WEDEMEYER, P.J.

Andres DelReal appeals from a judgment entered after a jury found him guilty of two counts of second-degree recklessly endangering safety, while armed, contrary to §§ 941.30(2), 939.05, and 939.63, Stats. He also appeals from an order denying his postconviction motion. DelReal claims that the trial court erred in ruling on the admissibility of certain evidence relating to gunshot residue tests. Because we conclude that the State failed to disclose exculpatory evidence relevant to the defense, we reverse and remand for a new trial.

I. BACKGROUND

At 9:00 p.m., on October 18, 1994, Louis Burnette and Michael Surprise were walking in an alley near South 10th Street in the City of Milwaukee. Both stated that a white car approached and stopped directly in front of them. They stated that a passenger in the car then aimed a handgun in their direction and fired several shots at them. Both victims later identified DelReal as the shooter and stated that DelReal was located in the back seat behind the driver at the time he fired the gun. Another eyewitness, Alicia Moreno, identified the car that had been involved in the shooting and also identified DelReal, but she indicated that he was in the front passenger seat at the time the car drove through the alley.

DelReal and three other people were in a car that was stopped about one hour and twenty minutes after the shooting. DelReal was subsequently arrested and taken to the police station where his hands were swabbed for gunshot residue. The gun was never found. DelReal pleaded not guilty and his case was tried to a jury. DelReal contends he was not the shooter.

The State's case consisted of the testimony of the victims and Moreno, together with the testimony of Police Detectives Octavio Delgado and Moises Gomez and other police officers involved in investigating the case. The issue in this appeal centers on the testimony of Detectives Delgado and Gomez. Gomez was the detective in charge of investigating the crime scene. Delgado was in charge of investigating the hospital scene and conducted the two line-ups that took place at the police station the day after the shooting. During cross-examination, Delgado testified that DelReal's hands had been swabbed for gunshot residue, but that he did not know what the results revealed.

Following this testimony, the State recalled Gomez, who testified that swabbing had not been done and that he would have known if, in fact, swabbing was done because he was the detective in charge of the scene. Subsequently, the trial court struck the testimony regarding the swabbing and gunshot residue, ruling that it was irrelevant. The jury convicted DelReal.

DelReal filed a postconviction motion seeking a new trial on the grounds that the State failed to disclose exculpatory evidence, and that he was thereby denied his rights to confrontation and to present a defense. Postconviction investigation revealed that DelReal's hands, in fact, had been swabbed for gunshot residue, but that the State never performed the tests to determine the results. At DelReal's request, the tests were performed postconviction and the results were "negative." A negative test result means there is not sufficient gunshot residue on the swab to result in a positive finding.

At the postconviction hearing, Police Officer Mark Lelinski testified that Detective Gomez instructed him to take the suspects "down and have their swabs taken." Janell Eccher, a Bureau of Identification technician, testified that she performed the swabbing of the suspects and that, although she has no independent recollection, the documents she filled out indicated that Detective Gomez had requested the swabbing. Despite Lelinski's and Eccher's testimony, and despite the documents reflecting that Detective Gomez ordered the swabbing, Gomez testified that he did not order these tests.

At the postconviction hearing, Raymond Lentz, an expert who performs gunshot residue tests, testified regarding what such test results reveal. He indicated that if the test is positive it means the subject either: (1) discharged a firearm; (2) handled a recently discharged firearm; or (3) was in close proximity to a firearm that was being discharged. If the test is negative, it is inconclusive because the shooter could have wiped off the gunpowder or, because of the type of gun and ammunition used, no residue may have been emitted onto the shooter's hands. The trial court denied DelReal's postconviction motion. He now appeals.

II. DISCUSSION

Whether to admit or exclude evidence is a discretionary decision that will not be reversed on appeal unless the decision constitutes an erroneous exercise of discretion. See State v. Morgan , 195 Wis.2d 388, 416, 536 N.W.2d 425, 435 (Ct.App. 1995). A trial court does not erroneously exercise its discretion if it considers the pertinent facts, applies the correct law and reaches a reasonable decision. See State v. Evans , 187 Wis.2d 66, 77, 522 N.W.2d 554, 557 (Ct.App. 1994). Further, the defense has a constitutional right to material exculpatory evidence in the hands of the prosecutor. See Brady v. Maryland , 373 U.S. 83, 86 (1963). Exculpatory evidence is material if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different. A reasonable probability is a probability sufficient to undermine confidence in the outcome. See United States v. Bagley , 473 U.S. 667, 682 (1985); State v. Garrity , 161 Wis.2d 842, 847-48, 469 N.W.2d 219, 221 (Ct.App. 1991). This court independently applies the Bagley constitutional standard to the undisputed facts of the case. See State v. Woods , 117 Wis.2d 701, 715-16, 345 N.W.2d 457, 465 (1984). Impeachment evidence casting doubt on a witness's credibility is material and subject to disclosure. See State v. Nerison , 136 Wis.2d 37, 54, 401 N.W.2d 1, 8 (1987).

We need not address whether the trial court's decision during the trial to strike the gunshot residue testimony was erroneous because we conclude that the additional evidence produced during the post-trial proceedings demonstrated that the State failed to turn over evidence of an exculpatory nature; i.e., the State failed to disclose that swabbing had in fact been performed, which would have provided DelReal the opportunity to have the swabs tested leading to a negative test result. This was relevant, exculpatory evidence because the negative test result would have some weight and its tendency could have supplied a favorable inference of DelReal's innocence to the jury.

Evidence is relevant if it has "any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence." Section 904.01, Stats. As noted, Detective Delgado testified that swabbing was done, but he had not seen the results. The State recalled Detective Gomez to rebut Delgado's testimony. As noted, he denied that swabbing was ordered or performed. Further, on cross-examination, Gomez testified that the gunshot residue test can show whether or not someone fired a gun. The trial court proceeded to ask Gomez some questions regarding swabbing and gunshot residue tests. Gomez testified that he had taken some classes regarding evaluating the results of this type of evidence. He further testified that these test results can show whether someone did not fire a gun, explaining that if a person had fired a gun, the tests would show gunpowder residue on the hands and if the person had not fired a gun, there would be no gunshot residue. Thereafter, outside the presence of the jury, the trial court expressed its concern about the accuracy of this testimony and explained that, based on its own experience, these tests cannot prove that the defendant did not fire a gun and cannot help the defendant in any way. The trial court analogized the gunshot residue tests to fingerprint evidence; that is, the presence of a fingerprint (or gunshot residue) is proof that someone touched something (or fired a gun), but the absence of a fingerprint (or the absence of gunshot residue) does not prove that the person did not touch something (or did not fire a gun).

When DelReal's counsel asked the trial court what, if anything, it planned to do as a result of its analysis, the trial court simply stated that it just wanted to explain why it asked questions of the witnesses about the gunshot residue tests. Thereafter, the trial court again raised the issue about the gunshot residue testing, suggesting that the State should bring in someone from the crime lab to testify regarding what these tests can show. DelReal's counsel suggested that the trial court should let the State decide how to present its case. The State indicated that it was not sure whether or when it could get a witness on this issue. The trial court then asked why this entire area of testimony should not be stricken. DelReal's counsel responded that the testimony was relevant and exculpatory because it may tend to show that DelReal was not the shooter and that one of the other individuals in the car actually was the triggerman. DelReal's counsel also argued that the evidence was relevant to challenge the credibility of the police and the investigation. The trial court indicated that this evidence was irrelevant and decided to strike the testimony. The decision was based in part on the trial court's belief that no swabbing was, in fact, done. The trial court ruled: "What is the evidentiary significance of a decision in this case not to do swabs of the defendant or anyone else? . . . [T]here is no evidentiary significance to that."

The postconviction proceedings clearly established the relevancy of the undisclosed evidence. The postconviction testimony revealed that Gomez was wrong. In fact, swabbing was performed, although the swabs were never tested for gunshot residue. This fact is relevant to attacking both Gomez's credibility and the quality of the police investigation. See Kyles v. Whitley , 514 U.S. 419, 444 (1995). The postconviction hearings also revealed that the gunshot residue tests of DelReal's hands were negative. That is, there were insufficient amounts of chemical elements present to yield a positive result. The defense certainly could argue from this result that the test reduces the probability that DelReal fired the gun. Similar to the fingerprint analogy, the test cannot conclusively prove that DelReal was not the shooter because he may have taken some action to eliminate any positive evidence, such as washing his hands to remove any residue, just as a defendant may take action to ensure his fingerprints do not remain at a scene by wearing gloves or wiping the surface clean. This, however, does not make the test or its results irrelevant or inadmissible. Rather, these factors are arguments with respect to the weight of the evidence. The negative evidence may not disprove a defendant's guilt, but it certainly has a "tendency" to make it "less probable." Section 904.01, Stats.

The United States Supreme Court in Kyles reached a similar conclusion on facts analogous to those presented here. In Kyles , the State failed to disclose to the defense a list of license plates that the police had gathered from a crime scene where the victim was shot and killed and the killer stole the victim's car. Id. at 450. The police theorized that, prior to fleeing with the victim's car, the suspect, Kyles, drove to the parking lot and left his car at the scene during the heat of the investigation. See id. However, Kyles's car's license plate was not contained on the police list. See id. The State argued that the list provided neither impeachment nor exculpatory evidence because Kyles could have moved his car before the police compiled the list or because the police list may have been incomplete. See id. at 450-51. The Supreme Court rejected this argument, ruling that the State's claim "confuses the weight of the evidence with its favorable tendency." Id. at 451. The Supreme Court ruled that the list, with Kyles's license plate absent, "would have had some weight and its tendency could have been favorable to Kyles." Id.

The same analysis applies in the instant case. The fact that swabbing was done, and/or the fact that the test results were negative, has some weight and its tendency could have been favorable to DelReal. As further explained in Kyles , the State's failure to disclose certain evidence to the defense prevented the defense from using the evidence to "throw the reliability of the investigation into doubt and to sully the credibility of" the lead detective in the case. Id. at 447. The same holds true here. In the interest of securing a fair trial, DelReal was entitled to challenge the reliability of the police investigation and to challenge the credibility of Gomez. If the State had disclosed that swabbing was done, but not tested, the defense certainly could have challenged the failure to perform the tests, as it has done on appeal.

Gomez's credibility was also fair game for defense attack. Gomez testified falsely at trial, either because he was mistaken or because he chose not to be truthful. The State used Gomez to attack the credibility of eyewitness Moreno, whose testimony supported DelReal's claim that he was not the shooter.

Other jurisdictions that have addressed the issue of whether gunshot residue test results are admissible have decided in favor of allowing the evidence despite the inconclusive nature of the results, ruling that the tests are relevant to show the probability that the defendant did or did not fire a gun or to show that the test results are consistent or inconsistent with the subject firing a gun. See United States v. Barton , 731 F.2d 669, 672 (10th Cir. 1984); Isbell v. State , 931 S.W.2d 74, 78 (Ark. 1996); Mills v. State , 476 So.2d 172, 176-77 (Fla. 1985); People v. Cole , 524 N.E.2d 926, 935-36 (Ill.Ct.App. 1988); State v. Velez , 588 So.2d 116, 133 (La.Ct.App. 1991); State v. Stephan , 941 S.W.2d 669, 674 (Mo.Ct.App. 1997); State v. Chatman , 383 A.2d 440, 442 (N.J.Super.Ct. App. Div. 197 8). The inconclusive nature of this evidence does not render it inadmissible, but rather, affects its probative value, which is for the jury to determine.

Based on the foregoing, we conclude that the State failed to disclose relevant exculpatory evidence. In the factual context of this case, the evidence was relevant for impeachment purposes, for challenging the police investigation, and for arguing the defense theory that DelReal was not the shooter.

The State argues that any error was harmless. We cannot agree. An error is harmless only if there is no reasonable possibility that the error contributed to the result in the case. See Dyess , 124 Wis.2d at 543, 370 N.W.2d at 231-32. Under the facts presented in the record, particularly during the post-trial proceedings, we cannot affirm this case under a harmless error analysis. The State's case against DelReal was by no means airtight. The case rested primarily on the testimony of the two victims. The shooting occurred at night in an alley. There was a dispute as to the quality of the lighting. Gomez testified it was "well-lit enough" to be able to see where you were going. Burnette testified there was one light pole and Surprise testified that the alley was lit only from outside house porch lights. Moreover, initial identifications by both witnesses were equivocal. Burnette initially stated that he thought DelReal looked "similar" to the shooter and Surprise said he "believed" DelReal was the shooter. Surprise also indicated he only saw the car for a second. The victims did not have a long time to examine what the shooter looked like as they were fleeing from gunfire. Thus, the two eyewitness identifications were not beyond challenge. Further, the third eyewitness, Moreno, placed DelReal in the front passenger seat, whereas the shooter was undisputedly in the back seat behind the driver. The gun was never found and the gunshot residue tests were not consistent with DelReal having discharged a firearm. The evidence presented here is not so overwhelming that the State's failure to disclose this relevant potentially exculpatory evidence was harmless.

III. CONCLUSION

In sum, we reverse and remand this case for a new trial.

By the Court. — Judgment and order reversed and cause remanded with directions.