins, dissenting.

Although I disagree with virtually all of the majority's analysis and conclusions, I write, in particular, to express my strong objection to the majority's creation and application of a new exception to our well established standard of review, an exception that allows this court to engage in a de novo review of live expert testimony presented in a habeas proceeding. This new standard is driven by the majority's disagreement with the habeas court's¹ credibility findings, which, if allowed to stand, are fatal to the claims raised by the petitioner, Richard Lapointe. The deferential clear error standard of review that applies to credibility assessments in this context bars the majority from substituting its own judgment for that of the habeas court; and the record, which contains ample support for the habeas court's findings, prevents a finding of clear error. Hopelessly trapped between its unwillingness to accept the habeas court's findings and its inability to overturn them under our clear error standard, the majority summons down its deus ex machina: a singular exception to our clear error standard of review, created just for this case and those cases “indistinguishable” from this one, that empowers the majority to reach its desired ending by retrying in this court the credibility issues settled in the habeas court. This court, however, is not a fact-finding court, and we do not judge the credibility of witnesses because we lack the constitutional authority to do so. By making its own findings about the credibility of expert testimony, the majority exceeds the limits of our jurisdiction and unjustifiably usurps the habeas court's role as fact finder, thereby turning the hearing in that court into little more than an exercise in futility.

The habeas court proceeding is more than just a conduit for logging the evidence, and that court's findings represent more than a suggestion. The habeas court is the sole trier of fact and assessor of credibility, with the exclusive power to determine the credibility of the testimony presented to it. See, e.g., Sanchez v. Commissioner of Correction, 314 Conn. 585, 604, 103 A.3d 954 (2014) ; Gaines v. Commissioner of Correction, 306 Conn. 664, 690, 51 A.3d 948 (2012). We consistently have deferred to its exclusive role in this regard, properly relegating our role to reviewing its fact and credibility findings only to determine whether they find support in the record. See, e.g., Sanchez v. Commissioner of Correction, supra, at 604, 103 A.3d 954.

This court's power to draw conclusions of fact and credibility is proscribed not only by practical considerations—unlike the trier, we do not receive evidence first-hand—but also by the limits of our constitutional function. For more than one century, this court has repeatedly recognized that it simply does not have the jurisdiction under our constitution to resolve disputed questions of fact or to assess the credibility of testimony, irrespective of the nature of the claim or the type of evidence presented to the trier of fact. Styles v. Tyler, 64 Conn. 432, 442, 30 A. 165 (1894). The bifurcation of the Superior Court from the appellate level courts leaves the Superior Court as the final arbiter of fact disputes and limits our role to resolving questions of law. Id., at 444–47, 30 A. 165. Consequently, we lack jurisdiction to resolve disputed questions of fact and credibility, regardless of how this court may attempt to cast the nature of its inquiry. Dexter Yarn Co. v. American Fabrics Co., 102 Conn. 529, 538, 129 A. 527 (1925). Adhering to this principle, this court has repeatedly rebuffed suggestions by litigants—and even some judges of this court—that we may substitute our judgment for that of the trier of fact when the trier of fact has made credibility findings after receiving and weighing the evidence. See, e.g., Skakel v. State, 295 Conn. 447, 487 n. 25, 991 A.2d 414 (2010) ( “[t]his court ... squarely has rejected the proposition that a less deferential standard than abuse of discretion should apply to review of decisions pertaining to evidence [even when such decisions are] not predicated on an assessment of the witness' demeanor”).

By allowing this court to substitute its own judgment for that of the habeas court on questions about witness credibility, the majority has exceeded our jurisdiction. Our law is clear that assessing new witness credibility in the Strickland ² prejudice /Brady ³ materiality context presents a question of fact for the habeas court and that, consistent with our jurisdictional limits, we defer to its findings. See, e.g., Sanchez v. Commissioner of Correction, supra, 314 Conn. at 604, 611, 103 A.3d 954 ; Gaines v. Commissioner of Correction, supra, 306 Conn. at 690–91, 51 A.3d 948. Our constitution and case law simply do not permit this court to usurp the primary and exclusive function of the Superior Court. By granting itself a license to review de novo the live expert testimony presented to the habeas court, the majority has placed itself outside the limits of our authority and installed itself as both the trier and reviewer —a duality of power that this court previously has labeled as “evil....” Styles v. Tyler, supra, 64 Conn. at 449, 30 A. 165.

Compounding the impropriety of its analysis, the majority has adopted a standard allowing it to find facts even though neither party has asked this court to do so. Because the majority has adopted its new standard sua sponte, neither party had any notice or opportunity to brief the propriety of such a standard. Furthermore, neither party—most notably not even the petitioner—has claimed that the habeas court's credibility findings were incorrect, so neither party has briefed the merits of the credibility issues decided by the majority. Nevertheless, the majority has not asked the parties to provide any supplemental briefing, leaving the majority to engage in its new fact-finding role, which it has created sua sponte, without any meaningful notice to or guidance from the parties, in violation of the principles recently set forth in Blumberg Associates Worldwide, Inc. v. Brown & Brown of Connecticut, Inc., 311 Conn. 123, 128, 84 A.3d 840 (2014) (Blumberg ).

Finally, although the majority arrogates to itself the responsibility to make its own findings of fact and determinations of credibility, it ultimately proves unequal to the task. Rather than expressing its own findings based on its review of the record, the majority launches a seemingly endless attack on the habeas court's findings—an unnecessary exercise in light of the majority's self-created fact-finding power. Furthermore, in carrying out its purportedly de novo analysis of the record, the majority utterly ignores testimony from the petitioner's own expert witnesses and other evidence that amply supports the habeas court's findings and casts serious doubt on the credibility of their testimony. In short, although the majority goes out of its way to retry the case, I am constrained merely to review it. Deferring to the habeas court's fact-finding role, my own review of the record convinces me that its findings were not clearly erroneous. In light of its findings that the petitioner's new evidence did not reliably establish a burn time that supported the petitioner's alibi defense, I am persuaded that the petitioner has not met his burden under Strickland and Brady.

My analysis in this opinion will proceed in three parts: first, I address why the majority's new standard for reviewing the testimony of expert witnesses contravenes our law; second, I address why the majority's decision to decide issues that the parties have not raised or briefed violates the principles in Blumberg; and, third, I explain why the record supports the habeas court's findings, which leads to the conclusion that the petitioner has failed to meet his burden under Strickland and Brady.

I

STANDARD OF REVIEW

The principal issue presented in this appeal concerns the petitioner's burden, under the prejudice/materiality element of a Strickland or Brady claim, to prove the credibility of new witness testimony, and the habeas court's role in determining whether a petitioner has met that burden. Assessing credibility of new witness testimony in the Strickland and Brady context requires the habeas court to determine whether there is a reasonable probability that a new jury would credit the testimony, and it presents a question of fact that an appellate tribunal reviews only for clear error. See, e.g., Sanchez v. Commissioner of Correction, supra, 314 Conn. at 604, 611, 103 A.3d 954. Although the habeas court in the present case declined to find the petitioner's new expert testimony credible and denied the petitioner's claim on that basis, the Appellate Court did not acknowledge the habeas court's role in making credibility assessments. See Lapointe v. Commissioner of Correction, 138 Conn.App. 454, 476–77 n. 17, 53 A.3d 257 (2012). Instead, the Appellate Court decided that assessing the credibility of new witnesses should be left entirely to a new jury, and not the habeas court. Id. By determining that a jury, and not a court, should assess credibility, the Appellate Court effectively removed any requirement that the petitioner make a credibility showing as part of his Strickland and Brady claim. As a result, the Appellate Court analyzed the petitioner's claim by hypothesizing what a jury could find, if it credited the new evidence. See id., at 476–78 and nn. 17–20, 53 A.3d 257. The Appellate Court granted the petitioner a new trial on the basis of the results of this speculation. Id., at 478, 480, 53 A.3d 257. The issue presented in this appeal is whether the Appellate Court properly concluded as it did.

The success of the petitioner's Strickland claim—which is based on his prior habeas counsel's failure to raise a Brady claim concerning the state's failure to disclose a certain note prepared by Detective Michael Ludlow of the Manchester Police Department (Ludlow note)—depends on the credit that may be afforded the testimony of the petitioner's expert witnesses. A finding that this expert testimony was credible was necessary if the petitioner was to have any hope of establishing that the state's failure to disclose the Ludlow note prejudiced him by preventing him from presenting a viable alibi defense at trial. To demonstrate that he would have been able to present such a defense, he relied on the testimony of two expert witnesses to establish the time period during which the fire that consumed the apartment of the victim, Bernice Martin, likely started, together with testimony of his former spouse, Karen Martin, to establish that he was at home during that same time period. To prove prejudice, the petitioner needed to show a reasonable probability that a jury would credit this expert testimony and that the new credible testimony, when weighed against the original trial evidence, gives rise to a reasonable probability of a different result. A credible alibi defense, when weighed against the trial evidence, certainly would support findings of prejudice under Strickland and materiality under Brady. If, however, a jury was not reasonably likely to credit the petitioner's new alibi evidence, the petitioner would have no valid evidence of the fire's start time, and his alibi claim would fail. Thus, without a credible alibi defense, the petitioner could not demonstrate that his prior habeas counsel's errors or the state's failure to disclose the Ludlow note caused him any harm.⁴ At the petitioner's habeas trial, the habeas court heard testimony from five fire investigation experts: the two experts who originally investigated the scene for the state shortly after the fire; the two experts called by the petitioner, and an expert called by the respondent, the Commissioner of Correction. The habeas court did not credit the opinions of the petitioner's experts, finding that their testimony could not credibly establish a burn time that was precise enough to support the petitioner's alibi claim. The court explained as follows: “The court ... notes that it finds less credible the attempts to provide more precise burn time estimations. Given the complexity of fires and the many factors or dynamics that impact a fire, a complexity which is patently evident from the testimony of [the petitioner's and the respondent's experts], it is impossible to establish with precision when the fire was set. Thus, the more precise the estimate, the less credible the court finds the opinion. What is clear from all the evidence in the record, the original trial testimony, crime scene photographs, reports, and the expert testimony presented to this court on the fire, is that the precise time the fire was set cannot be determined. At best, a range is established that includes that time period of 6:15 p.m. to [7 ] p.m. (fire could have been burning between 5:45 p.m. and 7:55 p.m.), when [Karen ] Martin cannot account for the petitioner's whereabouts and does not provide an alibi for him. ” (Emphasis added.) The habeas court's finding that the petitioner's burn time evidence was not credible led it to conclude that the petitioner had failed to meet his burden of proving his claims under Strickland and Brady, and the habeas court denied the habeas petition.

The Appellate Court reversed the judgment of the habeas court and, in doing so, analyzed the petitioner's claim without acknowledging the habeas court's role in assessing credibility of new witness testimony in this context. See Lapointe v. Commissioner of Correction, supra, 138 Conn.App. at 476–77 n. 17, 53 A.3d 257. The Appellate Court instead concluded that such an assessment must be left to a new jury, not a court. Id. The Appellate Court explained as follows: “If the Ludlow note had been disclosed to trial counsel ... it would have been the responsibility of the jury and not the court to weigh the credibility of the arson experts. Whether the burn time evidence, which was so critical in buttressing [the petitioner's] alibi defense, raised a reasonable doubt as to the petitioner's guilt would best be a determination left to the jury and not a habeas court.” (Emphasis added.) Id. After so concluding, the Appellate Court analyzed the case by hypothesizing about what a jury “could” find “[i]f” it credited the new testimony. Id., at 476, 477, 53 A.3d 257. For example, the Appellate Court stated: “At the ... habeas trial, the [petitioner's] two experts ... testified that the fire could not have been set any earlier than 7:30 p.m. If that

testimony had been presented at the criminal trial, and credited by the jury, the petitioner's whereabouts at and after 7:30 p.m. would have been critical to his defense.... If the jury credited Karen Martin's testimony, it could have concluded that the petitioner was at home watching television with her and their son when the fire had been set.” (Emphasis added; footnote omitted.) Id., at 476–77, 53 A.3d 257. The Appellate Court concluded by explaining: “With the burn time estimate provided by one of the state's fire marshals, [the petitioner's criminal] trial counsel testified that they would have retained the services of an arson expert and that Karen Martin would have testified as to the petitioner's whereabouts during the critical times of [the] evening [of the victim's murder]. That evidence, if believed by the jury, could have resulted in the jury's finding that it was temporally impossible for the petitioner to have committed the crimes [of] which he was convicted.” (Emphasis added.) Id., at 479, 53 A.3d 257. On the basis of its speculation, the Appellate Court concluded that the petitioner was entitled to a new trial because his evidence, if credited, would support his alibi defense. Id., at 479–80, 53 A.3d 257. Contrary to the majority's assertions, the Appellate Court did not analyze the habeas court's credibility findings; nor did it make its own findings with respect to whether there was a reasonable probability that a jury would credit the new testimony. See generally id., at 468–80, 53 A.3d 257. Rather, it left any credibility assessments to a jury, not a court, effectively relieving the petitioner of any burden to establish the credibility of his new witness testimony as part of his claim based on Strickland and Brady. See id., at 476–77 n. 17, 53 A.3d 257.

The respondent appealed to this court upon our grant of certification; Lapointe v. Commissioner of Correction, 307 Conn. 940, 940–41, 56 A.3d 948 (2012). The respondent claims that the Appellate Court improperly disregarded the habeas court's role in assessing credibility and did not properly consider all of the evidence in the record in considering the ultimate questions of prejudice under Strickland and materiality under Brady. Contrary to the majority's view, neither of the parties has claimed that the Appellate Court performed its own, de novo credibility assessment. Rather, both the respondent and the petitioner agree that the Appellate Court left the credibility issue to a new jury, not a court. Specifically, the respondent argues that, by leaving the credibility assessment of new witness testimony to a jury, the Appellate Court improperly disregarded the habeas court's role as fact finder and arbiter of credibility in habeas proceedings, and improperly reached its decision on the ultimate issues of materiality and prejudice by “speculating on the basis of what ‘could have’ happened ‘if’ certain evidence had been believed.” (Emphasis omitted.) The respondent cites cases holding that mere conjecture or speculation is not sufficient to demonstrate prejudice and argues that, “[i]n saying that jurors could have reached a conclusion if they believed certain testimony, the Appellate Court said no more than that this was possible or conceivable” and thus did not base its materiality decision on “that which was reasonably probable in light of the habeas court's finding regarding credibility and the entirety of the evidence [in the] record,” as the Strickland/Brady standard requires. In response, the petitioner acknowledges the respondent's claim that the Appellate Court improperly decided that the credibility issue was “a jury issue” and argues that it was not up to the habeas court to decide whether to credit his experts, each of whom was qualified to render an opinion.

The majority affirms the judgment of the Appellate Court but takes a much different approach than the Appellate Court did, an approach that neither party has requested and that the Appellate Court did not employ. Rather than analyzing the case by assuming that the jury would credit the evidence, as the Appellate Court did, the majority concludes that an appellate tribunal must make its own findings about the credit that may be given to the petitioner's new expert testimony, a standard of review that has no precedent in our law. Although the majority acknowledges the deference we owe to the fact-finding role of the habeas court, including in the Strickland/Brady context, it decides that no such deference is warranted in the present case because the “highly unusual” circumstances of this case require a “limited exception” to that rule. According to the majority, we defer to fact and credibility findings of the habeas court when those findings are “made on the basis of [the] firsthand observation of [a witness'] conduct, demeanor and attitude.” (Internal quotation marks omitted.) Text accompanying footnote 40 of the majority opinion. The majority then explains, however, that, when a finding is not based on those factors, this court is as well suited as the habeas court to assess the testimony of the petitioner's experts and, therefore, may evaluate that testimony de novo, without any deference to the findings of the habeas court. The majority, of course, fails to cite any authority from this state for this proposition because none exists—this court has never performed such a review, and it has uniformly rejected the notion that it could do so. In resolving disputed issues of credibility by making its own findings on the basis of its own assessment of the testimony and evidence presented to the habeas court, the majority has not just reviewed, but has retried the habeas case in this court. I emphatically disagree with the majority's decision to fashion a new standard of review just for this case.

A

Until today, our case law was clear that determining whether a jury is reasonably likely to credit new witness testimony in the Strickland/Brady context presents a question of fact for the habeas court, not a new jury, and most assuredly not for an appellate tribunal. Our review is limited to determining whether the habeas court's findings are reasonable in light of the record. This issue is not a question of law, and we do not substitute our judgment for that of the habeas court by reviewing the habeas court's credibility findings de novo.

At the outset, I note that I agree with the majority when it states, as it must, that the Strickland prejudice standard is identical to the Brady materiality standard, and that the respective roles of the habeas and reviewing courts are the same under both standards. See, e.g., State v. Dupigney, 295 Conn. 50, 60–63, 988 A.2d 851 (2010). I will therefore analyze the principles that apply to these claims together and, for simplicity, will refer to the standard as the “Strickland/Brady prejudice” standard.

The Strickland/Brady prejudice standard presents a mixed question of law and fact. To prove prejudice, a defendant must establish that, in the absence of the substandard performance of defense counsel or the state's suppression of exculpatory evidence, there is a “ ‘reasonable probability’ ” that the result of the criminal trial would have been different. United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985) ; accord Strickland v. Washington, 466 U.S. 668, 694, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). Resolving the ultimate question of prejudice requires a court to consider or weigh the new or different evidence—which was not presented at the original trial because of counsel's or the state's actions or omissions—against the original trial evidence to determine the likelihood of a different result. See, e.g., Adams v. Commissioner of Correction, 309 Conn. 359, 373, 71 A.3d 512 (2013). We exercise plenary review over the ultimate question of whether the defendant suffered prejudice, which requires an application of the legal standard to the facts. See, e.g., Bryant v. Commissioner of Correction, 290 Conn. 502, 510, 964 A.2d 1186, cert. denied sub nom. Murphy v. Bryant, 558 U.S. 938, 130 S.Ct. 259, 175 L.Ed.2d 242 (2009) ; State v. Ortiz, 280 Conn. 686, 720, 911 A.2d 1055 (2006). When, however, the historical facts supporting the defendant's claim are disputed, the resolution of these underlying disputes presents a question of fact for the habeas court, which we review only for clear error. See, e.g., Bryant v. Commissioner of Correction, supra, at 509, 964 A.2d 1186 ; State v. Ortiz, supra, at 720, 911 A.2d 1055. We defer to the trier's resolution of these underlying disputes unless it is shown to be clearly erroneous. Bryant v. Commissioner of Correction, supra, at 509, 964 A.2d 1186 ; State v. Ortiz, supra, at 720, 911 A.2d 1055.

Although the ultimate question of prejudice focuses on the impact that the new evidence might have on the original trial evidence, determining whether there is a reasonable probability of a different outcome also requires some threshold finding that the new evidence is credible—after all, if new evidence is not worthy of belief, it cannot lead to a different result at a new trial. To be sure, some cases do not present such credibility issues. For example, suppose the state withheld an exculpatory report by a government officer that appears to contradict testimony that the officer gave at trial. If the parties stipulate to the authenticity and content of that report, there are no threshold fact or credibility disputes to be resolved, and the habeas court need only consider the impact of that report on the original trial evidence, an analysis that is subject to de novo review by an appellate court. Frequently in Strickland cases, though less often in Brady cases, the new evidence is testimony from a lay or expert witness who was not called at the original trial. When, as in the present case, the petitioner claims that his counsel or the state deprived him of the opportunity to present certain testimony, the petitioner has the burden of proving that the testimony is credible. Sanchez v. Commissioner of Correction, supra, 314 Conn. at 598–601, 604, 103 A.3d 954. If the testimony is not credible, then it cannot support a finding of prejudice under Strickland or Brady. See id., at 611–13, 103 A.3d 954.

This threshold issue of whether new testimony may be credited in this context—whether lay or expert testimony—presents a question of fact for the habeas court to resolve. The habeas court, as the trier of fact in habeas proceedings, must assess the credibility of the new testimony, considering it together with other evidence and testimony presented to the habeas court, to determine whether there is a reasonable probability that a new jury would credit the new testimony. See, e.g., id., at 604, 611–12, 103 A.3d 954. The habeas court—with its unique ability to receive the new testimony and evidence firsthand and to consider it in the context in which it was presented—is unquestionably in the best position to perform this credibility analysis. After all, the habeas court is the only court hearing the petitioner's claims and the only one that can assess the new evidence in the same manner as a jury would.

When assessing a habeas court's fact and credibility findings, including in the context of Strickland and Brady claims, our role as an appellate tribunal is limited. We do not undertake our own de novo review of the habeas court's findings; rather, we review them under our well established clear error standard. We have often explained: “The habeas court is afforded broad discretion in making its factual findings, and those findings will not be disturbed unless they are clearly erroneous.... Thus, [t]his court does not retry the case or evaluate the credibility of the witnesses.... Rather, we must defer to the [trier of fact's] assessment of the credibility of the witnesses based on its firsthand observation of their conduct, demeanor and attitude.... The habeas judge, as the trier of facts, is the sole arbiter of the credibility of witnesses and the weight to be given to their testimony.” (Internal quotation marks omitted.) Id., at 604, 103 A.3d 954 ; see also Adams v. Commissioner of Correction, supra, 309 Conn. at 368–69 n. 14, 71 A.3d 512 (noting, in context of Brady claim, our “well established standard of review” requiring this court to defer to habeas court's findings unless they are clearly erroneous). Furthermore, our cases also demonstrate that we defer to the habeas court's findings regardless of whether that court credits or rejects the testimony of a petitioner's new witnesses. See, e.g., Gaines v. Commissioner of Correction, supra, 306 Conn. at 688, 690, 51 A.3d 948 (upholding decision to grant new trial when habeas court found new alibi testimony to be “credible and compelling”). When reviewing for clear error, we will reverse the trial court's or habeas court's findings “only in the clearest of circumstances, [when] its conclusion could not reasonably be reached.” (Internal quotation marks omitted.) Ventres v. Goodspeed Airport, LLC, 275 Conn. 105, 125, 881 A.2d 937 (2005), cert. denied, 547 U.S. 1111, 126 S.Ct. 1913, 164 L.Ed.2d 664 (2006). “We do not examine the record to determine whether the trier of fact could have reached a conclusion other than the one reached.” Pandolphe's Auto Parts, Inc. v. Manchester, 181 Conn. 217, 222, 435 A.2d 24 (1980). An appellate tribunal may not reject a finding “merely because the reviewing judges personally disagree with the conclusion or would have found differently had they been sitting as the [fact finder].” Kaplan v. Kaplan, 186 Conn. 387, 391, 441 A.2d 629 (1982).

The majority, and the Chief Justice in her concurring opinion, both suggest that we may treat our review of this credibility assessment differently from a review of credibility findings made at a trial because, in this context, the habeas court is rendering a probabilistic judgment about how a jury might assess certain evidence. This is incorrect.

Although determining how a jury might assess testimony requires some measure of a predictive judgment, we recently—and unanimously—made clear that this assessment is no different from any other credibility assessment and is entitled to the same deference by a reviewing court. In a decision penned by the author of the majority in this case and released on December 2, 2014, we explained that determining whether it is reasonably probable that a jury would credit new witness testimony presents a question of fact for the habeas court that we review only for clear error. Sanchez v. Commissioner of Correction, supra, 314 Conn. at 611–12 and n. 16, 103 A.3d 954 ; see id., at 611, 103 A.3d 954 (upholding denial of habeas petition based on new witness testimony because “perhaps most important, the habeas court concluded that a jury was unlikely to have found [the new witness ] credible, and we agree with the Appellate Court that the habeas court's finding in this regard [was ] not clearly erroneous ” [emphasis added] ). Unlike the majority's decision in the present case, we did not, in Sanchez, draw our own conclusions about the credibility of the new witness testimony. See id., at 602 n. 12, 611–12 and n. 16, 103 A.3d 954. We made clear that the habeas court's role in determining whether a jury might credit new evidence is no different from a traditional credibility analysis and that we will defer to the habeas court's resolution of this issue. See id., at 604, 611–12 and n. 16, 103 A.3d 954 ; see also id., at 602 n. 12, 103 A.3d 954.

Assessing credibility requires the habeas court to approach its analysis neutrally, as a second jury would, without considering the new testimony in a manner favorable to either party, and it must consider all of the evidence presented. See id., at 611 n. 16, 103 A.3d 954 (“we ... presume, in the absence of any indication to the contrary, that the [habeas] court considered all of the evidence when assessing [the new witness'] credibility”). This includes the attitude, candor and demeanor of the witnesses. Id., at 604, 103 A.3d 954. Just as with any other credibility analysis, the habeas court must consider the credibility of the new testimony in light of impeachment evidence and any other testimony or evidence presented to the court.

Contrary to the majority's position, we do not alter our standard of review, even when the petitioner's claims are based on new expert testimony. See Anderson v. Commissioner of Correction, 313 Conn. 360, 375, 380–81, 98 A.3d 23 (2014), cert. denied sub nom. Anderson v. Semple, ––– U.S. ––––, 135 S.Ct. 1453, 191 L.Ed.2d 403 (2015). In Anderson, this court deferred to the habeas court's credibility assessments of new expert testimony even though the habeas court did not expressly base its findings on the experts' demeanor. Id., at 371–72, 377–80 and nn. 7–8, 98 A.3d 23. In doing so, we disagreed with the conclusions of a dissenting Appellate Court judge who had reviewed de novo the credibility of the petitioner's new expert testimony. See id., at 373–74 n. 5, 98 A.3d 23. The petitioner in that case, Oscar Anderson, challenged his conviction of first degree sexual assault and risk of injury to a child. See id., at 362–63, 98 A.3d 23. Anderson claimed that his trial counsel rendered ineffective assistance by not presenting medical records and expert testimony to show that he suffered from one or more sexually transmitted diseases (STDs) during the time period that the assaults occurred, and expert testimony on the transmission rates of STDs. Id., at 362, 98 A.3d 23. Anderson argued that this testimony and evidence, coupled with the victim's lack of any STD during the relevant time period, would establish that he was prejudiced by his trial counsel's failure to present this evidence. See id., at 373, 98 A.3d 23. At the habeas hearing, Anderson presented the testimony of Timothy Grady, an expert witness. Id., at 366–69, 98 A.3d 23. Grady testified that Anderson's medical records showed that he was treated for STDs several times during the period when the assaults allegedly occurred and that, under the circumstances of the case, there was a 40 percent chance of Anderson transmitting an STD to the victim during each act of intercourse. Id.; see also id., at 379 n. 8, 98 A.3d 23. Grady admitted, however, that there were no positive lab results to demonstrate that Anderson was infected with an STD during the relevant period but that Grady's assessments were based on the fact that Anderson had been treated in a hospital's emergency department for STD related symptoms. See id., at 366–67, 98 A.3d 23.

The respondent in Anderson also presented expert testimony on this subject. Id., at 369–70, 98 A.3d 23. The respondent called Stephen Scholand, a physician, who testified that other ailments that Anderson suffered from could be responsible for his STD related symptoms. Id., at 369, 98 A.3d 23. Scholand also testified that, even if Anderson had an STD, that there was only a 30 percent chance of transmitting it from adult to adult during each act of intercourse. Id.

The habeas court found that Anderson's expert testimony lacked a sufficient factual foundation to establish that he suffered from any STD during the relevant time period. See id., at 371–72, 98 A.3d 23. That court determined that, without any test results to support Anderson's claims of an STD when the assaults occurred, his expert evidence could not credibly prove that he suffered from an STD at that time. See id. The habeas court also concluded that, even if Anderson had demonstrated that he had an STD during the relevant time period, there was only a 30 percent chance of transmission, implicitly rejecting the opinion of Anderson's expert that the rate would have been higher under the circumstances. See id., at 372, 98 A.3d 23. On the basis of these findings, the habeas court rejected Anderson's claims. See id., at 372–73, 98 A.3d 23.

On appeal, a divided Appellate Court affirmed. Anderson v. Commissioner of Correction, 128 Conn.App. 585, 598, 17 A.3d 1138 (2011). The Appellate Court majority, relying on the habeas court's findings, upheld that court's decision that Anderson's evidence fell short. See id., at 597–98 and n. 7, 17 A.3d 1138. Judge Borden, in dissent, thought, however, that Anderson had met his burden under Strickland. See id., at 598–99, 17 A.3d 1138 (Borden, J., dissenting). In analyzing the prejudice element, Judge Borden relied on Anderson's proffered expert testimony regarding his STD related symptoms and the transmission rate of STDs, which the habeas court had found to be not credible. See id., at 616–17, 17 A.3d 1138 (Borden, J., dissenting). Specifically, Judge Borden asserted that Anderson “was suffering from” STDs during the relevant time period. Id., at 610, 17 A.3d 1138 (Borden, J., dissenting). He also explained that “the jury would have had Grady's testimony, contrary to Scholand's, that there is a transmission rate of 40 to 50 percent for each sexual contact between an infected male and a woman, and an even higher rate for a female of the age of the victim,” even though the habeas court did not credit this evidence but, rather, concluded that the evidence showed only a 30 percent chance of transmission. Id., at 616–17, 17 A.3d 1138 (Borden, J., dissenting). Rather than confining his analysis to the habeas court's findings, Judge Borden disregarded the habeas court's findings and relied on testimony that the habeas court had not credited.

In a certified appeal to this court, Anderson argued that the Appellate Court majority improperly had decided the case against him. See Anderson v. Commissioner of Correction, supra, 313 Conn. at 362, 98 A.3d 23. In his arguments, Anderson, consistent with Judge Borden's reasoning, repeatedly relied on the evidence that the habeas court had found to be not credible, namely, that he was infected with one or more STDs and that the transmission rate would have been 40 percent or higher. See Anderson v. Commissioner of Correction, Conn. Supreme Court Records & Briefs, February Term, 2013, Petitioner's Brief pp. 22–25.

In response, the respondent took issue with Anderson's and Judge Borden's reliance on testimony that the habeas court had not credited. The respondent argued that “[Judge Borden] improperly assumed that it is reasonably probable that the jury would have found Grady's testimony and interpretations of the medical records to be credible and reliable when the fact finder ... did not.” Id., Respondent's Brief p. 19. The respondent explained in his brief that “[a]n appellate court cannot retry the case or evaluate the credibility of the witnesses.... Rather, [it] must defer to the [trier of fact's] assessment of the credibility of the witnesses based on its firsthand observation of their conduct, demeanor and attitude.... The habeas [court], as the trier of facts, is the sole arbiter of the credibility of witnesses and the weight to be given to their testimony.” (Internal quotation marks omitted.) Id. The respondent argued that Judge Borden's dissent contravened this well established standard of review because Judge Borden “wholeheartedly credited, and essentially deferred to, the opinions and interpretations of [Anderson's] expert,” which the habeas court did not credit. Id., pp. 19–20. The respondent further explained that “[a] reviewing court cannot simply choose to find credible particular evidence that the trier of fact itself did not credit and [that] is inconsistent with the trier's ultimate findings and conclusions.” Id., p. 20. In support, the respondent noted that determining prejudice on the basis of new witness testimony is “not simply ... an evaluation of the legal sufficiency of that evidence” but also requires an evaluation of “the overall credibility and persuasiveness of that evidence....” Id. Consistent with our case law, the respondent went on to explain that this credibility assessment is for the habeas court to make, not a reviewing court: “[O]nly the habeas court heard and considered all of the evidence.... Unlike a reviewing court, the habeas court is able to view the evidence as it is presented, including the demeanor of the witnesses, and, thus, the habeas court is in a far better position to appreciate how a jury is likely to hear and process that evidence....” (Citation omitted; emphasis in original; internal quotation marks omitted.) Id. The respondent acknowledged that a petitioner's evidentiary burden is lower in this context than in an ordinary criminal trial, at which proof beyond a reasonable doubt is required, or in a civil trial, at which proof by a preponderance of the evidence is required. See id., pp. 20–21. Nevertheless, the respondent explained that “the fact that the degree of proof is different does not mean that the proper roles of the habeas court and appellate court vis-á-vis the credibility, weight and ultimate persuasiveness of the evidence differ” in this context because “[the] habeas court is still in [a] position superior to an appellate court to make such an evaluation....” (Emphasis in original.) Id., p. 21. In conclusion, the respondent argued that “the habeas court, sitting in a position similar to that of the jury hearing the evidence, refused to accept, at face value, the opinion of [Anderson's] expert and his interpretation of the records.... [T]he habeas court was in a far better position than [Judge Borden] to view the demeanor of the experts and the confidence and certainty with which they rendered their opinions. Consequently, to the extent that [Judge Borden's] conclusions rely heavily—indeed, almost entirely—on [his] deference to Grady's testimony, [Judge Borden] fails to give proper consideration to the limits of appellate review in evaluating the reasonable probability that such evidence would have persuaded the jury to [find Anderson not guilty].” Id.

This court affirmed the judgment of the Appellate Court. Anderson v. Commissioner of Correction, supra, 313 Conn. at 384, 98 A.3d 23. We did not adopt the approach that Judge Borden had taken and we disagreed with his conclusions. See id., at 378 n. 7, 380 n. 9, 98 A.3d 23. In our decision, we set forth the habeas court's findings; id., at 371–72, 98 A.3d 23 ; and cited our well established standard of review that “[t]he habeas [court], as the trier of facts, is the sole arbiter of the credibility of witnesses and the weight to be given to their testimony.” (Internal quotation marks omitted.) Id., at 375, 98 A.3d 23. In our analysis, we upheld the habeas court's finding that Anderson's evidence failed to establish that he suffered from an STD during the time period that the assaults occurred. Id., at 377–78 and n. 7, 98 A.3d 23. We expressly disagreed with Judge Borden's conclusion to the contrary, and, in response to his arguments, we cited the habeas court's findings that contradicted his conclusion. Id., at 378 n. 7, 98 A.3d 23. Furthermore, we also decided that, “[e]ven if [it were] assume[d] that [Anderson] was suffering from nonspecific [STDs] during the relevant time frame ... Scholand's testimony established that there was only a 30 percent chance that the victim would have acquired [an STD] from the alleged abuse.” (Footnote omitted.) Id., at 378–79, 98 A.3d 23. In reaching this conclusion, we expressly deferred to the habeas court's findings in this regard, noting that the habeas court had the opportunity to view the testimony of the experts firsthand. Id., at 379 n. 8, 98 A.3d 23. Thus, contrary to the approach taken by Judge Borden, and consistent with the respondent's arguments, this court did not substitute its judgment for that of the habeas court regarding the credibility of Anderson's proffered expert testimony. The majority's decision to adopt an exception to our well established standard of review, especially in light of Anderson and Sanchez, marks a significant departure from the approach that we have taken in reviewing Strickland and Brady claims and from our role as an appellate tribunal.

B

Even though the majority acknowledges the deference that we give to a habeas court's credibility assessments, stating that it “agree[s] fully” that “the general rule is one of deference, even in cases involving claims under Brady, ” the majority has decided not to follow this rule in the present case. The majority proclaims, without citing a single Connecticut case, that we can undertake a de novo review of the credibility of expert testimony “when the habeas court's assessment of the expert testimony has nothing to do with the personal credibility of the expert witness but instead is based entirely on the court's evaluation of the foundational soundness of the witness' professional opinion....” This novel standard, which neither party has asked us to adopt and which even the majority concedes has no basis in our law, is wholly incompatible with our case law and our limited role as an appellate tribunal.

1

The majority contends that we may render our own credibility findings when the habeas court's findings were based on the “scientific underpinnings” of the expert's testimony and not the demeanor or “personal credibility” of the expert witness. Our case law does not draw such an unworkable distinction because it ignores the realities of judging credibility. When we speak of credibility, we do not refer only to a witness' personal character but to the broader question of whether the witness' testimony may be believed and given evidentiary weight. Evaluating the credibility of a witness' oral testimony is a complex process that always entails consideration of subjective factors like the attitude, candor and demeanor of the witness, and this assessment cannot be based solely on objective factors reflected in the printed record. E.g., Wyatt Energy, Inc. v. Motiva Enterprises, LLC, 308 Conn. 719, 737, 66 A.3d 848 (2013) (“[c]redibility must be assessed ... not by reading the cold printed record, but by observing firsthand the witness' conduct, demeanor and attitude” [internal quotation marks omitted] ). Separating the “substance” of oral testimony from the manner in which it is communicated ignores the fact that the meaning and import of spoken words depend on the manner in which the spoken words are relayed, and not just their substance. This is precisely why, when it comes to live testimony, “it is inappropriate to assess credibility without having watched a witness testify, because demeanor, conduct and other factors are not fully reflected in the cold, printed record.” (Internal quotation marks omitted.) Shelton v. Statewide Grievance Committee, 277 Conn. 99, 111, 890 A.2d 104 (2006). In assessing new witness credibility in the Strickland /Brady prejudice context, our law thus requires that the trial court observe the testimony of the new witness firsthand to consider the witness' demeanor while he or she testifies. See, e.g., Thomas v. Commissioner of Correction, 141 Conn.App. 465, 472–73, 62 A.3d 534 (habeas court must have opportunity to observe testimony of petitioner's new witnesses to evaluate their credibility in determining Strickland/Brady prejudice), cert. denied, 308 Conn. 939, 66 A.3d 881 (2013) ; Townsend v. Commissioner of Correction, 116 Conn.App. 663, 668, 975 A.2d 1282 (upholding denial of habeas petition when petitioner did not present testimony of new exculpatory witness, “which the habeas court properly found was insufficient to show prejudice because there was no opportunity to evaluate the testimony or credibility of the claimed witness”), cert. denied, 293 Conn. 930, 980 A.2d 916 (2009).

Judging credibility is a complex process that requires the trier to consider the testimony in light of all of the evidence presented in a case, including other testimony. It requires consideration of not only the witness' candor, demeanor and attitude, but also any factors that the trier deems relevant to its decision, which can include, but is certainly not limited to: the witness' ability to observe and comprehend information; the beliefs, prejudices, bias and assumptions that might impact the witness' perspectives; the witness' ability to accurately recall events; the plausibility of what the witness relays; the internal consistency of the witness' testimony and consistency with other testimony and evidence; and the witness' background, training, education and experience. See, e.g., State v. Reddick, 153 Conn.App. 69, 79 n. 6, 100 A.3d 439, appeal dismissed, 314 Conn. 934, 102 A.3d 85 (2014), and cert. denied, 315 Conn. 904, 104 A.3d 757 (2014) ; State v. Guess, 44 Conn.App. 790, 805, 692 A.2d 849 (1997), aff'd, 244 Conn. 761, 715 A.2d 643 (1998) ; State v. Nieves, 36 Conn.App. 546, 549, 653 A.2d 197, cert. denied, 232 Conn. 916, 655 A.2d 260 (1995) ; State v. Jaynes, 35 Conn.App. 541, 553–54 n. 7, 645 A.2d 1060, cert. denied, 231 Conn. 928, 648 A.2d 880 (1994).

Judging credibility is no different with expert witnesses, and attempting to distinguish between decisions based on the “substance” of expert testimony and those based on the expert's demeanor is just as unworkable. Because of the nature and purpose of expert testimony, subjective factors, including attitude, candor and demeanor, play just as great of a role in the evaluation of an expert witness' credibility. Our law allows expert opinion testimony to be admitted only when its subject matter is beyond the ordinary knowledge, comprehension and experience of the average fact finder. See, e.g., State v. George, 194 Conn. 361, 373, 481 A.2d 1068 (1984), cert. denied, 469 U.S. 1191, 105 S.Ct. 963, 83 L.Ed.2d 968 (1985). Thus, when considering expert testimony, the jury is evaluating testimony about subject matter that is outside of its knowledge, which puts the jury at a disadvantage when it comes to sizing up the soundness of the expert's scientific foundation and conclusions. When faced with conflicting expert testimony, the jury necessarily lacks firsthand knowledge with which to assess the soundness of the expert's methodology and opinions. Consequently, in deciding whether to credit an expert's testimony, a fact finder often must rely more heavily on subjective factors. As a result, our law recognizes the importance of demeanor evidence as a tool for evaluating an expert witness' credibility. “Expert testimony is considered, weighed and tested like any other evidence.” Aspiazu v. Orgera, 205 Conn. 623, 634, 535 A.2d 338 (1987) ; see also State v. Joly, 219 Conn. 234, 243, 593 A.2d 96 (1991) ( “the fact that a witness testifies as an expert does not compel the acceptance of his or her testimony as true”). “It is in the sole province of the trier of fact to evaluate expert testimony, to assess its credibility, and to assign it a proper weight.” State v. Jarzbek, 204 Conn. 683, 706, 529 A.2d 1245 (1987), cert. denied, 484 U.S. 1061, 108 S.Ct. 1017, 98 L.Ed.2d 982 (1988). In assessing expert testimony, the trier considers subjective factors, such as the demeanor of the expert, as well as other factors, including the expert's level of expertise, the reasonableness of how the expert reached the opinion, and the factual and scientific basis for the opinion: “The trier of fact [is] ... entitled to consider the basis of the testimony of the expert witness. In so doing it might weigh, as it [sees] fit, his expertise, his opportunity to observe [data] and to form an opinion, and his thoroughness. It might consider also the reasonableness of his judgments about the underlying facts and of the conclusions [that] he drew from them.” State v. Perez, 182 Conn. 603, 610, 438 A.2d 1149 (1981) ; see also id., at 609, 438 A.2d 1149 (“[t]he trier of fact was entitled to find that the cross-examination of [the expert witness] showed his testimony to be inadequately grounded in knowledge of the defendant's condition, medical data, and expertise of interpretation”). “The acceptance or rejection of the opinions of expert witnesses is a matter peculiarly within the province of the trier of fact and its determinations will be accorded great deference by this court.” (Internal quotation marks omitted.) Johnson v. Healy, 183 Conn. 514, 515–16, 440 A.2d 765 (1981). “The credibility and the weight of expert testimony is judged by the same standard [as that used to evaluate lay witness testimony], and the trial court is privileged to adopt whatever testimony [it] reasonably believes to be credible.... On appeal, [this court does] not retry the facts or pass on the credibility of witnesses.” (Internal quotation marks omitted.) United Technologies Corp. v. East Windsor, 262 Conn. 11, 26, 807 A.2d 955 (2002) ; see also Connecticut Bank & Trust Co. v. Incendy, 207 Conn. 15, 34, 540 A.2d 32 (1988).

The scope of our review on appeal does not change according to our perception of the basis for the trier's credibility findings. Because the trier must consider all of the evidence before it when making its credibility findings, including the witness' demeanor, it is impractical, if not impossible, for the trier—or this court—to neatly categorize or separate credibility findings on the basis of a witness' demeanor from those derived from the basis for the witness' opinions. Given the complexity of a credibility analysis, the innumerable factors that a trier considers, and the complete discretion afforded the trier to weigh these factors, it would be unreasonable to require the trier to enumerate each and every basis for its finding by describing every factor that it considered and the weight it accorded to it. For these reasons, we do not require the trial court to state every conceivable rationale for its credibility findings; we require only that the trial court state its ultimate finding of fact on the record. See Practice Book § 64–1. Even if a trial court chooses to enumerate certain reasons for its credibility findings, we still do not vary our scope of our review on the basis of those reasons because it is inevitable that the court considered other evidence not expressly identified in its decision. Rather, we presume that the trier considered all of the evidence in making its findings, and we review them only for clear error. See Sanchez v. Commissioner of Correction, supra, 314 Conn. at 611 n. 16, 103 A.3d 954 ; Gaines v. Commissioner of Correction, supra, 306 Conn. at 690–91, 51 A.3d 948. As a result, our standard of review does not change even when the trial or habeas court does not cite witness demeanor as a reason for its credibility findings. See Sanchez v. Commissioner of Correction, supra, at 611–12, 103 A.3d 954 ; Anderson v. Commissioner of Correction, supra, 313 Conn. at 372, 377–80 and nn. 7–8, 98 A.3d 23.

Our case law belies the majority's claim to the contrary. For example, in Sanchez, we upheld the denial of a habeas petition when the habeas court found that the new witness testimony was not likely to be credited by a jury. See Sanchez v. Commissioner of Correction, supra, 314 Conn. at 611–12. The habeas court, in making its credibility finding, explained only that the witness had a criminal record and “a motive to be deceptive....” (Internal quotation marks omitted.) Id., at 601, 103 A.3d 954. Neither of these factors related to the witness' demeanor. The witness' criminal history was an objective factor that was reflected in the record. With respect to the witness' “motive to be deceptive”; (internal quotation marks omitted) id.; we explained that this stemmed from the fact that others had implicated the witness in the crime in question, giving the witness a reason to minimize his own involvement when testifying before the habeas court. See id., at 611–12, 103 A.3d 954. This, too, was reflected in the printed record. See id. The habeas court said nothing about the witness' demeanor, nor did it state that it relied on any other subjective factors in deciding not to credit the testimony. See id. Thus, we presumably were in just as good a position as the habeas court in that case to assess the witness' credibility. Nevertheless, unlike the majority's opinion in the present case, we deferred to the trial court's assessment, as our law requires, and noted that we presume that the court considered all of the evidence in making its finding, not just its enumerated reasons. Id., at 611 n. 16, 103 A.3d 954. We did not conduct a de novo review of the witness' credibility, and we did not suggest that our standard of review could change depending on the trial court's stated basis for a credibility finding.

Even in cases involving new expert testimony in the Strickland/ Brady prejudice context, this court and the Appellate Court apply a deferential standard of review without pausing to consider whether the habeas court's findings were based on the credibility of the expert's opinion or the expert personally. See, e.g., Anderson v. Commissioner of Correction, supra, 313 Conn. at 372, 377–80 and nn. 7–8, 98 A.3d 23 (deferring, in Strickland prejudice analysis, to habeas court's decision not to credit petitioner's new expert testimony about disease transmission rates); Francis v. Commissioner of Correction, 142 Conn.App. 530, 540 n. 5, 66 A.3d 501 (upholding denial of habeas petition and deferring, in context of Strickland prejudice determination based on new expert testimony, to habeas court's finding that that testimony was unreliable), cert. denied, 310 Conn. 921, 77 A.3d 141 (2013). The majority has not cited a single case in which we changed our standard of review because of the stated basis for the trier's credibility findings, and it cannot do so because we have never made such an unworkable distinction. Moreover, because we cannot review live testimony firsthand, as the trier can, we are unable to consider factors such as attitude, candor and demeanor. Kaplan v. Kaplan, supra, 186 Conn. at 391, 441 A.2d 629 ; see Sanchez v. Commissioner of Correction, supra, 314 Conn. at 611–12, 103 A.3d 954. Thus, when it comes to credibility issues, we are in a worse position than the habeas court, and, by making our own findings, we would be substituting our judgment for that of the habeas court, even though our judgment is based on only part of the evidence. The truth is, we can never be in the same position as the trier with respect to oral testimony and can never perform a truly de novo review on the basis of all of the evidence, as we require the trier to do. Our review is therefore limited to testing the legal sufficiency of those findings by assessing whether they are reasonable in light of the record. See, e.g., Sanchez v. Commissioner of Correction, supra, 611–12 and n. 16, 103 A.3d 954 ; Anderson v. Commissioner of Correction, supra, 313 Conn. at 375, 377–80 and nn. 7–8 ; see also Ventres v. Goodspeed Airport, LLC, supra, 275 Conn. at 125, 881 A.2d 937.

The facts of this case also contradict the majority's attempted distinction. The majority claims that the habeas court's credibility assessments were based solely on the “soundness” of the experts' opinions and not on the demeanor or personal credibility of any witness. This simply is not true. Because we do not require our trial or habeas courts to articulate each and every factor considered and the basis for their credibility findings, we do not know the extent to which the habeas court in the present case considered demeanor in making its findings; the habeas court was silent on this. Because we must presume that the habeas court considered all of the evidence before it, including the demeanor of the witnesses; see Sanchez v. Commissioner of Correction, supra, 314 Conn. at 611–12 n. 16, 103 A.3d 954 ; we must presume that the habeas court based its findings, at least in part, on the demeanor and candor of the petitioner's witnesses and the demeanor and testimony of any other witnesses who testified before the court. To be sure, it is possible that the habeas court found that an expert's demeanor weighed in favor of crediting the expert's testimony. But, given the habeas court's silence, it is not possible for us to say either that the habeas court did not consider witness candor, attitude and demeanor, or that it did not base its findings on these factors. Indeed, if it were true that the habeas court did not consider the witnesses' demeanor and personal character in considering the credibility of their testimony, then the habeas court would have been acting improperly. We, of course, do not presume error; we presume that the habeas court considered these factors in the absence of a clear indication to the contrary, which the habeas court did not give in this case.

Assessing the credibility of the petitioner's expert witnesses also required the habeas court to consider the credibility of other witnesses who testified at the habeas hearing. The majority claims that the habeas court's assessment was “not dependent on any underlying factual findings requiring the ... court's firsthand observation and determination of the credibility or reliability of other witnesses.” Yet, the habeas court heard evidence that the conclusions of one of the petitioner's experts, John DeHaan, directly conflicted with the eyewitness observations of Stephen Igoe, another fire expert who, unlike the petitioner's experts, investigated the actual fire scene in person and within hours of the fire. The habeas court noted that, as part of Igoe's investigation, he conducted a test burn of the materials in the couch where the fire was supposedly ignited and principally burned. He determined that the couch materials burned “very, very slowly....” Despite Igoe's testimony before the habeas court about his firsthand observations of the materials involved in the fire, the petitioner's expert, DeHaan, who was not present at the fire scene, disagreed with Igoe's observations. Although DeHaan did not inspect the actual fire scene or the materials in the couch, and did not conduct a test burn, he based his burn time estimate on a contrary conclusion that the couch burned “very rapidly.” Assessing the soundness of DeHaan's opinion that the fire burned very rapidly therefore required the habeas court to evaluate the credibility of Igoe's contrary observation that the fire burned very slowly. Thus, the habeas court's findings necessarily depended on “underlying factual findings requiring the ... court's firsthand observation and determination of the credibility or reliability of other witnesses.”

Because the trier must consider all evidence when making credibility findings, it is not possible to separate those findings based on the soundness of an expert's opinion from those based on an assessment of the expert personally. The majority thus attempts to draw a dichotomy that is not grounded in reality and is not supported by our law or the facts of this case. If the majority nevertheless wants to draw this contrived distinction, then the appropriate remedy in this circumstance is to order an articulation from the habeas court rather than to make assumptions based on the habeas court's silence, which tells us nothing when we have never previously required the habeas court to speak on this matter or given any significance to this silence.

2

Even if it were possible to distinguish between credibility findings based on an expert's demeanor and those based on the substance of the testimony, our law still forbids a de novo review of such credibility findings by this court. In support of its new standard, the majority proclaims that we “need not, and will not,” defer to the trial court on any issue “when we are in as good a position as the trial court to decide the issue....” Text accompanying footnote 63 of the majority opinion. But this court has never before embraced such a sweeping principle, and we certainly have not accepted it as a valid excuse to retry a case by making our own findings from the record. To the contrary, this court has repeatedly rejected this tired suggestion because it is inconsistent with the constitutional limits on our role as an appellate tribunal. See, e.g., Styles v. Tyler, supra, 64 Conn. at 442, 30 A. 165. For example, we have determined that we cannot make our own credibility assessments, even when the trial court's assessment was based on a video recording of a witness' statement that we could view in the same manner as the trial court. See Skakel v. State, supra, 295 Conn. at 470, 486–87 and n. 25, 991 A.2d 414. In Skakel, the author of the majority opinion in the present case, in dissent, posited that there was “no occasion for ... deference” to the trial court's credibility determination when it was based on the viewing of a video recording and thus not otherwise based on subjective factors, including the witness' demeanor. Id., at 631, 991 A.2d 414 (Palmer, J., dissenting). The majority rejected this assertion, noting that “this court ... squarely has rejected” the notion that we may apply a less deferential standard of review even to a credibility decision “that is not predicated on an assessment of the witness' demeanor.” Id., at 487 n. 25, 991 A.2d 414.

We have also uniformly rejected this notion in other contexts. For instance, in State v. Lawrence, 282 Conn. 141, 156–57, 920 A.2d 236 (2007), we declined to undertake a de novo review of a trial court's credibility determination that was based on that court's review of a printed transcript, rather than an in-court assessment of the witness' testimony. Although the defendant in Lawrence claimed that deference to the trial court's findings was unwarranted “because the trial court did not have the opportunity to observe [the witness'] conduct, demeanor and attitude on the witness stand”; id., at 156, 920 A.2d 236 ; we declined to adopt this approach as it “misapprehends the fundamental distinction between the function of the fact finder, which is to make credibility determinations and to find facts, and the function of the appellate tribunal, which is to review, and not to retry, the proceedings of the trial court.” (Internal quotation marks omitted.) Id. We explained that, “[i]n light of our limited function, it would be improper for this court to supplant its credibility determinations for those of the fact finder, regardless of whether the fact finder relied on the cold printed record to make those determinations. ” (Emphasis added.) Id., at 157, 920 A.2d 236. In Besade v. Interstate Security Services, 212 Conn. 441, 562 A.2d 1086 (1989), we rejected a claim that this court need not defer to a trial court's findings, even when the evidence before the court was “largely documentary” and the findings were not based on “a personal appraisal” of any live testimony. Id., at 448, 562 A.2d 1086 ; see id., at 449, 562 A.2d 1086 (“[w]e have not heretofore distinguished between documentary and testimonial evidence in defining the role of appellate tribunals in reviewing findings of fact”).

Even in the context of expert testimony, this court has consistently declined to adopt a less deferential standard of review when the trial court's findings were based solely on the substance of the expert testimony and documentary evidence, not on witness demeanor. See, e.g., Pandolphe's Auto Parts, Inc. v. Manchester, supra, 181 Conn. at 221–22, 435 A.2d 24 (noting that our review of fact and credibility findings is limited to clear error review, and, “[b]eyond that, we will not go”). And over one-half century ago, we explained in Morgan v. Hill, 139 Conn. 159, 90 A.2d 641 (1952), that “[t]he case [on appeal] presents another of countless instances [in which] an unsuccessful litigant, still unconvinced, renews in this court his previous, fruitless effort to discredit the evidence submitted by his opponent. Nothing in our law is more elementary than that the trier is the final judge of the credibility of witnesses and of the weight to be accorded their testimony.... It was the prerogative of the [trial ] referee to accept the testimony of the plaintiff's experts rather than that offered by the defendant's. ” (Citations omitted; emphasis added.) Id., at 161, 90 A.2d 641.

3

We defer to the trial court's findings even when demeanor is not at issue because the reasons for our deference are not limited to practical considerations alone. Even though we often cite, as the reason for our deference, a trial court's unique ability to assess evidence firsthand and in the context in which it is presented, our scope of review is mandated by the constitutional limits on our appellate jurisdiction, which prohibit this court from resolving disputed questions of fact or determining the credit that may be given to a witness' testimony. Since this court's inception, its jurisdiction has been limited to reviewing issues of law, whereas jurisdiction to resolve issues of fact and credibility has been vested exclusively in the trial court. See, e.g., Styles v. Tyler, supra, 64 Conn. at 444–50, 30 A. 165. This division of responsibility is not merely for convenience; it is intended to prevent the “evil” that would result from having a single court act as final arbiter of issues of fact in addition to issues of law. Id., at 446, 449, 30 A. 165. Prior to the creation of this court, the Superior Court acted as the court of last resort for issues of fact, whereas the legislature held the power to finally resolve questions of law. Id., at 446, 30 A. 165. The legislature created the Supreme Court of Errors in 1784; id.; and its “jurisdiction was confined to questions of law arising upon facts found by the Superior Court [which is] the court of last resort for all matters of fact.” Id., at 447, 30 A. 165. This separation intentionally paralleled the “distinction as drawn under our system of jurisprudence ... between facts that the trial court must find from the testimony, and the application of the principles of law in reaching a judgment based [on] such facts.” Id., at 454, 30 A. 165.

The founders of our state's judicial system divided the jurisdiction of our courts to ensure that no single court could act as the final arbiter for both issues of fact and issues of law, which they feared could destroy the rule of law: “[T]he administration of justice is not safe when the court of last resort for the settlement of the law, in the exercise of an absolute and final power, can render judgment on the facts and law so intermingled that its decision is not simply the declaration of the law but may become the arbitration of the case.” Id., at 449, 30 A. 165. Vesting one court with jurisdiction to finally decide matters of fact and law will leave the results in our cases to the discretion of judges who can choose both the facts and the law that will govern a case, and “nothing but ... human wisdom and firmness on the part of its judges can prevent a court exercising such ... jurisdiction from eventually becoming one great arbitration that would [e]ngulf all the courts of law, and sovereign discretion [not the principles of law] would be the rule of decision.” (Internal quotation marks omitted.) Id. Such an expansive “discretion of a judge is the law of tyrants; it is always unknown; it is different in different men; it is casual, and depends [on] constitution, temper and passion. In the best, it is [often times] caprice; and in the worst, it is every vice, folly, and passion to which human nature is liable!” (Internal quotation marks omitted.) State v. Danforth, 3 Conn. 112, 122 (1819). Thus, “confining the jurisdiction of [this] court ... to the settlement of rules of law” alone was “[t]he most significant feature in the establishment of the court”; Styles v. Tyler, supra, 64 Conn. at 447, 30 A. 165 ; and reflected “a principle deemed vital to our judicial system.” Id., at 448, 30 A. 165.

Mindful of these concerns, the drafters of the 1818 constitution adopted the existing divide between the jurisdiction of the Superior Court and the Supreme Court of Errors. See id., at 449–50, 30 A. 165 ; see also Conn. Const., art. V, § 1. In doing so, they “expressed the conviction of the people that a jurisdiction of mixed law and fact vested in any court of last resort, exercising a supreme and uncontrolled power, was inconsistent with a sound system of jurisprudence and was dangerous to the administration of justice....” Styles v. Tyler, supra, 64 Conn. at 451, 30 A. 165. The Superior Court thus holds “supreme jurisdiction” to try cases and settle disputes of fact, whereas the Supreme Court of Errors, which subsequently became this court, holds “supreme and final jurisdiction” to settle issues of law. Id., at 450, 30 A. 165. This court thus “is not a supreme court for all purposes, but a supreme court only for the correction of errors in law....” Id. Accordingly, “it is the exclusive province of the trial court to judge ... the credi[bility] of witnesses.” Dexter Yarn Co. v. American Fabrics Co., supra, 102 Conn. at 539, 129 A. 527.

The constitutional limits on fact-finding and assessing credibility cannot be circumvented merely by rebranding the nature of our inquiry, as the majority attempts to do in the present case. “[S]ettling the credi[bility] of witnesses, weighing the evidence, ascertaining the truth from conflicting or incongruous evidential facts, is a function within the exclusive jurisdiction of the trial court under our system of law, and is not reviewable by [this court]....

“Our jurisdiction cannot be enlarged, to permit the retrial of facts by us, by legislative enactment or rules of court ... and obviously not by the consent or acquiescence of the parties. [When] an appeal involves a request for the correction of a finding ... by a weighing of the evidence, it is patent that the real substance of such a proceeding is nothing but a retrial of pure questions of fact settled by the final judgment of the Superior Court. Its actual nature cannot be changed by calling it a correction of the finding.” (Citations omitted; internal quotation marks omitted.) Id., at 537–38, 129 A. 527 ; see also Thresher v. Dyer, 69 Conn. 404, 410, 37 A. 979 (1897) ; Atwater v. Morning News Co., 67 Conn. 504, 525–27, 34 A. 865 (1896). “A [retrial on] the testimony ... by whatever name it may be called, is a trial of the facts in that cause, whether its effect be limited to ordering a new trial, or extends to the rendition of a final judgment on the facts so adjudicated; and is inconsistent with the primary distinction drawn by the [c]onstitution, between the jurisdiction original and appellate of courts for the full trial and adjudication of causes, and the jurisdiction of a court of last resort for correcting errors in law which may have intervened in the course of a trial.... The Superior Court is the court of last resort for such purpose.... Such a principle embedded in the [c]onstitution ... cannot be evaded through any plan for accomplishing by a mere form of words the very evils it was adopted to prevent.” (Emphasis added.) Atwater v. Morning News Co., supra, at 526–27, 34 A. 865. Consequently, “there can ... be no doubt but that the determination by this court, upon the evidence, of questions of pure fact, for the mere purpose of rendering its own judgment upon issues of fact, is inconsistent with [this court's] jurisdiction, and clearly obnoxious to that underlying principle which holds the security of the citizen and the certainty of the law as best served by confining the supreme and uncontrolled power vested in a court of last resort for the correction of errors to the determination of principles of law.” Styles v. Tyler, supra, 64 Conn. at 456, 30 A. 165.

The jurisdictional limits that prevent us from assessing credibility do not relax, even when a defendant's claim implicates rights guaranteed by the federal constitution. Although we have explained that we have a duty to scrupulously review the record in reviewing a constitutional claim, we have made clear that this review does not allow this court to engage in its own credibility assessments: “Notwithstanding our responsibility to examine the record scrupulously, it is well established that we may not substitute our judgment for that of the trial court when it comes to evaluating the credibility of a witness.... It is the exclusive province of the trier of fact to weigh conflicting testimony and make determinations of credibility, crediting some, all or none of any given witness' testimony.... Questions of whether to believe or disbelieve a competent witness are beyond our review. As a reviewing court, we may not retry the case or pass on the credibility of witnesses.” (Internal quotation marks omitted.) State v. DeMarco, 311 Conn. 510, 519–20, 88 A.3d 491 (2014) (deference in context of reviewing fourth amendment claim); see also Sanchez v. Commissioner of Correction, supra, 314 Conn. at 604, 606, 103 A.3d 954 (deference in context of reviewing Strickland claim); State v. Krijger, 313 Conn. 434, 447–48, 97 A.3d 946 (2014) (deference in context of reviewing first amendment claim); State v. Mullins, 288 Conn. 345, 365, 952 A.2d 784 (2008) (deference in context of reviewing fifth amendment self-incrimination claim); State v. Santiago, 245 Conn. 301, 312–13, 715 A.2d 1 (1998) (deference in context of reviewing Brady claim).

Contrary to the Chief Justice's assertions in her concurring opinion, our deference to the trier's role as the arbiter of credibility does not depend on whether the trier is making “ultimate” credibility findings, and we have never changed our standard of review on this basis. The Chief Justice suggests, and the majority agrees, that we need not defer to the trial court when it is not acting as the “ultimate” fact finder but is merely assessing the probability that a jury will credit testimony. This contrived distinction appears nowhere in our law, which presumably explains why neither the Chief Justice nor the majority has offered any citations to support it. Indeed, we rejected this attempted distinction in the Strickland/Brady context just months ago in Sanchez v. Commissioner of Correction, supra, 314 Conn. at 585, 103 A.3d 954 a unanimous decision in which every member of the panel in the present case joined. In Sanchez, we indicated that there was no difference between our treatment of an assessment of whether a jury would likely credit new witness testimony and our treatment of a traditional credibility finding. See id., at 602 n. 12, 611–12 and n. 16, 103 A.3d 954. In doing so, we expressly agreed with the Appellate Court majority's rejection of Judge Sheldon's suggestion in his dissenting opinion in the Appellate Court that we could review this assessment de novo. Id.

Furthermore, we do not engage in de novo credibility assessments even in other, similar contexts requiring a predictive judgment about how a jury might weigh testimony. For example, we defer to the trial court's assessment of new witness credibility in the context of claims for a new trial based on newly discovered evidence, which also require a trial court to assess the likelihood that a jury will credit the new testimony and reach a different result than that reached at the original criminal trial. See, e.g., Shabazz v. State, 259 Conn. 811, 827–28, 792 A.2d 797 (2002) ; see also Adams v. State, 259 Conn. 831, 842, 792 A.2d 809 (2002) (“[i]n deciding a petition for a new trial ... the trial court sits as fact finder in place of the jury and examines the newly discovered evidence independently, in order to determine whether it is likely to result in a different verdict in the event of a retrial”). Even though newly discovered evidence claims do not require what the Chief Justice deems to be an “ultimate” finding but, rather, an assessment of how a jury might evaluate new evidence, we do not substitute our judgment for the trial court's judgment. This is true even when the new evidence is of a type that does not call for any firsthand assessment of demeanor such that we could evaluate the evidence in the same manner as a trial court. See Skakel v. State, supra, 295 Conn. at 487 n. 25, 991 A.2d 414 (video-recorded statement from witness).

Claims for prejudgment remedies also require a trial court to consider the probability that a party will succeed, which includes a predictive assessment of how a jury might gauge witness credibility. See, e.g., TES Franchising, LLC v. Feldman, 286 Conn. 132, 137–38, 143, 943 A.2d 406 (2008). A prejudgment remedy proceeding requires a trial court “ ‘to determine probable success' ” by the moving party; id., at 142, 943 A.2d 406 ; and is “not contemplated to be a full scale trial on the merits....” Id., at 143, 943 A.2d 406. Even though the trial court's credibility assessments in these proceedings are merely probabilistic, we review them only for clear error. Id. (“judging the credibility of the witnesses is the function of the trier of fact and this court will not usurp that role” [internal quotation marks omitted] ); see also Augeri v. C.F. Wooding Co., 173 Conn. 426, 428, 378 A.2d 538 (1977) (noting that we defer to trial court's credibility findings even though prejudgment remedy proceedings do not “ultimately” decide parties' claims).

Accordingly, the long-standing limits on our jurisdiction require that “[w]e must accept ... the trial court's decision [on] the question of credibility. It is beyond our province to weigh evidence and decide questions of this character.” Swist v.

Swist, 107 Conn. 484, 487, 140 A. 820 (1928). We repeatedly have explained, in a variety of contexts, that “[a]n appellate court must defer to the trier of fact's assessment of credibility....” (Internal quotation marks omitted.) State v. Lawrence, supra, 282 Conn. at 155, 920 A.2d 236. “The evaluation of the credibility of witnesses is left to the sound discretion of the trial court.” State v. Jones, 193 Conn. 70, 80, 475 A.2d 1087 (1984). “We cannot retry the facts or pass [on] the credibility of the witnesses.” (Internal quotation marks omitted.) Pandolphe's Auto Parts, Inc. v. Manchester, supra, 181 Conn. at 220, 435 A.2d 24. “[N]othing in our law is more elementary than that the trier is the final judge of the credibility of witnesses and of the weight to be accorded their testimony.” (Internal quotation marks omitted.) Kervick v. Silver Hill Hospital, 309 Conn. 688, 717, 72 A.3d 1044 (2013). “In light of our limited function, it would be improper for this court to supplant its credibility determinations for those of the fact finder,” even if the basis for those findings is reflected in “the cold printed record....” State v. Lawrence, supra, at 157, 920 A.2d 236.

4

The majority, attempting to justify its new standard, quotes at length from a case from another jurisdiction, but that case is entirely inapposite because we have already squarely rejected the principle underlying its holding. Given the overwhelming authority from this state expressly rejecting the principles underlying the majority's new standard, it is no surprise that the majority seeks support, not from a Connecticut case, but from a single case from an intermediate appellate court in Indiana, namely, Bunch v. State, 964 N.E.2d 274 (Ind.App.), trans. denied, 971 N.E.2d 1215 (Ind.2012). But whether appellate courts in Indiana can review the credibility of expert testimony de novo says nothing about whether a Connecticut appellate court can do so, especially in light of the restraints imposed by our state constitution. Also, contrary to the majority's assertion that the Indiana case is “factually and procedurally indistinguishable,” the portion of the case that the majority cites at length does not even involve a Brady or Strickland claim but an Indiana state law claim for a new trial based on newly discovered evidence. Id., at 288–89, 293–97. The court in Bunch applied a de novo standard of review because it felt it was in as a good a position as the trial court to decide credibility on the basis of a review of transcripts of new witness testimony presented live to the trial court when the trial court's assessment was not based on witness demeanor. See id., at 293. We rejected this principle in Skakel when we held that this court could not perform a de novo credibility review of newly discovered evidence, even when the trial court's analysis did not involve a firsthand assessment of witness demeanor. Skakel v. State, supra, 295 Conn. at 487 n. 25, 991 A.2d 414 (declining to conduct de novo review of findings concerning witness' video-recorded statement because we do not vary our scope of review even for findings “not predicated on an assessment of the witness' demeanor”). We have also made clear that credibility assessments of new witness testimony in the context of newly discovered evidence claims cannot be made by reviewing a printed transcript of a witness' testimony. Adams v. State, supra, 259 Conn. at 844, 792 A.2d 809 (credibility of newly discovered evidence “must be decided on the basis of the trial court's own assessment of credibility, not on the type of cold transcript utilized by the trial court”); cf. State v. Lawrence, supra, 282 Conn. at 156–57, 920 A.2d 236. If we cannot make credibility judgments even when the trial court bases its assessment on a video recording or a printed record, it is preposterous to conclude that we may do so when, as in the present case, the court based its findings on live testimony. The majority, justifying its reliance on Bunch, explains that newly discovered evidence claims require “exactly the same analysis as claims under Brady and Strickland, as they entail the same considerations.” Text accompanying footnote 62 of the majority opinion. Given this acknowledgment by the majority, it seems to me that we should look to our own law on newly discovered evidence claims, which rejects the majority's approach, rather than to rely on a lone case from another state to justify a new standard of review for this case.

5

The majority's creation of its new standard carries with it the danger of unwarranted expansion of appellate powers. In support of its new standard, the majority sweepingly proclaims that an appellate tribunal “need not, and will not,” defer to a trial court on any issue whenever “we are in as good a position as [that] court to decide the issue....” Text accompanying footnote 64 of the majority opinion. It does so without any citation or attempt to reconcile this statement with our prior decisions and constitutional principles, all of which expressly reject this assertion. See, e.g., Anderson v. Commissioner of Correction, supra, 313 Conn. at 375, 377–80 and nn. 7–8, 98 A.3d 23 ; Skakel v. State, supra, 295 Conn. at 487 n. 25, 991 A.2d 414 ; State v. Lawrence, supra, 282 Conn. at 156–57, 920 A.2d 236. Relying on this principle, the majority suggests that we can review credibility assessments de novo whenever this court deems itself as fit as the trial court to do so. Based on the majority's assertions that its new standard applies to the type of predictive fact-finding characteristically reserved for Brady claims, it applies at least to new witness testimony in the context of Strickland, Brady and newly discovered evidence claims, and perhaps also to prejudgment remedy proceedings. In light of the number of Strickland, Brady and newly discovered evidence claims based on new witness testimony, the majority's new standard would increase the burden on our appellate tribunals and litigants by requiring them to retry issues settled by the habeas court. Under the majority's new standard, appellate courts would have not only the ability, but also the duty to undertake a de novo review of witness credibility when the criteria for applying the majority's new standard are met. This new standard would require our appellate courts to make their own credibility findings whenever the habeas court does not base its decision on subjective factors, like candor or demeanor. It would also require litigants to retry these issues in our appellate courts by briefing factual disputes in addition to the legal issues. If the appellate tribunal can make its own credibility findings in this context, the habeas court's purpose in these cases will be relegated to little more than receiving and recording the evidence, and then issuing a recommended ruling for the appellate tribunal to mull over, accept or cast aside, as it sees fit.

Recognizing the dangers associated with the new standard it applies in this case, the majority immediately sets out to contain the damage done by trying to prevent its application in future cases. The majority acknowledges that we must defer to the trier's credibility assessments but nevertheless asserts that its holding is but a narrow exception to our ordinary standard of review. The majority acknowledges that its approach is unprecedented in our law but claims that this is because “we previously have not had a case on all fours with this one.” Text accompanying footnote 61 of the majority opinion. It characterizes its new standard as a “limited exception” necessitated by the “uncommon” and “highly unusual” circumstances of the present case, and explains that it will apply only in the “rare” cases that are “indistinguishable in any material respect from this one.” But the majority does not explain exactly what makes this case so exceptional. Nor does it explain why this case requires us to depart from decades of cases explaining that credibility is a matter for the trier without regard to the purported basis for the trier's findings. Moreover, the majority does not square its claim that the holding in this case is limited with its broad pronouncement that we will perform a de novo review of any issue whenever we deem ourselves to be in as good a position as the trial court to do so—a principle that we heretofore have rejected unequivocally. If it is true that we may review any question de novo whenever we deem ourselves to be as well situated as the trial court, how can the majority simultaneously claim that its holding in this case is limited to the facts of this case? Why is this case so exceptional? II

REVIEWABILITY

The majority's decision to undertake a de novo review of witness credibility injects issues into the case that have not been raised or briefed by the litigants, which ordinarily requires us to allow the parties to submit supplemental briefing on before we go sifting through the record in order to make our own credibility findings. The majority, on its own initiative, has changed our law by adopting a de novo review of credibility findings and has proceeded to undertake that review even though neither party raised any arguments in this court or in the Appellate Court with respect to the habeas court's credibility findings, and neither party has asked us to change our standard of review to allow this court to conduct its own credibility assessments. As a result, neither party has weighed in on the propriety of the majority's new standard of review or the underlying fact and credibility disputes that the majority resolves.

The principal issue actually raised in the present case is one of law. It concerns the petitioner's burden to establish the credibility of new evidence in order to prove Strickland/Brady prejudice and the habeas court's role in determining whether the petitioner has met that burden. Although the habeas court did not credit the petitioner's expert testimony, the Appellate Court decided that credibility of the new witnesses should be left to a new jury, not to the habeas court, a conclusion that the respondent has challenged on appeal to this court. None of the parties discussed, either in their briefs or at oral argument, whether an appellate tribunal can make its own credibility assessments. The arguments in this court regarding the proper legal standard assumed the propriety of leaving credibility assessments entirely to a new jury, and focused on the extent of the habeas court's role in judging new witness credibility and the standard by which the habeas court should determine whether a new jury should hear the new testimony at a new trial.

In addition, despite the habeas court's decision not to credit the burn time evidence from the petitioner's experts, the petitioner did not raise any claim attacking those findings before the Appellate Court or before this court. The respondent noted the absence of any such claim in his brief to this court and at oral argument, and the petitioner did not object to these repeated assertions. Indeed, in his brief to the Appellate Court, the petitioner agreed with the habeas court's conclusion that his evidence could not credibly establish the fire's burn time to a degree that would bolster his alibi defense. The habeas court concluded as follows: “What is clear from all the evidence in the record, the original trial testimony, crime scene photographs, reports, and the expert testimony presented to [the habeas] court [concerning] the fire, is that the precise time the fire was set cannot be determined. At best, a range is established that includes that time period of 6:15 p.m. to [7] p.m. (fire could have been burning between 5:45 p.m. and 7:55 p.m.), when [Karen] Martin cannot account for the petitioner's whereabouts and does not provide an alibi for him.” (Emphasis added.) The petitioner, tracking the habeas court's language, agreed, stating “it is true that the experts could not determine the exact amount of time the fire burned ” and that “the fire's burn time could not be precisely determined ....” (Emphasis added.) Rather than attacking the habeas court's findings, the petitioner has instead argued that the Appellate Court properly concluded that it is up to a jury, not the habeas court, to determine whether to credit his experts, each of whom was qualified to render an opinion. Accordingly, neither party has raised any claim of error based on the habeas court's credibility findings. Nor has either party suggested that an appellate court should make its own credibility assessments.

Failure of a party to raise a claim on appeal results in abandonment. See, e.g., State v. Cyrus, 111 Conn.App. 482, 487, 959 A.2d 1054 (2008), aff'd, 297 Conn. 829, 1 A.3d 59 (2010). In the absence of any challenge to the habeas court's credibility findings, or any argument by the parties that we should make our own findings, the issues that the majority decides are not properly before this court. See, e.g., State v. Crumpton, 202 Conn. 224, 231–32, 520 A.2d 226 (1987).

To dampen the criticism that it receives from the dissenting justices for deciding unraised issues, the majority declares that the issues it decides were already decided by the Appellate Court and thus are not new to this appeal. To that end, the majority repeatedly asserts that the Appellate Court performed its own de novo assessment of the credibility of the petitioner's new expert testimony and that the Appellate Court rendered its own finding that there was a reasonable probability that a jury would credit it. The majority quotes at length from the Appellate Court's decision but missing from its quotation is any de novo assessment of the “soundness” of the expert testimony or credibility findings by the Appellate Court. The majority has not provided its readers the courtesy of a citation to what page or pages this review and these findings appear in the Connecticut Appellate Reports. The fact is, the Appellate Court did not perform its own de novo credibility assessment, nor did it make its own findings.

Any question on this point is dispelled by a fair reading of the Appellate Court's decision. Absent from the Appellate Court's decision is any mention that it is undertaking a de novo review of the credibility of expert testimony. Nor does the Appellate Court give any justification for such a review. Given that such a review by an appellate tribunal in this state would be unprecedented, at least until today, it would be surprising for the Appellate Court to sail into uncharted waters without at least warning its readers that it is doing so. Also missing from the Appellate Court's decision are any findings resulting from its purported de novo credibility review. Instead, the Appellate Court repeatedly conditioned its prejudice analysis by stating that the petitioner's evidence, “if believed by the jury, ” could lead to a different result. (Emphasis added.) Lapointe v. Commissioner of Correction, supra, 138 Conn.App. at 479, 53 A.3d 257 ; see also id., at 476–77, 53 A.3d 257. I, for one, would expect that, if the Appellate Court were performing its own unprecedented de novo review of witness credibility, it would at least state its findings on the record, as our trial courts are required to do. Rather than making its own findings, the Appellate Court stated that assessing the credibility of the expert witnesses was a task best left to a jury: “If the Ludlow note had been disclosed to trial counsel, however, it would have been the responsibility of the jury and not the court to weigh the credibility of the arson experts. Whether the burn time evidence, which was so critical in buttressing [the petitioner's] alibi defense, raised a reasonable doubt as to the petitioner's guilt would best be a determination left to the jury and not a habeas court. ” (Emphasis added.) Id., at 476–77 n. 17, 53 A.3d 257. Furthermore, neither party argued in the Appellate Court that it should be performing its own de novo credibility review. Nor did the parties brief the issue of whether such a review by an appellate tribunal would be appropriate. As I previously mentioned, the petitioner did not challenge the habeas court's findings, but, instead, the parties argued over the proper standard to be applied by the habeas court.

In light of these statements by the Appellate Court, I simply do not agree with the majority's conclusion that the Appellate Court performed its own credibility assessment. By leaving the credibility question to a jury, the Appellate Court freed any court considering the petitioner's Strickland/ Brady claim of any need to inquire into the credibility of the expert testimony or the need to make any findings of its own. Deciding that a jury and not a habeas court should assess credibility in this context is markedly different from deciding that an appellate tribunal may do so. Both the habeas court and a jury properly may serve as the trier of fact and arbiter of credibility, whereas an appellate tribunal, except for the majority in the present case, may not. Consequently, the considerations that are needed to resolve these two issues are wholly distinct. I thus disagree that the Appellate Court would undertake such an unprecedented review without even stating that it was doing so or attempting to justify such an analysis.

Because neither party has claimed, either in this court or the Appellate Court, that an appellate tribunal can properly make its own credibility assessments, or that the Appellate Court in this case did so, we have no briefing from the parties concerning whether such an analysis by an appellate tribunal is proper. The lack of a de novo review by the Appellate Court, and any suggestion by the parties that such a review is proper, is fatal to the majority's claim that the parties raised and briefed arguments about the propriety and the merits of such a review by an appellate tribunal. Neither party has briefed the issue of whether an appellate tribunal may properly render its own credibility findings in this context. The lack of any argument on this issue stands in stark contrast to the briefing in Anderson, in which the respondent expressly argued that an appellate tribunal could not render its own credibility findings, a position that we ultimately followed in that case. Instead, the respondent argues in this case precisely what I have explained in this opinion, namely, that the Appellate Court improperly disregarded the habeas court's role, left the credibility assessment to a jury, and then reached its materiality determination by “speculating on the basis of what ‘could have’ happened ‘if’ certain evidence had been believed.... In saying that jurors could have reached a conclusion if they believed certain testimony, the Appellate Court said no more than that this was possible or conceivable,” which is insufficient to establish Strickland/Brady prejudice. (Citations omitted.) Whether credibility assessments properly may be left to a new jury, rather than a habeas court, is a much different issue than whether an appellate tribunal may make its own credibility assessment. In light of the respondent's briefing on this issue in Anderson, I find it hard to believe that the respondent would remain silent on this same issue in the present case if it were truly presented to us in this appeal.

Moreover, because neither party made any claim of error regarding the merits of the credibility questions put into issue by the majority, neither party has provided briefing on that topic. The only discussion of the credibility of the experts comes from a single, half page footnote in the respondent's brief. Most tellingly, that footnote was appended to a sentence in the text that explained that the petitioner had not raised any claim of error concerning the habeas court's credibility findings at any point during the appeal. Furthermore, in that footnote, the respondent explained that the record did not permit a finding of clear error; he mentioned nothing about what an appellate tribunal could find if it performed a de novo review. It would be naive to think that, if the respondent truly had notice that this court would conduct its own de novo review of the credibility of new witness testimony, the respondent would have dedicated only a single footnote to this fact intensive inquiry, especially considering that such an analysis by this court is unprecedented in our law and requires consideration of numerous days of testimony and exhibits from the petitioner's habeas hearing and the criminal trial. The respondent's decision to relegate any discussion of expert witness credibility to a footnote and to couch it in terms of a clear error analysis is proof of only one thing: the parties did not raise this issue and were not on notice that this court might perform its own credibility assessment. Even the petitioner recognized that the respondent was arguing not that the Appellate Court performed a de novo review, but that the Appellate Court acted improperly by deciding that the credibility assessment was a “jury issue” and not an issue for the habeas court. The majority's uninvited change to our law has therefore left us with a task that we have never before undertaken and without any guidance from the parties as to how we should resolve their dispute.

Before we alter our well established and heretofore unquestioned standards of review and take the unprecedented step of rendering our own credibility findings, we must at least provide the parties with an opportunity to submit supplemental briefs on these issues. In Blumberg Associates Worldwide, Inc. v. Brown & Brown of Connecticut, Inc., supra, 311 Conn. 123, 84 A.3d 840 we determined that we may, sua sponte, consider issues not preserved or raised by the parties, subject to certain restrictions. See id., at 128, 161–64, 84 A.3d 840. We explained that a reviewing court may, but is not required to, review “issues involving plain error or constitutional error sua sponte, as long as the court provides an opportunity for the parties to be heard by way of supplemental briefing and the other threshold conditions for review are satisfied.” Id., at 161–62, 84 A.3d 840. “Of course, as we have explained in connection with a reviewing court's consideration of a claim raised for the first time on appeal, if a party objecting to the reviewing court's sua sponte consideration of the claim can demonstrate that it would be unfairly prejudiced by such consideration, it would be inappropriate for the appellate tribunal to consider such a claim. Furthermore, once the objecting party makes a colorable showing of unfair prejudice, the burden shifts to the other party to establish that the objecting party will suffer no such prejudice.... [W]e also reiterate that our system is an adversarial one in which the burden ordinarily is on the parties to frame the issues, and the presumption is that issues not raised by the parties are deemed waived.” Id., at 163–64, 84 A.3d 840. The majority has not explained what considerations led it to address issues not raised by the parties. Even if a de novo review of the habeas court's findings might be warranted, the principles delineated in Blumberg require this court to first assess whether addressing the issue would be unfairly prejudicial to the respondent. This prejudice analysis is a necessary predicate to addressing an unraised issue. Even if we find no prejudice to the respondent in addressing this claim, we must still give the parties an opportunity to address the issues that the majority decides.

Reaching this issue on our own and without any input from the parties harms the parties and undermines the fairness of our judicial process. Due process compels us to give the parties notice and an opportunity to be heard. The record in this case is voluminous. It consists of more than eighty volumes of transcripts with thousands of pages of testimony from a probable cause hearing, suppression hearing, three habeas hearings, and a lengthy criminal trial. There are hundreds of exhibits from these proceedings. The significance of certain evidence may not be obvious to us from our own, unguided review of such a voluminous record. The parties have greater knowledge of the evidence in the record and how it got there. They also have a more complete perspective of the context in which evidence was presented and its import. Facts viewed in isolation are not as powerful as facts woven into a coherent and compelling argument. That is the purpose of briefing. By denying the parties notice and a chance to brief the issue that the court decides, we may be silencing valid arguments not obvious from our own review. Even if that briefing does not change the majority's conclusions, the losing party deserves the solace of knowing that it has been fairly heard.

The harm to the respondent that results from a denial of notice and the right to be heard is not just theoretical—it is demonstrated in the majority opinion. In several places in its opinion, the majority correctly observes that the respondent was silent about the credibility of the petitioner's experts, noting, for example, that “the respondent makes no attempt to defend the ... habeas court's assertion that there was a material difference” in the substance of the burn time experts' testimony. But this silence is significant only to demonstrate that the respondent was not afforded the notice that he needed to address this unraised issue. Does one seriously doubt that, if the respondent had been afforded notice that the habeas court's findings were under attack, he might have addressed these claims before this court rather than remaining silent on these points? The respondent briefed his case on the understanding that the petitioner had not challenged the habeas court's findings in the Appellate Court or in this court, and specifically explained in his brief that the petitioner had not raised such a claim. The majority is thus deciding an issue that was not raised and is holding the respondent's silence on that issue against him. The harm caused by silencing the respondent, as a representative of the state, is especially acute in this case. The majority's decision to order a new criminal trial could well be the functional equivalent of a directed judgment of acquittal. The events underlying this case occurred more than one quarter century ago. Memories may have faded, remaining evidence may be of a questionable quality, and key witnesses, some of whom were advanced in years at the time of the petitioner's criminal trial, might have passed away. The majority's opinion leaves the state in the difficult position of having to retry this case long after the events at issue occurred. Certainly, if the petitioner's conviction was obtained unfairly, then he deserves to have his conviction vacated. But the state, through the respondent, is, at a bare minimum, entitled to be heard before we charge its appointed representatives with withholding exculpatory evidence and vacate the petitioner's conviction on that basis.

For these reasons, the majority should either decide only the issues raised by the parties or seek additional briefing in light of its decision to address unraised issues.

III

THE MAJORITY'S CREDIBILITY FINDINGS

Even if I assume, for the sake of argument, that we can review the habeas court's findings de novo, I still am not persuaded that the petitioner is entitled to a new criminal trial. The majority finds the petitioner's new expert testimony to be credible after a purportedly de novo review of the record. But it spends much of its analysis on a lengthy diatribe against the habeas court's findings—an unnecessary endeavor in light of the majority's assertion that no deference to those findings is warranted. Indeed, rather than looking for evidence to prove credibility, the majority seems to presume that the new testimony is credible and searches for reasons to discredit it. The majority begins by simply attacking the habeas court's findings. Then, after determining that the habeas court's “stated reasons for discrediting the burn time estimates of [the petitioner's experts] are baseless,” the majority determines whether there is some “other apparent reason why a jury would be apt to discredit their testimony.” The majority finds no such reason but does so after briefly considering only the experts' qualifications and the fact that the respondent has not challenged their credibility. This is insufficient to establish credibility. First, determining that the habeas court's reasons underlying its credibility determinations were incorrect does not necessarily establish that the testimony of the new experts is credible. Second, an expert's qualifications to render an opinion, while necessary to secure admission of expert testimony, do not automatically prove that there is a reasonable probability that it will be credited; that determination requires a review of all of the evidence and testimony presented at the habeas trial, which the majority does not perform. Third, as I mentioned previously, the respondent's silence on this issue is hardly a basis for making a credibility finding, given that this issue was not raised. The respondent could not seriously be expected to know that this court would raise this issue sua sponte after relying on an inapposite case from another jurisdiction. Furthermore, by looking to the respondent for criticisms of the petitioner's experts, and finding the new testimony credible in their absence, the majority improperly shifts the burden of proof to the respondent.

Even if I assume that the petitioner had challenged the habeas court's credibility findings, my review of the record persuades me that the findings were correct. My review reveals that the new expert testimony simply is not reliable because it lacks adequate foundations in fact and science, and because the opinion testimony of the petitioner's two experts substantially conflicted in material respects. First, as the petitioner conceded in the Appellate Court, his own experts' testimony demonstrated that the fire's burn time cannot be reliably determined to any helpful degree. The petitioner's experts, who were hired more than one decade after the fire, relied on speculation and unproven methods to reach their burn time estimates. They also relied on factual assumptions that were refuted by eyewitness testimony. Additionally, the petitioner's experts disagreed with each other about key data relevant to determining the fire's burn time and conceded that their inability to analyze the fire scene firsthand could impact their conclusions. Second, other fire investigators, including the only two investigators that examined the actual fire scene, testified that the fire could have burned for several minutes to several hours—a time period that does not support the petitioner's alibi—and further testified that there was not enough data available to reliably develop a more specific determination with respect to the fire's burn time.

Without a valid and reliable factual and scientific foundation, the petitioner's expert testimony would not even be admissible, let alone credible. See, e.g., State v. Porter, 241 Conn. 57, 74, 698 A.2d 739 (1997), cert. denied, 523 U.S. 1058, 118 S.Ct. 1384, 140 L.Ed.2d 645 (1998) ; State v. John, 210 Conn. 652, 677, 557 A.2d 93, cert. denied, 493 U.S. 824, 110 S.Ct. 84, 107 L.Ed.2d 50 (1989). To be admissible, an expert's opinion must have a basis in facts that are supported by the evidence; the factual basis for an expert's opinion may not be based on speculation. See Viera v. Cohen, 283 Conn. 412, 449, 927 A.2d 843 (2007). “In order to render an expert opinion the witness must be qualified to do so and there must be a factual basis for the opinion.... [When] the factual basis of an expert opinion is challenged ... the question before the court is whether the uncertainties in the essential facts on which the opinion is predicated are such as to make an opinion based on them without substantial value.” (Citations omitted; internal quotation marks omitted.) State v. John, supra, at 677, 557 A.2d 93. Thus, without an adequate basis in fact, an expert's opinion cannot be admitted into evidence, let alone credited. See, e.g., id. Additionally, the methodology that an expert uses to reach an opinion on scientific issues must be scientifically valid. See State v. Porter, supra, at 83–87, 698 A.2d 739. An expert's personal experience alone, even if extensive, is not sufficient to establish the scientific validity of the expert's methodology. Klein v. Norwalk Hospital, 299 Conn. 241, 263, 9 A.3d 364 (2010). With these principles in mind, I turn to my review of the expert testimony presented at the petitioner's habeas hearing and the testimony and exhibits from the criminal trial relating to the testimony of the expert witnesses.

A

Testimony from the Petitioner's Experts

1

Turning first to the testimony of DeHaan, I note that the record demonstrates that his estimate was unreliable because the principal factual underpinnings that DeHaan relied on in support of his burn time estimate are contradicted by eyewitness testimony, as well as his own testimony, and thus are not supported by the evidence. DeHaan estimated that the fire burned for at least twenty-five minutes but no more than sixty minutes. DeHaan based his estimate on two key factual assertions. First, DeHaan assumed that the fire on the couch “burned very readily,” quickly consuming the available oxygen in the apartment, and then was reduced to a smolder. Second, DeHaan estimated that temperatures inside the apartment were about 400 degrees when the first firefighter on the scene, Michael Tomkunas, kicked in the front door because the high heat in the apartment prevented him from entering it. But each of these two critical foundations for DeHaan's burn time estimate is contradicted by testimony from eyewitnesses, and not supported by any other evidence.

First, DeHaan's belief that the fire on the couch “burned very readily” is contrary to the observations of a fire investigator who actually witnessed the speed at which the couch burned. Igoe, the state fire marshal who originally investigated the fire, testified at the criminal trial and again before the habeas court. Igoe arrived at the scene within a few hours of the fire and remained on the scene for about five hours to conduct his investigation. During his investigation, Igoe examined the couch where the fire originated and the inside of the victim's apartment to assess the damage and the fire's burn pattern. He also conducted a test burn of the couch to determine how it burned by burning the actual material from the couch in an oxygen rich environment. Igoe testified at the habeas hearing that, even with plenty of oxygen, the couch material “burned very, very slowly and it emitted heavy black smoke.” (Emphasis added.) In his report issued shortly after the fire, Igoe explained: “Examination of the living room couch leads this investigator to conclude that the burning to the couch was the slow smoldering type burning and not the rapid burning which occurs with some materials .” (Emphasis added.)

In reaching his own conclusions, however, DeHaan simply rejected these eyewitness observations about the speed at which the couch burned, even though DeHaan had neither examined the couch materials nor witnessed the speed at which they burn. The petitioner, in his posttrial brief, acknowledged this rejection, explaining that “DeHaan disagreed with Igoe substantively on the fire's characteristics, most particularly that it was a slow burning fire.” DeHaan believed, from looking at photographs of the couch, that it was made of a material that burned rapidly and that it would not have been treated with flame retardant. But the materials involved in the fire were discarded or destroyed long before DeHaan became involved in the case. DeHaan thus did not examine the couch firsthand, could not undertake an analysis of its materials to determine their composition, and did not personally witness or conduct any test burns of the couch's materials. Nor did he explain how his theory was consistent with the results of Igoe's burn test. DeHaan's dismissal of Igoe's testimony, despite DeHaan's lack of any personal knowledge of how it burned, places the factual basis for his conclusions in direct conflict with eyewitness evidence presented to the habeas court.

Second, the factual basis for DeHaan's estimate of the temperature in the apartment when Tomkunas first arrived is contradicted by Tomkunas' testimony from the criminal trial. DeHaan based his estimate of the entry level temperature on his understanding that Tomkunas could not enter the apartment even though he suspected someone was inside. Tomkunas' inability to enter the apartment suggested to DeHaan that temperatures were between 300 and 400 degrees when Tomkunas first arrived, with the temperatures most likely closer to 400 degrees. This led DeHaan to conclude that the fire could not have been started more than one hour before Tomkunas tried to enter, or else the high temperatures would have dissipated and Tomkunas would not have been deterred from entering the victim's apartment.

The majority relies heavily on DeHaan's understanding that Tomkunas could not enter the apartment because of the heat to call into question the habeas court's reliance on the testimony of the respondent's expert, Robert Corry. According to the majority, DeHaan testified “that the temperature inside the hot gas layer was likely between 300 and 400 degrees when Tomkunas entered the apartment. DeHaan explained that this estimate was based, in part, on Tomkunas' testimony at the petitioner's criminal trial that, when he arrived at the victim's apartment, the outside of the front door was hot to the touch, and ‘the temperatures and ... hot gases he encountered at near floor level were untenable, and he couldn't go in, even though he suspected there was a victim inside. ’ ” (Emphasis in original.) DeHaan also testified that, if the temperature had been lower than 400 degrees, “Tomkunas would not have necessarily been dissuaded from entering.” DeHaan's burn time conclusion, therefore, was based on his understanding that Tomkunas could not enter the apartment at all because of the high temperatures and that Tomkunas entered the apartment only after the firefighters vented it, allowing the heat to escape. The habeas court discredited DeHaan's estimate of entry level temperatures based on Corry's testimony that Tomkunas entered the victim's apartment before it was vented and was not burned, indicating that temperatures could not have been as high as DeHaan had suggested. Contrary to DeHaan's understanding that Tomkunas “couldn't go in[to]” the apartment because of the heat until after it was vented, Tomkunas testified that he was able to enter the apartment when he first arrived and before it was vented. Tomkunas, a volunteer firefighter, was driving in his personal vehicle less than one mile from the victim's apartment when he received a dispatch reporting the fire. Tomkunas, who was on a date at the time, did not have any protective gear or breathing apparatus with him. Nevertheless, he drove directly to the victim's apartment and arrived there about one minute after the dispatch. The petitioner directed Tomkunas to the apartment, and he went straight to the front door and kicked it in. Tomkunas was met with what he called a “wall of smoke.” It was nighttime, and the apartment was filled with dense smoke, making it difficult for Tomkunas to see inside the apartment. Despite these conditions, Tomkunas was able to enter the apartment on his first attempt. He dropped to his knees in an attempt to stay below as much of the smoke as possible and crawled in. He was able to crawl about eight feet into the apartment and within only three feet of the burning couch. Tomkunas saw a small flame on the still smoldering couch. Tomkunas estimated that he was in the apartment for about fifteen to twenty seconds before the heat and smoke conditions convinced him that rescue efforts would be “a lot easier” if the apartment was vented. Tomkunas crawled back out the front door and asked another firefighter who had arrived to open the back door of the apartment to allow the heat and smoke to escape. Tomkunas immediately crawled back into the apartment to continue his search. He estimated that he was outside of the apartment for less than five seconds before reentering. During his second entry into the apartment, Tomkunas again crawled to within only three feet of the burning couch. This time, he found the victim and removed her from the apartment with the help of another firefighter. At no time during his testimony did he say that he was prevented from entering the apartment as a result of the heat. DeHaan's assertion that the high heat prevented Tomkunas from being able to enter the apartment altogether is, therefore, at odds with Tomkunas' actual testimony.

In addition, Tomkunas' ability to remain in the apartment for fifteen to twenty seconds before it was vented and without suffering any burns completely contradicts DeHaan's claim that the apartment was close to 400 degrees when Tomkunas first kicked in the door. Corry testified that, according to data from the National Fire Protection Association, a person with uncovered skin will sustain first or second degree burns when exposed to 212 degree temperatures for fifteen seconds; burns will occur in even less time at higher temperatures. Although DeHaan estimated that Tomkunas would have experienced temperatures near 400 degrees “to the bare skin,” Tomkunas did not sustain any burns or injuries from the heat or smoke on either of the two occasions that he entered the apartment.

Tomkunas's testimony directly contradicts DeHaan's and the majority's understanding that the 400 degree heat prevented Tomkunas from even entering the apartment. The habeas court discredited DeHaan's testimony in part because it credited Corry's testimony that entry level temperatures could not have been close to 400 degrees as DeHaan estimated because Tomkunas was able to enter the apartment without suffering any burns. The majority concludes, however, that this finding “lacks support in the record.” According to the majority, the approximately 400 degree temperatures prevented Tomkunas from even entering the apartment before it was vented, and that was why Tomkunas did not get burned. The majority explains that “Tomkunas had testified ... that the heat and smoke prevented him from entering the apartment when ... he first attempted to do so,” and it criticizes Corry for “simply ... refus[ing] to accept” this fact. (Emphasis in original.) But Tomkunas' testimony makes clear that he did enter the apartment on his first attempt and remained in there for about fifteen to twenty seconds without suffering any burns.

I also note that Corry's opinion received additional support from Corry's own interviews of Tomkunas. As part of his investigation of the fire, Corry personally interviewed Tomkunas twice about his actions on the night of the fire and the heat conditions that Tomkunas encountered. Corry testified during the habeas hearing that, “when [he] asked [Tomkunas] to compare [the heat he experienced upon entering] to something that he was familiar with, he said at most it was similar to entering a sauna, which would be about 150 degrees to 190 [degrees].” DeHaan, on the other hand, did not speak to Tomkunas or even attempt to contact him; he instead claims to have relied on Tomkunas' trial testimony, which, as I explained previously, is inconsistent with DeHaan's understanding of the conditions that Tomkunas experienced.

DeHaan's assumptions in support of his estimates of entry level temperatures are thus completely refuted by Tomkunas' testimony. We do not know what impact, if any, Tomkunas' ability to enter the apartment had on DeHaan's conclusions because DeHaan's stated understanding of Tomkunas' actions simply does not line up with Tomkunas' testimony. When, as in the present case, an expert's opinion lacks a valid basis in fact, it is nothing more than speculation and may not be admitted into evidence, and, as a result, cannot be credited. See, e.g., State v. John, supra, 210 Conn. at 677, 557 A.2d 93.

In addition, DeHaan's belief that entry level temperatures were too high for Tomkunas even to enter the apartment conflicts with DeHaan's own testimony about the fire's energy level. DeHaan testified that the fire would have produced a relatively small amount of heat and would have been approachable, even at its maximum intensity. DeHaan explained that the fire's intensity “would probably be about the same as an average fireplace fire. It would be pumping a lot of heat into this room, but not so much that you couldn't—you couldn't approach it, for instance, to try to extinguish it, and if you were there at the time ... it reached its maximum ....” (Emphasis added.) DeHaan did not explain how the fire could have been approachable at its maximum intensity but unapproachable when Tomkunas tried to enter the apartment, which, according to DeHaan, likely occurred after the fire passed its peak and the apartment had begun to cool. DeHaan's unsupported and conflicting testimony fully supports the habeas court's conclusion that DeHaan's burn time estimate was too speculative to be reliable, rendering it inadmissible.

2

Turning next to the testimony of the petitioner's other burn time expert, Gerard Kelder, Jr., I am persuaded that the record supports a finding that his estimate was also unreliable. The petitioner did not provide any evidence to show that Kelder used scientifically valid methods to investigate the fire, rendering his testimony not credible and inadmissible.

Kelder testified that he developed his burn time estimate principally by reviewing the fire damage depicted in photographs and a video recording. During the habeas hearing, the respondent's attorney asked Kelder what scientific method he followed to investigate the fire. Kelder initially testified that he relied on his personal experience and that his procedures tracked those described in chapter four of the National Fire Protection Association's Publication 921, a widely used fire investigation guide. After being shown a recent edition of Publication 921, Kelder said that he had an older edition and would have to review the publication to recall its basic procedures. After reviewing it, Kelder changed his testimony about his method and admitted that he did not follow those procedures to investigate the fire but used his own procedure based on his personal experience instead.

Kelder did not explain whether the method he contrived from his own personal experience was a scientifically valid and reliable method for determining a fire's burn time. Kelder acknowledged that his “estimate of the ignition time of the fire was based largely on the amount of damage that [he had] observed,” but there was no other testimony to establish that a review of fire damage through photographs and a video recording of the scene is a valid and reliable method for determining how long a fire burned. Additionally, when describing his method, Kelder stated that the method he used to determine the burn time was based on his experience in “determining [the] cause and origin ” of a fire; he said nothing about his experience or method for determining how long a fire burned. (Emphasis added.) The only evidence in the record about whether this method was valid came from Corry, who testified that examining photographs of fire damage was not a viable method for determining a fire's burn time.

The respondent attacked Kelder's testimony before the habeas court on this very basis. In his posttrial brief, the respondent argued that Kelder “could identify no scientific principle that would support his assertion that he [could] determine the duration of the fire simply by viewing photographs of the damage that it caused. Indeed, when Kelder's testimony is considered in its entirety, it becomes clear that it was nothing more than pseudo-scientific babbling. His conclusion, therefore, should not be credited....” Not surprisingly, the petitioner, in his posttrial brief, cited to Kelder's testimony only once.

A court must preclude an expert's opinion when there is no evidence in the record to support the validity of the expert's methods other than the expert's own “ipse dixit” that those methods are valid and reliable; a witness' personal experience is not enough. Klein v. Norwalk Hospital, supra, 299 Conn. at 263, 9 A.3d 364 ; see also id., at 262–63, 9 A.3d 364 (“[T]he defendant made no showing that [the expert witness'] methodology had been subjected to peer review, nor was [the witness] able to identify a likely rate of error for his chosen methodology. [Although] neither of these determinations is a talismanic requirement for satisfaction of the Porter requirements, their absence is ... determinative of the inadequacy of the defendant's proof of the methodology's reliability.... Without these or any other meaningful indicia of reliability, [the witness'] conclusion was without basis in an assuredly reliable methodology; without any stated support for its reliability other than his own personal expertise, it was nothing more than his ipse dixit. ” [Citation omitted; emphasis added.] ).

The habeas court in the present case was fully justified in concluding that Kelder's opinion would be excluded from a criminal trial altogether. See id., at 263, 9 A.3d 364 (“[n]othing ... requires a ... court to admit opinion evidence that is connected to existing data only by the ipse dixit of the expert” [internal quotation marks omitted] ). The petitioner's failure to present facts to establish the validity and reliability of Kelder's method would render his testimony not credible and inadmissible.

3

Other than the inherent unreliability of their estimates, Kelder and DeHaan disagreed with each other about key data points relevant to a determination of the fire's burn time, including the maximum temperature in the apartment, the temperature inside the apartment when Tomkunas first arrived, and the materials in the couch where the fire principally burned.

At the habeas hearing, Kelder testified that the maximum temperature in the apartment as a result of the fire would have been between 1800 to 2000 degrees at ceiling level. When asked how he reached this estimate, Kelder replied, “[i]t's scientific proof.” DeHaan testified, however, that the temperature in the apartment could not possibly have exceeded 400 to 450 degrees. According to DeHaan, there were a number of objects throughout the apartment, including within the hot smoke layer, that “had been affected, but not ignited,” and, if the temperature had been any higher than 450 degrees, “a lot of other things would have been ignited and burned.”

Kelder and DeHaan also disagreed about the temperature of the apartment when Tomkunas first kicked in the front door. Each of the petitioner's experts agreed that this was an important factor in determining the fire's burn time. As I discussed previously, DeHaan estimated that the entry level temperature was about 400 degrees. Kelder, on the other hand, initially testified that “there was not a lot of heat in that room at the time [the firefighters] entered” because, if there had been, there would have been a “flashover ... of fire from the new oxygen coming into the room” when the firefighters first opened the door. Later in his testimony, however, Kelder gave an estimate of 600 to 800 degrees, which far exceeded DeHaan's estimate of both the entry level temperature and the maximum temperature that the apartment could have reached as a result of the fire. Yet, under further questioning, Kelder testified that he was “not really sure” what the entry level temperature was because he “wasn't at the fire scene,” so any estimate of the entry conditions would be “an assumption on [his] part.” In addition to demonstrating the inconsistent and contradictory nature of the petitioner's evidence, Kelder's testimony on this point is yet further evidence—from the petitioner's own expert —that DeHaan could not make a reliable estimate of the entry level temperature.

Furthermore, the petitioner's experts disagreed over whether the couch cushions were made of rubber foam or polyurethane foam, which burn at different rates, a difference that could have affected the fire's burn time. DeHaan testified that he was sure that the couch cushions were made with rapidly burning polyurethane foam, which led him to reject Igoe's eyewitness observations that the fire burned slowly. According to DeHaan, if the couch cushions were made of slower burning rubber foam, the remnants of the cushions would have appeared much differently than that depicted in the photographs, and rubber was not widely used as cushion material when the victim's couch was likely constructed. Nevertheless, the petitioner's other expert, Kelder, contradicted DeHaan's opinion and concluded, instead, that the couch cushions were made of rubber foam, which burns relatively slowly and does not easily ignite. Like DeHaan, Kelder also based his conclusion on his review of the photographs of the couch. And, unlike DeHaan, Kelder did not dispute Igoe's report that the couch material burned “very, very, slowly....” In fact, when asked whether the couch cushions could have been made from polyurethane foam, which burns more rapidly than rubber foam, Kelder testified that he was “sure” that they were not made of polyurethane. He stated: “I can tell you it wasn't polyurethane foam because, if that were the case, any type [of] poly foam like that would ignite and burn rapidly. This did not [based on] the test done by ... Igoe.”

The petitioner's experts also admitted that their late arrival to the case, as well as their resulting inability to study the actual fire scene in person, could affect the validity of their conclusions. Neither Kelder nor DeHaan investigated the actual fire scene at the time of the fire. They were hired more than one decade after the fire to formulate burn time estimates based only on a review of photographs, a video recording and the testimony of certain witnesses. But the photographs and video recording were taken after emergency personnel had disturbed the original fire scene. For example, photographs of the couch where the primary fire burned show the couch outside of the victim's apartment on the lawn, even though all experts agree that the couch was in the living room at the time of the fire. Emergency personnel moved the couch outside before the first fire investigators arrived on the scene and before any photographs of the scene were taken. Moreover, neither Kelder nor DeHaan interviewed any of the witnesses to the fire, and neither of them personally examined or analyzed any of the materials involved in the fire.

Kelder acknowledged that important information about the conditions contributing to the fire's duration was unavailable to him because he joined the case long after the fire occurred. Kelder explained that he developed his burn time estimate by looking at photographs, a video recording and a diagram of the scene, and by reviewing certain testimony. He did not do anything further to gather additional information to help his investigation—for example, by interviewing witnesses—and explained that he was not asked to do this. Nevertheless, he agreed that it would have been a good investigative practice to do so, if he had been able to investigate the actual fire scene in “1987 versus 2000....” He also acknowledged that his inability to review, firsthand, the actual fire scene and materials involved in the fire prevented him from using standard fire investigation methods. When asked whether he knew the ignition temperatures of the materials involved in the fire, Kelder conceded that he “would have to test the materials” but that he “didn't have that availability” and therefore could not conduct his own tests. And when Kelder was asked whether it would be important for his investigation to determine exactly what type of material was involved in the fire, he answered: “Yes, if I had the availability of being at the scene, I think I would have sent it to the lab and determine what components were in it.” Kelder also explained: “I could not make that determination. I wasn't at the scene.” Finally, Kelder also acknowledged that it would have been “a good investigative practice to try to determine what the temperatures were when the firefighters entered the apartment” but that he could not do so because he “wasn't at the fire scene.”

In addition, DeHaan acknowledged the limited data available to him could impact his conclusions. When DeHaan was asked at the habeas hearing whether there was “anything [he] could have learned by examining any items that were retained from the fire,” he explained: “If the [couch] had been retained, yes, that would have been useful to physically examine it and specifically identify the fabrics and filler rather than by photographs.” Although DeHaan based his burn time opinion primarily on his estimate of the temperature that Tomkunas was exposed to when he first arrived at the apartment, DeHaan acknowledged that he was able to estimate that temperature “[o]nly very approximately....” And, in his written report, DeHaan concluded that, “[d]ue to the limited data available as to the insulation and ventilation conditions existing in [the victim's apartment], it is not possible to predict the postfire duration of high temperatures. ” (Emphasis added.) Nevertheless, DeHaan ultimately based his opinion on an estimate of the postfire duration of high temperatures. When confronted during the habeas hearing with this apparent inconsistency, DeHaan admitted that it is not possible to determine the fire's duration “[w]ith any accuracy, that's right, with precise accuracy, I should say,” and explained that his estimate “was a very general conclusion that that was the—that was the time frame” based on a “prediction of the ventilation conditions in the room.” (Emphasis added.)

The petitioner had the burden of proving the elements necessary to prevail on his Brady claim, which required him to demonstrate prejudice by providing reliable burn time evidence in support of his alibi defense. The unfounded, speculative and contradictory nature of the testimony of the petitioner's burn time experts leads me to conclude that the petitioner's evidence is not reliable. Consequently, there is no reasonable probability that it would have been credited.

B

Testimony from Other Experts

Apart from the faulty testimony of the petitioner's own experts, the habeas court also heard testimony from three other fire experts. Their testimony reinforced what the petitioner's experts acknowledged: that investigators lacked the data needed to make a reliable burn time estimate, including proper measures of temperature in the apartment during the fire, ventilation conditions, and the materials in the couch where the fire primarily burned.

As I noted previously, the habeas court heard testimony from Igoe, who investigated the actual fire scene firsthand. On the basis of his analysis of the fire scene and his observations of the couch and its burn rate, he concluded that the primary fire started on the couch and burned “very, very slowly....” Igoe testified that, given the slow burn rate, the time the fire took to spread after ignition could have varied depending on ventilation and insulation conditions in the apartment. He explained, however, that he had difficulty assessing actual conditions at the time of the fire because the scene had been disturbed by emergency personnel. Igoe noted that the “[c]onditions of the [apartment] were horrible” and that the apartment was “a mess” after the fire. On the basis of his firsthand observations, Igoe determined that the slow burning fire could have burned for several minutes to several hours, depending on the exact conditions and materials in the objects that burned. Igoe testified that the only information he would feel comfortable using to determine the fire's duration would be eyewitness testimony about the fire's start time, which did not help narrow the window of time within which the fire could have started.

The habeas court also heard from Joseph Roy, another fire investigator from the state fire marshal's office, who assisted Igoe in his investigation at the scene. Roy's testimony supported Igoe's conclusion. Roy testified that any reliable estimate of the fire's burn time would require additional facts and scientific analysis of the materials involved. That analysis requires a lab analysis of the combustible materials in the home to determine how long those materials could burn, an analysis that was not performed in this case.

The respondent's expert, Corry, also testified that the burn time could not be reliably determined. Corry agreed with Igoe that the fire could have burned for several minutes to several hours, depending on the exact conditions at the time of the fire, which are not known to any of the investigators. Corry reached his conclusions on the basis of his own observations about the insulation conditions of the actual apartment. Of the three hired experts, Corry is the only one who visited the actual fire scene. He noted that the victim's apartment had remained vacant since the incident and still smelled strongly of smoke. He observed the insulation properties of the apartment and interviewed a maintenance employee at the apartment complex who had worked there since before the fire. The maintenance employee explained that the original door had been cleaned and left in place but that the windows had been replaced after the fire with the same type that had previously been installed. From his observations, he concluded that the apartment was well insulated. He also noted that photographs of the scene show limited smoke damage on the outside of the apartment, which led him to conclude that the apartment was not well ventilated at the time of the fire. Corry concluded that the fire's burn time could not be determined to a reasonable degree of probability. Corry believed that the couch burned slowly after ignition. According to Corry, burn patterns indicated that objects on the couch, along with a lack of ventilation, inhibited the spread of the fire. Like the other experts, Corry opined that the fire eventually consumed available oxygen and reduced to a smolder. Given the lack of knowledge of the type of materials involved in the fire and the exact ventilation conditions at the time of the fire, Corry concluded that “[i]t's very difficult to say [with any] degree of certainty” how long the fire lasted. Corry opined that the fire could have smoldered for some time, even with a relatively small amount of ventilation bringing in fresh oxygen to sustain the fire. Because Corry believed that the apartment was well insulated, he believed it could have retained heat and smoke from the fire for a “significant period of time....” For these reasons, Corry testified that the fire could have lasted anywhere from several minutes to several hours. Because of a lack of data needed to render a more specific and scientifically reliable burn time estimate, Corry determined, similar to Igoe, that one would have to rely on eyewitness testimony to determine the time frame within which the fire likely started. In this case, the eyewitness testimony, together with Corry's conclusion that the fire could have burned for several hours, led Corry to conclude the fire could have started anytime during a window of about two hours or more, which the habeas court determined would not support the petitioner's alibi.

This is not a case in which a fact finder, presented with equally compelling but divergent expert testimony, simply chose to credit one expert over others. The petitioner's unreliable burn time evidence and the testimony from other experts convinces me that there is not a reasonable probability that a jury would credit the testimony. How the majority can conclude that the habeas court's findings are “baseless” in light of this evidence is beyond me.

Without any credible burn time evidence to weigh against the original trial evidence, the petitioner cannot establish prejudice under Strickland and Brady. On the record, I disagree that we can conclude, as a matter of law on appeal, that the petitioner has met his burden.

IV

CONCLUSION

For the foregoing reasons, I cannot join in the majority's decision to uphold the Appellate Court's decision to order a new trial for the petitioner. I therefore would reverse the Appellate Court's judgment and remand the case with direction to affirm the habeas court's judgment denying the habeas petition.

Accordingly, I respectfully dissent.

ESPINOSA, J., dissenting. I join Justice Zarella's thorough and well reasoned dissent, which explains why the applicable standard of review and governing principles of law require this court to reverse the judgment of the Appellate Court, which improperly concluded that the petitioner, Richard Lapointe, is entitled to a new trial. There is no need for me to duplicate his compelling analysis. I write separately to emphasize what I view to be the central and troubling flaw in the majority opinion—it constitutes unfettered judicial activism and reflects a complete misunderstanding of the proper role that this court should play within the rule of law. With no reservations whatsoever, the majority usurps the fact-finding role of the trial court in defiance of the constitutional limits on this court's jurisdiction, ignores this court's own recently established guidelines that were intended to mitigate the damage to our system of justice and the prejudice to the parties when this court raises claims sua sponte on appeal, and blatantly serves as an advocate for the petitioner, Richard Lapointe. In other words, in a gross parody of judicial economy, the majority functions as fact finder, counsel and reviewing court. And the majority accomplishes all of this apparently with good intent and in the name of justice.

John Rawls explained the relationship between the rule of law, justice and the legal system: “[T]he conception of formal justice, the regular and impartial administration of public rules, becomes the rule of law when applied to the legal system.” J. Rawls, A Theory of Justice (1971) § 38, p. 235. Within the legal system, therefore, the rule of law is justice. One of the greatest dangers to a just society is presented when one in power acts outside the rule of law in order to vindicate a personal view of what justice requires. Such action should be exceedingly rare, and undertaken only when compelled by necessity. Significantly, immediately after defining the interrelationship between these foundational components of a just society, Rawls provided an example of injustice to illustrate that even seemingly innocuous, well-intentioned departures from impartiality and regularity pose a danger to the rule of law: “One kind of unjust action is the failure of judges and others in authority to apply the appropriate rule or to interpret it correctly. It is more illuminating in this connection to think not of gross violations exemplified by bribery and corruption, or the abuse of the legal system to punish political enemies, but rather of the subtle distortions of prejudice and bias as these effectively discriminate against certain groups in the judicial process.” Id.

Justice is not achieved by suspending the rules in order to benefit a single individual through a judicial decision. Justice is not served when a reviewing court expands its role to include fact-finding, a role properly and constitutionally reserved to the trial court. And justice is most certainly not attained by doffing one's judicial robe and donning an advocate's suit. That, however, is precisely what the majority has accomplished through today's decision. By resolving the appeal on a basis not argued by either of the parties—indeed, on an issue that, as Justice Zarella points out, was expressly abandoned by the petitioner—without allowing them the opportunity to brief the issue, the court flouts the principle that legal rules should be applied in a regular manner. By refusing to defer to the habeas court's predicate credibility findings and its ultimate factual finding that the burn time of the fire could not be determined, the majority acts without jurisdiction and in defiance of this court's constitutional role within the judicial system. By advocating on behalf of the petitioner, the majority appears to abandon any pretense of impartiality. The rule of law has been damaged by today's decision, which casts a cloud over the court, and it is reasonable to wonder if that cloud portends an approaching storm.

Unfortunately, there is nothing in the majority's departure from the rule of law that is even remotely innocuous. As Justice Zarella's dissenting opinion demonstrates in detail, the majority resolves this appeal on a basis not argued by the petitioner, either at the Appellate Court or this court, without allowing the parties the opportunity to brief the issue. The majority itself summarizes the claim raised by the petitioner to the Appellate Court, namely, that “the state's failure to disclose the Ludlow note¹ deprived him of due process of law and that his first habeas counsel had rendered ineffective assistance under Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), for failing to pursue and prove that claim.” (Footnotes altered.) The majority concedes that at the habeas court “the sole issue with respect to the Ludlow note was whether it was material.” The majority further acknowledges that this court has expressly stated that the applicable standard of review is that “a trial court's determination as to materiality under [Brady v.

Maryland, 373 U.S. 83, 87, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963) ] presents a mixed question of law and fact subject to plenary review, with the underlying historical facts subject to review for clear error.” (Internal quotation marks omitted.) State v. Ortiz, 280 Conn. 686, 720, 911 A.2d 1055 (2006). The majority then blithely applies plenary review to the entire materiality determination, including the habeas court's factual findings.

As Justice Zarella points out in his dissent, neither of the parties has raised the issue on which the majority resolves this appeal. Certainly, the respondent, the Commissioner of Correction, does not ask this court to reconsider a well established standard of review that in this instance favors the state. Justice Zarella aptly contrasts the respondent's brief in the present case with the brief of the respondent in Anderson v. Commissioner of Correction, 313 Conn. 360, 375, 98 A.3d 23 (2014), cert. denied sub nom. Anderson v. Semple, ––– U.S. ––––, 135 S.Ct. 1453, 191 L.Ed.2d 403 (2015), in which the state was on notice that this court would be reviewing the question of whether an appellate tribunal properly may revisit the factual findings of the habeas court, because in that case a dissenting judge at the Appellate Court had done precisely that. Accordingly, in the respondent's certified appeal to this court in Anderson, Justice Zarella notes that the respondent fully briefed the issue of the appropriate standard of review. Justice Zarella's implied question to the majority, therefore, is crucial—if the respondent is on notice, as the majority purports, that this court will reconsider the appropriate standard of review, why did the respondent fail to brief the issue?

As to the petitioner, not only has he failed to challenge the factual findings of the habeas court, he has expressly abandoned any claim that those findings constituted clear error by stating that he agrees with the habeas court's ultimate factual findings that “the experts could not determine the exact amount of time the fire burned” and that “the fire's burn time could not be precisely determined.” Additionally, the petitioner has not argued that this court should adopt a de novo standard of review of a habeas court's factual findings and, indeed, the petitioner could not prevail in such an argument. As Justice Zarella explains cogently and thoroughly in his dissenting opinion, it is beyond this court's jurisdiction to review factual findings de novo, and it is beyond this court's power to allow to itself what our constitution forbids. Given the petitioner's acquiescence to the factual findings of the habeas court, one would expect the majority to take those findings as the starting point for its consideration of whether the court properly concluded that the Ludlow note was not material. The habeas court's factual findings, however, are problematic for the majority, because those findings do not favor the petitioner. The habeas court stated that “as the finder of fact” it assigned “far more credit” to the respondent's expert, Robert Corry, than it did to the petitioner's experts, Gerard Kelder, Jr., and John DeHaan. Predicated on that subordinate factual finding, the court ultimately found that “the precise time the fire was set cannot be determined.” (Emphasis added.) The court grounded its legal conclusion that the Ludlow note was not material on its ultimate factual finding that the evidence did not establish that the burn time could be determined with sufficient precision to support the petitioner's proposed alibi defense. Finally, as to the petitioner's proposed alibi defense, the court also found that the petitioner's only alibi witness, his former wife, Karen Martin, without whose testimony the Ludlow note and the testimony of the burn time experts would be irrelevant to the petitioner's Brady claim, lacked credibility.

One might well ask why these findings are problematic for the majority. The answer is quite simple—the majority apparently takes as its starting point the conclusion that the petitioner is innocent.² It is not necessary to engage in any “divination” to discern the impetus driving the majority's decision. My conclusion that the majority begins with the conviction that the petitioner is innocent, and only constructs its analysis after it has arrived at that conclusion, is grounded on three observations. First, as I explain in this dissenting opinion, the majority has not grounded its conclusion on any valid legal principle. If the majority's conclusion is not compelled by the law, and is in fact prohibited by it, it is reasonable to question what has led the majority to arrive at its conclusion. Second, the tinted lens through which the majority views the petitioner's case suggests an answer to that question—that is, the biased language that the majority uses to describe the petitioner's case supports the conclusion that the majority is not viewing the case objectively. Finally, the majority attempts to confine its new rule to the facts of the present case, thus indicating that it has crafted this new rule specifically for this petitioner. For example, the majority emphasizes that the case “presents a highly unusual scenario....” Most tellingly, in response to Justice Zarella's justified concerns regarding the subsequent application of the majority's new rule, the majority states that its reliance on Bunch v. State, 964 N.E.2d 274, 293 (Ind.App.2012), is justified because that case is “factually and procedurally indistinguishable in any material respect from this one.”

The findings of the habeas court are irreconcilable with the majority's apparent conviction that the petitioner is innocent, and, if left standing, those findings would compel the conclusion that the habeas court properly concluded that the Ludlow note was not material. There are occasions when the factual findings of the trial court are determinative—this is one of those occasions. Because the majority is unwilling to acknowledge this limitation on its power, the majority's task is clear—somehow, some way, those findings must fall.

The factual findings of the habeas court create both procedural and substantive hurdles for the majority. First, as a procedural matter, because the parties have not been notified that the appeal will be resolved on a claim raised sua sponte by this court—resolving the appeal by revisiting the factual findings pursuant to a radical new standard of review—the parties must be allowed to submit supplemental briefs. See Blumberg Associates Worldwide, Inc. v. Brown & Brown of Connecticut, Inc., 311 Conn. 123, 161–62, 84 A.3d 840 (2014) (Blumberg ). Second, as a substantive matter, and as Justice Zarella explains, because this court lacks jurisdiction to render judgment on the facts of a case, as a matter of law, the majority cannot revisit the factual findings. Undaunted, the majority nevertheless sets to its task, first making quick work of the habeas court's ultimate factual finding that the burn time of the fire could not be determined with any precision, and the court's finding that Martin lacked credibility, by simply ignoring those findings. This solution is both elegant and ingenious—if a finding is problematic, simply pretend it is not there.

Because the majority cannot claim that the habeas court made no factual findings, it is forced to acknowledge that the court did indeed make credibility “findings.” Not to worry—confident that it is up to the task, the majority dusts off its hands and sets to the messy work of dismantling the habeas court's credibility findings, which it recasts as findings of persuasiveness, rather than credibility. The majority appears to believe that by substituting the word “persuasive” for “credible” it has cleverly recast the habeas court's credibility findings as somehow not really credibility findings. The majority also suggests that determinations of expert witness credibility in particular are not actually credibility findings, because when a trial court evaluates the credibility of an expert, it is “assess[ing]” the foundation on which the opinion is based, not making a finding as to the “personal credibility” of the witness. But see Anderson v. Commissioner of Correction, supra, 313 Conn. at 375, 98 A.3d 23 (in appeal arising from denial of petition alleging ineffective assistance of trial counsel in violation of Strickland v. Washington, supra, 466 U.S. at 686, 104 S.Ct. 2052 habeas court's factual findings as to expert witness' credibility “will not be disturbed unless they are clearly erroneous” [internal quotation marks omitted] ).

The majority's attempt to selectively categorize credibility findings as to expert witnesses as somehow distinct from all other credibility findings is simply not reconcilable with the way that a fact finder evaluates the testimony of witnesses, as it supposes that such a finding may be neatly dissected, somehow separating a fact finder's observations of the demeanor of a witness from its evaluation of the substance of the witness' testimony. The majority's standard is inconsistent with the fact that trial courts routinely instruct juries that they are to consider the credibility of expert witnesses in the same manner that they consider the credibility of any other witness. See State v. Borrelli, 227 Conn. 153, 174, 629 A.2d 1105 (1993) ; Connecticut Criminal Jury Instructions (4th Ed.) instruction 2.5–1, available at http://www.jud.ct.gov/JI/criminal/part2/2.5–1.htm (last visited March 27, 2015).

It is for the fact finder to consider that testimony, using its “ ‘best judgment,’ ” and to determine whether to give any weight to the testimony, and, if so, how much weight to give to it. State v. Borrelli, supra, 227 Conn. at 174, 629 A.2d 1105. A fact finder cannot determine credibility without considering demeanor, and, in the case of juries, they are specifically instructed to consider a witness' demeanor during testimony, regardless of whether that witness is an expert or lay witness. See Connecticut Criminal Jury Instructions (4th Ed.) instruction 2.4–2, available at http:// www.jud.ct.gov/JI/criminal/part2/2.4–2.htm (last visited March 27, 2015). The majority's rule, which requires the habeas judge to say the magic word “demeanor,” in order for this court to apply the standard of review we have applied without exception to a fact finder's credibility findings, is not only inconsistent with the way that our courts have understood the evaluation of the credibility of witnesses, and irreconcilable with the fact finder's discretion to disregard entirely the testimony of any witness, including an expert witness, it is divorced from the reality of what happens at the trial court level. It is irreconcilable with the fact that, pursuant to our criminal jury instructions, the fact finder, whether it be a judge or a jury, “may disregard the [expert's] testimony in whole or in part.” Connecticut Criminal Jury Instructions, supra, instruction at 2.5–1. This oversimplification reflects either an inability to understand or an unwillingness to acknowledge the nuances involved in the fact-finding process. The process of determining credibility is an exercise of human judgment that cannot be accomplished by reviewing a cold piece of paper, but that is the fiction upon which the majority's new standard is based.

In an attempt to blunt the argument of the dissenting justices in this appeal that the parties had no notice that the majority would abandon our established standard of review for credibility findings, the majority turns to the analysis of the Appellate Court, which it claims applied de novo review to the habeas court's credibility findings. There is no support whatsoever for the majority's reading of the Appellate Court decision, which merely concluded that “if” a jury on retrial were to credit both the testimony of the petitioner's burn experts and the testimony of Martin, then, the jury could have found “that it was temporally impossible for the petitioner to have committed the crimes for which he was convicted.” Lapointe v. Commissioner of Correction, 138 Conn.App. 454, 479, 53 A.3d 257 (2012). The holding of the Appellate Court, therefore, is framed in the form of a hypothetical.³ Id. That is, the Appellate Court did not answer the question of whether it is reasonably probable that a jury would credit the petitioner's new evidence. Accordingly, it did not apply any standard of review to the habeas court's factual findings. It simply concluded that if a jury were to credit that evidence, the jury could conclude that the petitioner had successfully established an alibi defense. It is much more useful for the majority, however, to assert, without any support, that the Appellate Court applied de novo review, because that reading supports the majority's refusal to allow the parties to submit supplemental briefs.

The cornerstone of the majority's strategy, on both the procedural and substantive fronts, is its casual announcement that the applicable standard of review of credibility findings—but only as to expert credibility findings, and only in the Brady and Strickland contexts, and only when the habeas court omits to state expressly that it considered the demeanor of the witness—is de novo. As support for this jaw-dropping statement, which is directly contradicted by more than one century of binding precedent of this court, the majority cites to an intermediate appellate court in Indiana, which the majority apparently discovered after scouring every jurisdiction in the country for a decision that would provide support for its astonishing new rule. In connection with a claim based on newly discovered evidence, not Brady material, the Indiana Appellate Court stated in 2012 that it would not defer to a trial court's findings regarding the credibility of experts because that type of credibility determination is not based on a “firsthand evaluation of [the witness'] demeanor” but instead is grounded on an assessment of the credibility of the “foundation” of the expert opinion. Bunch v. State, supra, 964 N.E.2d at 293. The majority opinion does not explain how we arrived in Indiana, or why we are ignoring the binding precedent of this court. See Anderson v. Commissioner of Correction, supra, 313 Conn. at 375, 98 A.3d 23 (expert credibility findings in Strickland context reviewable only for clear error); State v. Lawrence, 282 Conn. 141, 157, 920 A.2d 236 (2007) (“it would be improper for this court to supplant its credibility determinations for those of the fact finder, regardless of whether the fact finder relied on the cold printed record to make those determinations”); State v. Ortiz, supra, 280 Conn. at 720, 911 A.2d 1055 (habeas court's factual findings regarding materiality in Brady claim subject to review only for clear error).

Perhaps most astonishing of all is the majority's failure to mention even once in its 123 page decision a recent Connecticut decision, which, like Bunch, discussed the appropriate level of deference afforded by this court to the credibility determinations of the trial court in considering a claim predicated on newly discovered evidence. See Skakel v. State, 295 Conn. 447, 991 A.2d 414 (2010). Not surprisingly, the majority in Skakel flatly rejected as “unprecedented” and contrary to our law; id., at 487 n. 25, 991 A.2d 414 ; the suggestion of the dissenting justice that this court should not defer to the trial court's credibility findings with respect to the videotaped testimony of a witness, where the trial court had expressly stated that because the witness had not testified at the hearing on the petition for a new trial, that court was “unable to evaluate his ‘demeanor and manner....’ ” (Emphasis omitted.) Id., at 630, 991 A.2d 414 (Palmer, J., dissenting). The majority's dismissal of the dissenting justice's novel theory was not even remotely ambiguous—this court characterized that notion as one that this court “squarely has rejected....” Id., at 487 n. 25, 991 A.2d 414. Yet, today's majority not only fails to distinguish this court's decision in Skakel —it does not even deign that decision worthy of mention, and instead relies on Indiana case law.

Nor does the majority explain how its reclassification of expert credibility findings made within the Brady and Strickland contexts somehow overcomes the jurisdictional bar established by our constitution, or why, for that matter, a different standard of review of credibility findings should apply in Brady and Strickland cases. Instead, the majority acts as though it just discovered an already existing rule dictating the standard of review, a rule that magically allows the majority to find facts without admitting that it is doing so. And the best part is that because this is not, according to the majority, a new standard of review, there is no need to allow the parties to brief the issue of whether the court should adopt a new standard.

With a single decision by an intermediate appellate tribunal in a different state, ⁴ the majority has solved two vexing problems—both the procedural problem of how to resolve the appeal on a claim raised sua sponte by this court without allowing the respondent to brief the issue, and the substantive problem of how to find facts, notwithstanding that doing so requires acting without jurisdiction in violation of our state constitution and ignoring legal precedent that has guided this court for more than one century. It is truly astounding, in light of the overwhelming authority cited by Justice Zarella in his dissenting opinion explaining the jurisdictional limits on this court's authority and documenting the many cases in which we have recognized and held ourselves bound by those limits, and in light of the fact that we recently and expressly rejected in Skakel v. State, supra, 295 Conn. 447, 991 A.2d 414, the very principles on which the Bunch court relied, that the majority has the audacity to deny that we have repeatedly rejected the very standard of review that it now claims is appropriate in the present case.

The answer to the majority's insupportable claim is simple and irrefutable. Our decisions and our state constitution, which clearly prohibit precisely what the majority does today—engaging in de novo review of the habeas court's expert credibility findings—are binding authority. When a decision from another jurisdiction, which constitutes persuasive authority only, is contradicted by our binding precedent, our decisions control. Instead of applying that principle of black letter law, the majority looks to Indiana for permission, finds facts and simply insists that it is not doing so. It applies an astonishing new standard of review and pretends that the rule is well established. The majority raises a new claim sua sponte and expresses surprise when the dissenting justices cry foul.

Under these circumstances, it is unconscionable to fail to order supplemental briefing. The majority clumsily attempts to conceal the fact that it is adopting a radical new standard of review that is beyond the power and contrary to the very essence of this court, in order to justify revisiting the habeas court's factual findings in the absence of a claim of clear error, without allowing the respondent the opportunity to brief the issue. All the while, the majority acts as though there is nothing extraordinary in applying de novo review of factual findings. This sleight of hand is reminiscent of the Wizard of Oz exhorting Dorothy to “[p]ay no attention to that man behind the curtain!” At that point in the movie, no child was fooled, and the majority should not even try to convince itself that the reader will be fooled by its shell game.

The chilling aspect of the majority's brazen maneuver, however, is that we should have seen this coming. This court has been on a discernible path toward precisely this type of abuse of judicial power, and it began down that course by lightly tossing aside the rule of law in a case in which no necessity compelled such extreme action. Until recently, this court considered it to be a “bedrock principle of our adversarial system that courts decide only those claims that the parties have raised.” State v. Lenarz, 301 Conn. 417, 532, 22 A.3d 536 (2011) (Palmer, J., dissenting). Ironically, when Lenarz was decided, Justice Palmer took the majority to task for “[ignoring] this principle in resolving the present case on the basis of a claim that the defendant never has raised and that the state never has had a chance to address.” Id.

Enter Blumberg Associates Worldwide, Inc. v. Brown & Brown of Connecticut, Inc., supra, 311 Conn. 123, 84 A.3d 840. In a stunning and unnecessary departure from this basic principle of appellate procedure, this court announced in Blumberg that pursuant to our supervisory power, we, and the Appellate Court, may set aside that “bedrock principle” and sua sponte resolve an appeal on an issue not raised by the parties. Id., at 155, 84 A.3d 840. Perhaps in recognition of the fact that we were blurring the line between judging and advocacy, we limited our power by claiming that we would exercise it only upon the satisfaction of three conditions. First, there must be “exceptional circumstances” that justify reviewing the new issue.⁵ Id., at 128, 84 A.3d 840. Second, we must give the parties the opportunity to be heard on the issue. Id. Third, there must be “no unfair prejudice to the party against whom the issue is to be decided.” Id. The court specifically emphasized that although all three conditions are necessary before we may sua sponte raise and consider a new claim, “they are not alone sufficient. ” (Emphasis in original.) Id., at 157, 84 A.3d 840. The decision also stated that the reviewing court raising a new claim may not rely on a general “in the interests of justice” explanation; id., at 160, 84 A.3d 840 ; for the departure from the “general rule” that such claims are not reviewable. Instead, the reviewing court “should provide specific reasons, based on the exceptional circumstances of the case, to justify a deviation from the general rule that unpreserved [and unraised] claims will not be reviewed.” Id., at 161, 84 A.3d 840.

Despite this court's suggestion in Blumberg that it was simply making explicit what was already implicit in our decisions, reaction to the decision has ranged from surprised yet polite wariness to open dismay. See, e.g., C. Tait & E. Prescott, Connecticut Appellate Practice and Procedure (4th Ed.2014) § 8–2:5, pp. 453–55 (describing Blumberg as “stunning” decision that “greatly expanded the right of the appellate courts to reach out and decide an issue that was neither preserved by the parties, nor raised on appeal,” and expressing view that reviewing courts should exercise new authority under Blumberg “if at all, on exceedingly rare occasions,” warning that “[o]nly time will tell as to whether Blumberg is the first step down an unwarranted path of issue spotting by appellate courts”); D. Klau, “Two Points Make a Line (And Suggest a Troubling Trend),” Appealingly Brief!, July 1, 2014, available at http://appealinglybrief.com/2014/07/01/two-points-make-a-line-and-suggest-a-troubling-trend (last visited March 30, 2015) (Blumberg and other decisions in which this court has exercised its supervisory authority suggest that this court “has abandoned longstanding norms that have governed the operation of our adversarial system of justice—norms like, ‘[i]t is a bedrock principle of our adversarial system that courts decide only those claims that the parties have raised’—in favor of a philosopher-king model of appellate judging”); C. Ray & M. Weiner, “Mueller v. Tepler, 312 Conn. 631 (2014) : The Appellate Court Gets ‘Blumberged,’ ” Connecticut Lawyer, Vol. 25, No. 3 (October 2014), p. 31 (observing that “as the Blumberg vessel drifts into the murky waters of sua sponte identification and review of legal issues, the more tethered to the strictures set forth in Blumberg that the [c]ourt remains, the more comfortable we, as appellate practitioners, will feel”).

Until today's decision, this court has waded only tentatively into the Blumberg waters. Since that decision was released last year, this court has invoked Blumberg three times, and has generally adhered to the guidelines established therein. See Lane v. Commissioner of Environmental Protection, 314 Conn. 1, 15–16 n. 16, 100 A.3d 384 (2014) (resolving defendant's claim that we could not address question of whether statute retroactively applied to plaintiffs' conduct because it was not encompassed by certified question, on basis that proper construction of statute as applied to plaintiffs necessarily must resolve question of retroactive application, but also noting in dictum our authority under Blumberg ); Mueller v. Tepler, 312 Conn. 631, 643–46, 95 A.3d 1011 (2014) (allowing parties to brief unpreserved claim raised by plaintiff because Appellate Court ruling was reversible under plain error doctrine); State v. Henderson, 312 Conn. 585, 595–96, 94 A.3d 614 (2014) (sua sponte raising and allowing parties to brief issue of whether trial judge made finding that enhanced sentence was in public interest, on basis that claim was alternative ground for affirmance likely to arise on remand). Today, however, the majority reveals that it will no longer remain tethered.

Voltaire is credited with stating that with great power comes great responsibility. As the highest court in the state, it is undeniable that we have great power. Blumberg is rooted in our supervisory authority, which we historically have characterized as a power that should be used sparingly. See, e.g., State v. Medrano, 308 Conn. 604, 648, 65 A.3d 503 (2013) (Norcott, J., concurring) (“[o]ur supervisory powers are invoked only in the rare circumstance where ... traditional protections are inadequate to ensure the fair and just administration of the courts” [internal quotation marks omitted] ). The reason for our restraint is obvious. On the one hand, our supervisory powers serve an essential purpose, reflecting our recognition that, although the rule of law ensures justice within the legal system, there are some instances when justice is more properly aligned with principles of equity. In those rare instances, the uniformity of legal rules must yield to equity, thereby achieving justice.

On the other hand, our extraordinary authority to act outside the limits of the rule of law is unquestionably a “great power,” one that carries with it both great risk and attendant responsibility. Our supervisory authority allows us to reach down and announce a rule or result from on high. As the highest court in the state, once we have invoked that authority, our use of it is virtually unreviewable—with few exceptions, we are answerable only to ourselves. Accordingly, because of the lack of outside checks on that power, we have a duty to resort to that authority only when we must—disciplining ourselves to rely on it rarely. Otherwise, we risk injecting arbitrariness and capriciousness into the rule of law.

Blumberg runs afoul of these basic principles, and, in the present case, the majority takes another step down the dangerous path we have set for ourselves. The court in Blumberg exercised our extraordinary supervisory power arrogantly, without an awareness of its accompanying responsibility. The practical effect is obvious. By literally drawing a road map of all of the instances in which we now have given ourselves license to skirt around bedrock principles of appellate law, Blumberg encourages abuse of our supervisory authority. Today's majority has accepted that invitation with abandon.

On one level, it is not surprising that Blumberg has led us to the abuse of power in the present case, where the majority has raised a new claim and then simply denied that it has done so. The habeas court's findings, if given the proper deference, would not support the result that the majority reaches today, so those findings cannot be allowed to stand. The majority, therefore, whips up a new rule that allows it to make its own findings, and contorts the respondent's routine and cursory assertion that the Appellate Court's decision cannot be reconciled with the applicable standard of review to constitute a request to this court to revisit that standard of review, notwithstanding the fact that we are constitutionally prohibited from doing so.⁶ And the majority does it all without invoking Blumberg even once. As I mentioned earlier, the seeds were present in Blumberg, and we should have seen this coming.

On another level, however, today's decision could never have been foreseen, even by Blumberg 's most vocal critics. In Blumberg, this court revealed that it will dare to exercise its supervisory authority too broadly, too readily and too often. In today's decision, the majority reveals that not even a jurisdictional bar in the state constitution will stand between it and what it views as a just result. No one could have imagined that within one year, Blumberg would seem to be a relatively modest abuse of power.

The majority's unwillingness to abide even by the modest constraints on its power set forth in Blumberg appears to stem from a lack of impartiality, evident from the language of the opinion itself. From the outset, it is clear that the lens through which the majority focuses on the facts of the case is obscured by its apparent bias in favor of the petitioner. I will not undertake a detailed critique of the majority's discussion of the merits of the petitioner's appeal. As I have stated, Justice Zarella already has accomplished that task admirably in his dissenting opinion. I emphasize only some aspects of the majority's discussion in order to highlight the fact that the majority's determination to ignore the rule of law today is driven by the result that it sought from the beginning, handing a get out of jail free card to a man who the majority apparently has become convinced is innocent, despite the fact that the majority was not present at the original trial or at any of the habeas trials. I am mindful of the principle that in dissecting the majority opinion to demonstrate its bias, I am limited, as are we all, by the fact that one can never “[look] into another's mind” to know that person's intent. See Connecticut Criminal Jury Instructions (4th Ed.) instruction 2.3–2, available at http://www.jud.ct.gov/JI/criminal/part2/2.3–2.htm (last visited March 30, 2015). Absent a statement setting forth a person's intent, we always must rely on inference to discern it. I therefore look to what was said, and to what was not said, as evidence of the majority's bias.

I begin with a few examples of the slanted language employed in the majority opinion. Despite never having met the petitioner, the majority states that “he seemed physically, mentally and temperamentally incapable of the brutal crime.” The majority attempts to justify its disregard for the rule of law by referring to “the sinking discomfort that comes with the realization that an injustice may have occurred.” In referring to the Ludlow note, it is not sufficient for the majority to refer to it as merely exculpatory evidence—it labels it “exonerating....” As for the content of the note, which consisted of the cryptic notation “30–40 mins. [p]oss.,” the majority asserts that this vague notation constitutes “details” concerning the length of time that the fire burned in the victim's apartment, and that the note was “consistent” with the testimony of the petitioner's two experts during the habeas trial. The majority also claims that the police “focused their suspicions” on the petitioner, rather than their investigation.

Apparently, deeming the term “confession” to be too negative, the majority characterizes the petitioner's three confessions admitting that he raped and killed the victim and then set fire to her home, as “suspect admissions,” in which he “purported” to take responsibility for the crime. If all of this does not clarify for whom the majority speaks, one need only review part I of the majority opinion, the facts section, which includes as a statement of fact the argument of the petitioner's trial counsel that his confessions “were the product of a highly manipulative interrogation of an extremely vulnerable and impaired man, who had spent his entire life accommodating and agreeing with others in an effort to gain favor and to avoid conflict.” These confessions are the very same ones that this court upheld on appeal. State v. Lapointe, 237 Conn. 694, 730–35, 678 A.2d 942, cert. denied, 519 U.S. 994, 117 S.Ct. 484, 136 L.Ed.2d 378 (1996).

In a striking display of its utter loss of perspective regarding the role of this court and the functioning of our system of justice, the majority makes several statements that suggest it is willing to step beyond even the role of advocate, and take a thirteenth, oversized seat in the jury box. For example, in regard to the petitioner's confessions, the majority states: “Indeed, no fair-minded person who is familiar with the evidence in the present case can read the petitioner's statements and feel confident that they represent a true and accurate account of the victim's murder by the person responsible for her death.” I can think of a group of people who, after months of listening to evidence, did have such familiarity and did have such confidence: The jury of the petitioner's peers. The majority repeatedly ignores the fact that the petitioner had the opportunity to present whatever evidence he chose in support of his defense, to a jury that he agreed would be fair and impartial. In front of that jury, the petitioner had the opportunity to confront the witnesses against him, and to argue his theory of the case through counsel. That jury, unlike the majority, heard all of the evidence and convicted the petitioner.

In another illustration of the same problem, the majority first briefly catalogues some of the evidence relied on by the state at the petitioner's trial, including: upon supposedly discovering that the victim's apartment was on fire, the petitioner took a longer route than necessary to use the neighbor's telephone; when he used the neighbor's telephone, he called his wife and the victim's daughter instead of the police; he claimed that the door to the victim's apartment had been locked, but it was not; he repeatedly asked the police over the years whether he was a suspect in the case; and, he gave conflicting accounts about how often he had left his home and how often he had seen the victim on the day of the murder. Astonishingly, the majority then states: “Suffice it to say that we do not believe that a jury would necessarily find any of this conduct particularly odd or suspicious, even for the average person, and would likely find it much less so for the petitioner,” despite the fact that a jury already has found all of this behavior, taken together with a confession, to be quite a bit more than odd and suspicious. Indeed, the jury found it to be proof of guilt beyond a reasonable doubt.

The majority's brief catalogue of the state's evidence further serves as an illustration of the primary method by which the majority tries to dilute the strength of the state's case against the petitioner in the original trial in order to bolster the materiality of the Ludlow note. Rather than consider all of that evidence together, the majority addresses it piecemeal, as demonstrated by the majority's suggestion that the conduct listed in its brief catalogue should be evaluated without reference to the fact that the petitioner confessed to the crime. The majority uses this device repeatedly in its evaluation to exaggerate the weaknesses and minimize the strengths of the state's case.

In addition to considering the state's evidence in a piecemeal fashion, the majority unabashedly considers information that was not presented at the original trial or to the habeas court on the petitioner's second habeas petition, such as the opinion of Richard Leo, a professor who testified regarding false confessions in support of the petitioner's unsuccessful actual innocence claim in his first habeas trial. Leo's opinion has no bearing on the materiality of the Ludlow note. The majority claims that it is aware that its “focus, of course, is the import of the burn time evidence relative to the strength of the [state's] case,” but, nevertheless, “recite[s]” Leo's statements to underscore its finding that the confessions were unreliable. See footnote 86 of the majority opinion. Related to the majority's improper reliance on Leo's opinion is its perplexing discussion of the research it has reviewed regarding the problem of false confessions, research that was never presented or considered in connection with the second habeas petition.⁷ The majority fails to provide any explanation as to how its research is connected to this appeal, reinforcing the conclusion that the majority is simply convinced that the petitioner's confessions were false, and therefore the majority is authorized to act outside the rule of law. Additionally, by offering up the research that it has itself reviewed, the majority appears to be trying on yet another hat, that of an expert witness.

The majority also relies on certain testimony that the petitioner unsuccessfully had offered in support of his ineffective assistance of counsel claim in his first habeas petition, specifically, the testimony of a Manchester resident that saw someone running from the area of the crime scene at approximately 8 p.m. on the night of the murder. That testimony has no bearing whatsoever on the strength of the state's case, notwithstanding the majority's sly remark that it will not comment on the admissibility of this testimony at a new trial. The first habeas court made a factual finding that the petitioner had failed to establish a link between the unidentified runner and the crime. Presumably, that finding will stand absent a successful claim of clear error. See State v. Arroyo, 284 Conn. 597, 609–10, 935 A.2d 975 (2007) (explaining that proffered evidence of third-party culpability is inadmissible if that “[e]vidence ... would raise only a bare suspicion that a third party, rather than the defendant, committed the charged offense”).

The majority's reliance, in its assessment of the strength of the state's case, on information that was never introduced into evidence at trial is even more astonishing when considered in conjunction with the majority's selective failure, in that same assessment, to give weight to evidence that was actually presented by the state at trial. Most notably, the primary basis on which the majority concludes that the state's case against the petitioner was “tenuous,” is its finding that the petitioner's confessions were false. In making its case, the majority relies heavily on the fact that the petitioner, who has an IQ of 92, suffers from Dandy–Walker syndrome. The petitioner's trial counsel, however, presented ample evidence of his mental disability at trial in an attempt to call into question the voluntariness and reliability of his confessions, including the testimony of a clinical psychologist and a psychiatrist who both testified, over the course of five days during the trial, regarding the petitioner's mental disability. The jury also had the opportunity to consider this evidence in the context of its observations of the petitioner's testimony over the course of three days, and his demeanor during the two months of trial. The jury considered that evidence and concluded nonetheless that the confessions were both voluntary and reliable.

The majority also contends that the testimony offered by the state at trial detailing the petitioner's police interview and resulting confessions “provides significant insight into ... the petitioner's state of mind during questioning.” I agree that the testimony provides insight into the petitioner's state of mind. I observe that the jury obviously drew different inferences than those drawn by the majority, which, once again, reveals its apparent bias through its slanted account of the evidence. Specifically, although the majority acknowledges that the petitioner himself explained to Detective Paul Lombardo of the Manchester Police Department the reasons for his reluctance to provide the details of his crime, the majority substitutes its own theories as to why the petitioner vacillated. The petitioner was very clear about the basis for his resistance to providing details—he did not want people to view him as a “sex fiend” and was concerned that Martin would leave him once she found out what he had done to her grandmother, the victim, Bernice Martin. The logic of the petitioner's thinking is fairly compelling—brutally raping and murdering the eighty-eight year old grandmother of your then wife, then setting a fire to destroy the evidence, is horrific. It is one thing to confess to doing it; it is quite another to state out loud the details of what you did, knowing that whatever you say will become public knowledge.

The petitioner's thinking reveals a level of sophistication that is inconsistent with the majority's portrayal of him as “slow-witted, easily confused, child-like and gullible....” As the habeas court observed following its review of the trial transcripts, a reasonable interpretation of the petitioner's confessions is “that the petitioner intentionally gave the police a mixture of both truthful and misleading information. The petitioner's behavior may be nothing more than manipulation and duplicity.” That court further observed that its “thorough review of the transcripts of the petitioner's testimony, both during the trial and the extensive motion to suppress hearing, reflect an individual who answered questions quite well but nevertheless was often evasive, selective in his recall and bordering on so incredible as to be not believable....” Both the majority and the concurring opinions rely on the fact that some of the details of the petitioner's third, more detailed statement to the police were not corroborated by the evidence at the crime scene. For example, both opinions point to the fact that the petitioner incorrectly described the method of strangulation and the clothing the victim had been wearing on the night of the murder. It is ironic that the majority and the concurring opinions have such high expectations of the petitioner's ability to recall details two years after the crime, considering much of their theory of the case depends on their view that the petitioner suffers from a mental disability so severe that it renders his confession involuntary.

Consistent with the petitioner's statements that he was withholding information because he was concerned with the consequences, he provided his most detailed confession only after he had been assured that Martin had been informed that he had committed the crime, and that she still supported him and would stand by him. In a masterful display of advocacy, however, the majority decides not to accept the petitioner's own explanation as to why he was withholding details and vacillating during the interview. Instead, the majority supplies its own explanation for the petitioner's evasiveness, postulating that his resistance supports the conclusion that the confessions were false.

It is also significant that the majority dismisses out of hand a powerful piece of corroborative evidence relied on by the state at the petitioner's original trial. The majority summarizes that evidence very briefly, acknowledging that the state relied on the fact that, “before any information regarding a possible sexual assault became known to the police or the public, the [petitioner] stated in a conversation with ... a friend of the Lapointe family ... that ‘it was a shame they killed an old lady, but they didn't have to rape her, too.’ ” State v. Lapointe, supra, 237 Conn. at 699, 678 A.2d 942. The majority then opines that the petitioner's testimony at trial offered a perfectly good explanation as to how he came by that information—the petitioner stated that he had overheard this information at the hospital. As elsewhere, the information omitted by the majority reveals its true agenda. This court summarized the relevant facts: “When asked in a June, 1989 interview by [Lombardo] how [the petitioner] had learned that the victim had been sexually assaulted, the [petitioner] responded that he had been informed by a doctor at the hospital on the night of the murder that the victim had been strangled, stabbed and sexually assaulted. The medical personnel who had attended to the victim unanimously testified, however, that they did not check the victim for sexual assault trauma when she was at the hospital that night and, further, that it would have been highly unusual for them to have done so under the circumstances. Other family members who had been present at the hospital corroborated the testimony of the medical personnel who said that there had been no mention of sexual assault at the hospital.” Id., at 699–700, 678 A.2d 942. The majority ignores the weight of the testimony regarding what was revealed at the hospital, and instead highlights the testimony of a single witness that supports its conclusion. This is advocacy, not adjudication.

Perhaps the most deceptive aspect of the majority opinion is its failure to give any weight to the fact that the petitioner did present expert testimony at the original trial regarding the burn time of the fire. The majority strategically mischaracterizes that testimony in considering the petitioner's Brady material in the context of the evidence presented at the original trial. Specifically, at trial, the petitioner

called Christopher Marvin, who had served with the Manchester Fire Department for twenty-two years, the last nine or ten years of which he had held the rank of Deputy Fire Marshal. Although the majority dismisses Marvin by stating that he was “not a professional firefighter much less a forensic fire expert,” Marvin was a certified fire marshal who completed an initial 160 hours of required training, and maintains certification by completing ninety hours of required training every three years. He testified that he had taken arson and fire investigation courses through the state fire marshal's office, the national fire academy and also through college courses. At the time of trial, he was attending college working toward his degree in fire science. Considering Marvin's long years of service as a fire marshal, as well as his extensive training, the majority's reliance on his status as a volunteer is deceptive by misdirecting the reader from what is important, that is, his education and experience as a fire marshal. Marvin was qualified to offer an expert opinion, and he did.

As part of its effort to downplay the significance of Marvin's trial testimony, the majority flatly misrepresents it, suggesting that the petitioner did not call him in order to question him regarding his investigation of the fire, and implying that he was questioned regarding the fire's burn time only during redirect examination. He was questioned extensively regarding his investigation of the fire during direct examination by defense counsel, and he testified specifically regarding his conclusions, based on the evidence that he had observed, regarding the burn time of the fire.

Marvin testified regarding his investigation of the fire, the likely temperatures inside the apartment during the fire, the rate at which the fire burned, and his conclusions about the fire's burn time. Significantly, during direct examination by the petitioner's trial counsel, Marvin testified that he estimated that the fire had burned for approximately fifteen to twenty minutes before it reached its “highest heat buildup” at approximately 8:10 p.m. That time frame corresponds, in all material respects, to the time frame mentioned in the Ludlow note and the estimates given by the petitioner's new burn time experts. During cross-examination, Marvin admitted that his estimate was rough and that the state fire marshal who investigated the fire, Stephen Igoe, was of the opinion that a burn time could not be determined. Although the majority relies on Marvin's characterization of his estimate as a rough one in its attempt to discount his testimony, his admission during cross-examination that the number of variables involved in determining the burn time of a fire makes any estimate a rough one is consistent with the habeas court's factual finding that a burn time could not be estimated with precision—of course, the majority cannot mention this fact, since it is assiduously pretending that the habeas court did not make that factual finding. Marvin's conclusion that it was not possible to offer more than a very rough estimate was shared by the other two fire investigators, Igoe and Joseph Roy, who, he testified, also did not draw conclusions as to any particular temperature range. Marvin explained on redirect examination that the reason that his estimate was very rough was because certain testing that could have been performed at the scene—computer modeling and computer testing, which would have permitted the investigators to determine the exact temperatures of the fire—had not been performed.

The majority also conveniently ignores the fact that Marvin provided a very specific basis for his estimation of the burn time, stating that he observed that there were heat buildup marks on the kitchen clock that had occurred when the clock hands had been at 8:10 p.m., leading him to conclude that the heat had reached maximum temperature at that time. He based his burn time estimate on that specific observation, which he stated provided a “base point” for determining the burn time of the fire.

On redirect examination, the petitioner's trial counsel asked questions to expand on Marvin's qualifications as a fire investigator, asked a number of questions about the scientific basis for his estimates, and also asked about the fire's burn rate and, again, about his burn time estimate. Accordingly, the principal aspects of the fire discussed during Marvin's testimony at the criminal trial (peak temperatures, significance of the smoke damage, materials in the couch, speed at which the fire burned on the couch, and duration of the fire) were essentially the same facts disputed among the new burn time experts almost two decades later at the habeas hearing. The best that can be said, therefore, about the petitioner's Brady claim, is that he contends that had the Ludlow note been disclosed to him, he would have presented different, and in his view, better expert witnesses to testify as to the burn time of the fire, in effect getting a second bite at the expert apple. Presumably, in weighing the effect of the Brady material, the majority should have considered that an expert already testified to exactly what the petitioner now claims he would have an expert testify. The majority's treatment of the petitioner's expert testimony, however, suggests that this argument would be entirely new, provides yet another demonstration of the majority's assumption of the role of advocate on behalf of the petitioner.

Ultimately, the most important question that the majority must answer is this: Why is the Ludlow note sufficient to undermine its confidence in the jury's verdict? No one contends that the note has any independent significance. Indeed, no one actually knows what the note means. And the testimony of the petitioner's experts regarding the burn time of the fire, standing on its own, is irrelevant to the case. It is important to recall that the only reason that the Ludlow note and the testimony of the petitioner's experts are relevant at all is because they lead to the potential alibi testimony of Martin. An assessment of her credibility is crucial to the materiality of the petitioner's Brady claim, which is the only issue in this case. The significance and persuasive force of the petitioner's burn time estimates rise and fall on the strength of Martin as a witness. Unfortunately for the petitioner, the habeas court found Martin to be a terrible witness.

Although the majority claims that the resulting alibi defense that the petitioner would be able to assert at a new trial, predicated on the testimony of the new burn time experts and Martin, would constitute “a complete and potentially compelling alibi,” that claim is not supported by the habeas court's credibility findings as to Martin. Martin testified before the habeas court in this matter, so the habeas judge had the opportunity to evaluate her credibility and the substance of her testimony. The court expressly found that Martin's testimony did not provide the petitioner with “anything that remotely amounts to an alibi.” During the habeas hearing, Martin repeatedly testified that she does not remember details from the day of the murder, she appeared easily confused during the questioning, and she contradicted some of her earlier testimony from the suppression hearing.

At one point during the hearing, Martin indicated that the petitioner may have, in fact, been out of the house walking the dog when she came downstairs after bathing her son. When asked where the petitioner was when she was bathing her son, Martin replied: “I don't know. I didn't—I don't know. I went up to get Sean, my son, ready, and I got him ready for bed. When we came down, he wasn't there, so I think he took the dog. I don't know. I'm—don't want it—I don't actually know if—you know he went out. He probably went—could have—he could have gone for a walk.” When the petitioner's counsel began to ask a follow-up question to this answer, Martin interrupted and said: “Or I don't know where he was.” After showing Martin her prior testimony from the suppression hearing, in which she had testified that the petitioner was at home when she came downstairs, the petitioner's counsel asked her whether the petitioner was at home, and she replied: “If he was there when I came downstairs, I came down probably about the time my aunt called and then he just left and went to my grandmother's. If I—I'm saying, I'm not—you know I'm not 100 percent positive.”

The habeas court found Martin's testimony to be “harmful to any alibi defense premised on [her testimony].... [Martin] testified that she did not know whether [the petitioner] had been in the house (on the night of the murder) while she was upstairs, if the petitioner had been outside she would not have known, and that she had no way of knowing if he had left the house. It is difficult, if not impossible, to conclude that [Martin's] presence on another floor, while giving her son a bath, not seeing or hearing the petitioner downstairs ... provides the petitioner with anything that remotely amounts to an alibi.”

The habeas court's finding regarding Martin's lack of credibility as a witness is devastating to the majority's conclusion that the petitioner has satisfied his burden to prove the materiality of the Ludlow note. It is astonishing that the majority has expended so much time and effort in 123 pages arguing that the petitioner should be able to present the “new” burn time evidence, and has crafted a brand new rule for reviewing determinations of expert credibility, specifically so that the petitioner can present Martin's “compelling” alibi testimony. At the end of the majority's yellow brick road is Martin—who is not, by the way, an expert witness, so the majority's new rule does not apply—and who has been found by the habeas court, as a matter of fact, to lack credibility. Always prepared, however, the majority simply employs its favorite strategy for dealing with problematic findings—it simply pretends that the habeas court did not find that Martin lacked credibility. The majority's failure to address, even once in its lengthy discussion of materiality, the habeas court's crucial finding that Martin lacked credibility, provides further evidence for my conclusion that the majority began not with the applicable legal principles, but with its apparent conviction that the petitioner is innocent.

The majority's transparent advocacy merely illustrates the greater danger presented by today's decision. The majority acts without jurisdiction, flouts precedent of this court from the prior two centuries, sandbags the respondent, alternately ignores inconvenient findings by the habeas court (i.e., the habeas court's finding that the burn time could not be determined with precision, and its finding that Martin was not credible) and fails to accord deference to the remainder of those findings (i.e., the habeas court's expert credibility findings), and brazenly insists that it is acting within the rule of law. And today's decision does not stand in isolation, but is part of an emerging and disturbing pattern in this court's jurisprudence, beginning with Blumberg Associates Worldwide,

Inc. v. Brown & Brown of Connecticut, Inc., supra, 311 Conn. at 123, 84 A.3d 840 and leading us to this point.

Today's decision could be viewed as illustrating precisely the type of “subtle [distortion] of prejudice and bias” of which Rawls warned, thus giving rise to a risk of the appearance of impropriety. J. Rawls, supra, § 38, p. at 235. Prejudice and bias can discriminate against groups not only by departing from the impartial rules to apply a more negative rule to a certain group, but also by a departure from those rules, as in the present case, to apply a more favorable rule to a single individual. While bestowing a benefit that is not required by and contrary to the rules on an individual may appear to be benign, the flaw in that overly simplistic view is demonstrated by comparing the case of the fortunate individual to others who did not benefit from the special treatment.

In the present case, in an extreme departure from the impartial administration of public rules, the majority does not defer to the credibility findings of the habeas court. The majority bobs and weaves its way around the applicable standard of review, obfuscating its distortion of the law in a manner that somewhat conceals the drastic nature of its departure from those rules. Ironically, the Chief Justice, by succinctly stating the new standard in her concurring opinion—namely, that we properly may revisit the findings of the habeas court because: (1) its credibility findings are not “ultimate” factual findings, but rather assessments of whether there is a reasonable probability that a jury could credit the testimony of the witnesses; and (2) because the habeas court did not expressly state that it relied on the demeanor of the witnesses, allowing this court to infer that it relied only on the substance of their testimony—highlights what the majority opinion obscures. Not only is this standard of review new, it is an outlandish distortion of basic principles of appellate procedure and is pretextual in nature. The Chief Justice cites to no authority in support of the rule, for the simple reason that none exists. This is a made up rule that benefits one person—Richard Lapointe.

Although the majority and concurring opinions contort both logic and the law in order to justify their departure from hitherto unquestioned rules of appellate procedure, this court, in a decision affirming the judgment of the Appellate Court, very recently rejected an invitation from a different petitioner to do precisely what it does today. See Sanchez v. Commissioner of Correction, 314 Conn. 585, 602 n. 12, 611 and n. 16, 103 A.3d 954 (2014). As Justice Zarella emphasizes in his dissent, this court in Sanchez properly adhered to the established rule that whether there is a reasonable probability that a new jury would credit new witness testimony presents a question of fact for the habeas court that we review only for clear error. In adhering to this principle, we specifically considered and properly rejected, in a unanimous decision, the reasoning of the dissenting judge at the Appellate Court; id., at 602 n. 12, 103 A.3d 954 ; who had drawn a distinction between a traditional credibility determination, as made by a jury, and the credibility determination made by a habeas court, which he characterized as merely “assessments of the likelihood” that the jury would credit the witnesses. Sanchez v. Commissioner of Correction, 138 Conn.App. 594, 605 n. 2, 53 A.3d 1031 (2012) (Sheldon, J., dissenting).

Our proper refusal to depart from the impartial and regular administration of the applicable legal principles as to Jorge Sanchez, the petitioner in Sanchez, stands in sharp contrast to the special rule that the majority has made up in the present case for Richard Lapointe, whose case has been a cause cèlébre for decades. I emphasize that our decision in Sanchez was proper—it is only when contrasted with the majority's superhuman efforts to benefit the petitioner in the present case that our adherence to the law in Sanchez casts light on the unfairness of the majority's manufactured rule in the present case. Undoubtedly, Jorge Sanchez would have welcomed a court that saw fit to depart from the rules in order to benefit him individually. Instead, he received the impartial administration of the applicable legal rules, which meant that this court and the Appellate Court properly deferred to the habeas court's credibility findings. Indeed, Jorge Sanchez might well ask: “If this is not a new rule, why was it not applied to me?” I do not offer my observations of the majority opinion, or of our prior precedent, lightly. I do so because I owe a duty to this court and to the rule of law. I believe that Blumberg was an ill-conceived exercise of our supervisory authority, and the abuse of power accomplished by the majority's decision today makes Blumberg appear tame by comparison. These decisions jeopardize the faith of the bar and the public in this court as an institution that serves the law, rather than rules over it. This court has been following an alarming path. We have done damage to the rule of law, which we have a duty to protect. At this point, we have a choice: Shall we crown ourselves as philosopher-kings or resume our proper role as servants of the law? One can only ask where the path will lead to next.