EASTERBROOK, Circuit Judge.

Medtest Corporation has a single asset: the intellectual property in a self-administered cervico-vaginal cytology testing process. Patrick Manning devised the process and together with Donald West, a lawyer, formed Medtest in December 1981 to obtain a patent and undertake development to make the process commercially attractive. Manning held 31% of the stock, West 26%, and the remainder was scattered among friends and relatives. In 1982 Manning sold some of his stock in Medtest to Robert and Anna Lisa Pommer: 250 shares in September for $25,000, and later another 2,750 shares for $175,000. The Pommers thus acquired 3% of Medtest's outstanding stock, which is valuable only to the extent the firm pays dividends, goes public, or is acquired by a third party.

None of these things has happened, and the Pommers believe that they are the victims of fraud. A jury agreed in this action under § 10(b) of the Securities Exchange Act of 1934, 15 U.S.C. § 78j(b), and the SEC's Rule 10b-5, 17 C.F.R. § 240.10b-5. It awarded the Pommers more than $300,000 in damages, representing the purchase price of the stock plus interest. A magistrate judge, presiding by consent under 28 U.S.C. § 636(c), set aside the verdict and entered judgment for the defendants. She concluded that none of the representations made to the Pommers was materially false.

I

Given the verdict, we must take all of the evidence in the light most favorable to the Pommers. A jury could have concluded that West told the Pommers, while they were negotiating to buy the stock, that Medtest had a U.S. patent on the process and that a sale of Medtest to Abbott Laboratories, at a price between $50 million and $100 million, was imminent ("almost a finished deal"). A 3% interest in Medtest would have been worth between $1.5 million and $3 million had such a sale been consummated. In fact Medtest did not have a patent at the time. Counsel informed West in December 1981 that the process was patentable; the firm filed an application on September 30, 1982, and the patent issued on August 14, 1984. Medtest was not in the last stage of negotiation with Abbott Laboratories; it raised the subject with Abbott, and Abbott's employees mentioned a price in the $50 to $100 million range, but no details had been discussed, no hands had been shaken — and on October 28, 1982, Abbott sent Medtest a letter stating that it was not interested in acquiring Medtest. Since then Medtest has been developing the process on its own.

The magistrate judge concluded that the representations about the existence of a patent were not material — and therefore did not support relief under the securities laws — because counsel had told West that the process was patentable, and Medtest obtained a patent in due course. The judge continued: "It is true that Pommer testified that he would not have purchased his shares had he known the patent had not actually issued, but he never explained this self-serving conclusion or suggested any reason why it would make a difference to him." As for the sale to Abbott, the judge wrote: "Pommer's own testimony demonstrated that he knew that the time of sale and its terms were indefinite. . . . . . . [A]ny person of Mr. Pommer's intelligence and sophistication [knows that nothing under negotiation is] absolutely certain. . . . Negotiations involving a price range of between 50 and 100 million dollars indisputably indicate that a lot remains to be negotiated and decided."

Considerations of this kind show that a verdict in defendants' favor could not be disturbed. They do not, however, show that no reasonable juror could think the statements material. A statement is material when there is "a substantial likelihood that the disclosure of the omitted fact would have been viewed by the reasonable investor as having significantly altered the `total mix' of information made available." TSC Industries, Inc. v. Northway, Inc., 426 U.S. 438, 449, 96 S.Ct. 2126, 2132, 48 L.Ed.2d 757 (1976), applied to § 10(b) actions by Basic Inc. v. Levinson, 485 U.S. 224, 231-32, 108 S.Ct. 978, 983, 99 L.Ed.2d 194 (1988). West told the Pommers that Medtest had a patent, doubtless recognizing that in selling stock as in other endeavors a bird in the hand is worth two in the bush. Counsel's belief that the process is patentable is a fair distance from a patent. The examiner may disagree or insist that the applicant limit the claims in a way that affects the commercial value of the invention. If we take counsel's belief as signifying an 80% chance that Medtest will obtain a patent on the central claims, still the difference is material. Even a small probability of a bad event may be material, if that event is grave enough. And it does not matter that Medtest obtained the patent two years later. The securities laws approach matters from an ex ante perspective: just as a statement true when made does not become fraudulent because things unexpectedly go wrong, so a statement materially false when made does not become acceptable because it happens to come true. Wielgos v. Commonwealth Edison Co., 892 F.2d 509, 518 (7th Cir. 1989); Jordan v. Duff Phelps, Inc., 815 F.2d 429, 440-41 (7th Cir. 1987); Escott v. BarChris Construction Corp., 283 F. Supp. 643, 669-70, 675 (S.D.N.Y. 1968). Good fortune may affect damages, but it does not make the falsehood any the less material.

So too with the negotiations to sell Medtest to Abbott Laboratories. Until the last minute a deal may collapse, but some deals are more likely to close than others. Probabilities determine the value of stock. At a 90% chance of a buyout for $50 million, the Pommers' stock was worth $1.35 million; at a 10% chance it was worth $150,000. A jury could determine that West conveyed to the Pommers a substantially higher probability than the facts supported — that although West represented that the parties were just about to sign on the dotted line, actually there had been no more than superficial discussions, and Abbott had no serious interest in acquiring Medtest. There is "a substantial likelihood" that the Pommers (and any other "reasonable investor") would have viewed the truth about the negotiations with Abbott as "significantly alter[ing] the `total mix' of information made available."

Defendants insist that until the parties agree on the price and structure of a transaction, the issuer may say what it pleases. For this proposition they cite Flamm v. Eberstadt, 814 F.2d 1169 (7th Cir. 1987), overlooking three important ("material") facts about Flamm. First, we discussed whether an issuer has a duty to speak even though it would prefer silence; Medtest was not silent. 814 F.2d at 1174-79. Second, the Supreme Court disapproved the price-and-structure approach when issuers make false statements about ongoing negotiations. Basic, 485 U.S. at 232-36, 108 S.Ct. at 983-86. Third, contemporaneously with Flamm we decided Jordan, which holds that the price-and-structure rule does not apply to closely held corporations such as Medtest. 815 F.2d at 434-35. Jordan observed that in aftermarket transactions duties to disclose come from state law, and states regularly apply a special rule requiring disclosure when a manager of a closely held firm sells stock to (or buys it from) an outsider in a face-to-face transaction. Defendants do not address the effect of Jordan, or for that matter Basic; they cite neither case.

One could recast the magistrate judge's conclusion about the sale to Abbott as a finding that the Pommers knew enough of the truth that West's lies did not significantly affect their appreciation of the situation. An issuer that utters a mixture of truth and falsehood may have furnished the information necessary for an accurate appreciation of the securities' value — especially when the truth is written and the lie is oral. Teamsters Local 282 Pension Trust Fund v. Angelos, 762 F.2d 522, 530 (7th Cir. 1985). "While a misleading statement will not always lose its deceptive edge simply by joinder with others that are true, the true statements may discredit the other one so obviously that the risk of real deception drops to nil." Virginia Bankshares, Inc. v. Sandberg, ___ U.S. ___, 111 S.Ct. 2749, 2760, 115 L.Ed.2d 929 (1991). Written disclosures of the truth are not only more compelling but also more verifiable. Memories may play tricks (or the parties may be less than candid about what they heard or said); writings are available for inspection. In shaping implied private rights of action, courts properly favor rules that simplify litigation, discourage unverifiable (and therefore irrefutable) claims, and reward issuers that commit their representations to writing. Cf. Blue Chip Stamps v. Manor Drug Stores, 421 U.S. 723, 95 S.Ct. 1917, 44 L.Ed.2d 539 (1975).

The Pommers must have been aware that a deal with Abbott was not just around the corner. A substantial range of price implied that hard negotiations lay ahead. West told them that Medtest needed to secure foreign patents, and that with these in hand it would be an attractive acquisition candidate. As the magistrate judge remarked, foreign patents cannot be obtained in a flash. The Pommers waited more than a year before beefing; would they have sat quietly so long if they really thought they had been promised a bonanza in a few weeks? The price of the securities also implies that a sale was a long shot: Why would Manning sell 3% of Medtest to the Pommers for $200,000 if the stock soon would be worth $1.5 million and up? Even if Manning desperately needed to raise capital, better terms must have been available. Robert Pommers said at one point that he viewed the investment in Medtest as his chance to hit the Lotto — yet he persuaded the jury that the defendants sold him a sure thing.

Whether these things would be enough to neutralize the representations about Abbott we need not say. The falsehood about the patent remains. More, we have emphasized that an issuer needs to disclose the truth clearly before a lie becomes immaterial. Acme Propane, Inc. v. Tenexco, Inc., 844 F.2d 1317 (7th Cir. 1988). It is not enough that the other party must have recognized a risk. Risks are ubiquitous. Disclosures assist investors in determining the magnitude of risks. Even savvy investors may recover when a bald lie understates the gravity of a known risk. Astor Chauffeured Limousine Co. v. Runnfeldt Investment Corp., 910 F.2d 1540 (7th Cir. 1990). A jury could find that defendants never told the truth. Although they direct us to an agreement between Manning and the Pommers, that two-page document contains no information beyond a boilerplate warning that Medtest's process "has only speculative value at the present time and may prove to be totally worthless, in which event the Shares purchased by Buyer pursuant hereto may also become worthless." Such generic warnings do not enlighten investors about the status of patent applications, negotiations to sell the business, and the like; they do nothing to disabuse an investor influenced by false oral statements purporting to describe the status of the firm's affairs. The truth, or an express disclaimer of reliance on oral representations, see Jackvony v. RIHT Financial Corp., 873 F.2d 411 (1st Cir. 1989), protects a seller, but this record discloses neither.

II

Although defendants are not entitled to judgment notwithstanding the verdict, they sought a new trial in the alternative. The magistrate judge accepted their principal argument but did not formally grant the motion, deeming the subject moot. The Pommers ask us to disapprove the judge's conclusion, and thus avert any possibility of a new trial. Whether to grant a new trial is a question for the trial judge in the first instance, with deferential appellate review. E.g., EEOC v. Century Broadcasting Corp., 957 F.2d 1446, 1460 (7th Cir. 1992) (collecting cases). The Pommers therefore must offer a compelling reason before we would forbid the magistrate judge from deciding on remand that a new trial would best serve the interests of justice.

The Pommers put before the jury Abbott's letter of October 28, 1982, disclaiming interest in Medtest. They contended that defendants' failure to show them this letter, while they were still paying for the stock, is an independent securities fraud. Defendants contended that the use of the letter should be restricted. The magistrate judge told the lawyers to argue their versions to the jury but after trial wrote that she should have kept the letter out of evidence, because the Pommers agreed to buy the stock before Abbott sent the letter. In contending that the judge's second thoughts are wrong, the Pommers again have an uphill contest, for our review of evidentiary matters is deferential. See Taylor v. National R.R. Passenger Corp., 920 F.2d 1372, 1375 (7th Cir. 1990).

The truth (or falsity) of defendants' statements, and their materiality, must be assessed at the time the statements are made, and not in the light of hindsight. We made exactly this point about the patent issued in 1984. The letter arrived after the Pommers agreed to purchase the stock. It was on this ground that the magistrate judge concluded that she should have restrained the Pommers' use of the document. Nonetheless, the Pommers contend, the judge was right the first time, because, if they had known that Abbott had broken off negotiations, they would have stopped paying.

If the Pommers were saying that they did not sign a contract to purchase the stock until after October 28, 1982, they would have a point. The statute of frauds is enforced rigorously in securities cases. UCC § 8-319; Kagan v. Edison Brothers Stores, Inc., 907 F.2d 690 (7th Cir. 1990). If not bound by contract to pay for all 3,000 shares, the Pommers were making a fresh investment decision every time they sent a check. But the magistrate judge found that on October 19, 1982, they signed a note promising to pay $200,000 for the stock, and the Pommers do not rely on the statute of frauds. Once they signed that note, the Pommers were committed under the law of contracts. There was no remaining investment decision. Roberts v. Peat, Marwick, Mitchell Co., 857 F.2d 646, 650-51 (9th Cir. 1988); Hill v. Equitable Bank, 655 F. Supp. 631, 638-39 (D.Del. 1987), affirmed on other grounds, 851 F.2d 691 (3d Cir. 1988). They might choose to balk and litigate, of course, but "[w]e have held repeatedly in recent years that the securities laws do not ensure that people will receive information sufficient to make correct decisions about filing or pursuing lawsuits. E.g., Harris Trust [ Savings Bank v. Ellis, 810 F.2d 700 (7th Cir. 1987)], 810 F.2d at 704; Ray v. Karris, 780 F.2d 636, 641-43 (7th Cir. 1985); Metlyn Realty Corp. v. Esmark, Inc., 763 F.2d 826, 833-34 (7th Cir. 1985); O'Brien v. Continental Illinois National Bank, 593 F.2d 54, 60-61 (7th Cir. 1979)". LHLC Corp. v. Cluett, Peabody Co., 842 F.2d 928, 932 (7th Cir. 1988). Cf. Virginia Bankshares, 111 S.Ct. at 2765-66.

According to the Pommers, Goodman v. Epstein, 582 F.2d 388, 409-14 (7th Cir. 1978), required defendants to show them the letter of October 28. Goodman held that a partnership making a call for further capital must reveal current, material information because each new contribution is a fresh investment decision. As the ninth circuit properly observed in Roberts, however, Goodman depends on the fact that the partnership agreement did not compel the investors to comply with each request for funds. They had to decide, and to decide needed accurate information. The Pommers were not deciding anew to buy Medtest stock every time they wrote a check. They were deciding, instead, whether to keep or break their promise. They may have reasons to do either, but the nature of the asset covered by the contract — 3,000 shares of stock rather than, say, 3,000 crates of rutabagas — no longer matters. Once the written commitment is firm, we held in LHLC, the subject passes into the domain of contract law. The magistrate judge therefore acted within her discretion in concluding (on second thought) that the letter of October 28, reflecting an adverse turn of events after the sale of stock, should not have been used as substantive evidence. The judge may elect in her discretion to award the defendants a new trial on remand.

To say that withholding the letter from the Pommers is not an independent fraud is not (necessarily) to say that the letter is inadmissible. Although the parties stipulated that Medtest and Abbott conducted some negotiations, the letter may support an inference that these never became serious. Such an inference would assist the Pommers in demonstrating that the defendants acted with intent to defraud. Whether the letter may be used with appropriate cautions to the jury is a subject for the trial judge in the first instance.

III

Several other issues may come up on remand, and we comment briefly on them to speed this litigation along.

1. Manning sold the stock, yet Medtest and West are the defendants. West may be liable for his own misstatements, if made "in connection with" Manning's sale of stock. In addition to the "connection" language of both statute and rule, see Blue Chip Stamps. Whether West knew that Manning was dealing with the Pommers is an important issue that the trier of fact should resolve.

2. If West was ignorant of Manning's dealings or stood to gain nothing from them, then his own statements to the Pommers look more like boasting over the coffee table (they were social friends). Scienter, another element of liability under § 10(b) and Rule 10b-5, would be hard to establish. See Ernst Ernst v. Hochfelder, 425 U.S. 185, 96 S.Ct. 1375, 47 L.Ed.2d 668 (1976). The parties have not discussed this question, although parts of the magistrate judge's opinion imply that she doubted whether West acted with intent to defraud. The subject remains open on remand.

3. Vicarious liability remains a possibility. But the chain — Manning's statements attributed to Medtest, and then back to West — is weak at both links. Section 20(a) of the '34 Act, 15 U.S.C. § 78t(a), extends liability to "[e]very person who . . . controls any person liable under any provision of this chapter . . . unless the controlling person acted in good faith and did not directly or indirectly induce the act or acts constituting the violation or cause of action." Manning and West both exercise control over Medtest and so might be liable, if Medtest is liable.

But what makes Medtest liable? Section 20(a) does not establish the liability of controll ed entities. It takes common law vicarious liability to reach Medtest, and such liability would depend on a finding that Manning or West acted as Medtest's agents, and within the scope of their authority. Manning sold for his own account, not for Medtest's. Investors, acting as investors, are not agents of the corporation. The Pommers do not contend that Manning bought the stock with a "view to a distribution" and so was Medtest's underwriter. Under § 2(11) of the Securities Act of 1933, 15 U.S.C. § 77b(11), a control person is an "issuer" for the purpose of defining an underwriter, but that does not help — it would make the Pommers the underwriters! At all events, the Pommers have not argued that Medtest should have registered the issue of stock, perhaps because under § 13 of the '33 Act, 15 U.S.C. § 77m, they sued too late. The magistrate judge thought it sufficient that "West and Manning had apparent authority to act on behalf of Medtest and to sell stock", but she did not mention the fact that Manning was not selling "on behalf of" Medtest. He sold his own shares, a fact conspicuously disclosed to the Pommers. The subject bears a fresh look on remand.

So does the imputation of any of Manning's wrongs to West via § 20(a). West is a control person, all right, but if he "did not directly or indirectly induce the act or acts constituting the violation or cause of action" he is not liable under that section. Schlifke v. Seafirst Corp., 866 F.2d 935 (7th Cir. 1989), holds that the plaintiff must show that the "control person" had or exercised control over the specific transaction complained of. It may be that West exercised control over Manning's sale to the Pommers, but the jury was not asked to make such a determination. The scope (if any) of vicarious liability therefore remains open to inquiry.

4. This suit is problematic under the federal statute of limitations that now governs actions under § 10(b) and Rule 10b-5. See Lampf, Pleva, Lipkind, Prupis Petigrow v. Gilbertson, ___ U.S. ___, 111 S.Ct. 2773, 115 L.Ed.2d 321 (1991); Short v. Belleville Shoe Manufacturing Co., 908 F.2d 1385 (7th Cir. 1990). The Pommers bought their stock in late 1982 and did not file suit until 1985, substantially more than a year after receiving a document in late 1983 that, the Pommers say, demonstrated that what they had been told in 1982 is untrue. Section 476 of the FDIC Improvement Act of 1991, Pub.L. 102-242, adds a new section 27A to the '34 Act, providing that the limitations period for actions pending on June 19, 1991 (the day before the Supreme Court decided Lampf Pleva) "shall be the limitation period provided by the laws applicable in the jurisdiction, including principles of retroactivity, as such laws existed on June 19, 1991." 15 U.S.C. § 78aa-1(a). The rule in this circuit on June 19 was the federal period of one year from discovery (and no more than three years from the sale), established by Short in 1990. As of June 19, 1991, the retroactivity of Short was an open issue.

The same day it decided Lampf Pleva, the Court issued James B. Beam Distilling Co. v. Georgia, ___ U.S. ___, 111 S.Ct. 2439, 115 L.Ed.2d 481 (1991). Although no more than three Justices signed any of the opinions in that case, the disparate expressions make it clear that six members of the Court concur with Justice Souter's conclusion that "[o]nce retroactive application is chosen for any assertedly new rule, it is chosen for all others who might seek its prospective application. The applicability of rules of law are not to be switched on and off according to individual hardship". 111 S.Ct. at 2447-48 (lead opinion). To apply a new rule in the first case is to prescribe its application in all cases. The Court applied the federal period of limitations in Lampf Pleva, over the objection of two Justices who thought the decision should not be made retroactive. 111 S.Ct. at 2785-88 (O'Connor, J., joined by Kennedy, J., dissenting). We applied the federal rule to the parties in Short. Putting Short together with Beam means that the law in effect in this circuit on June 19, 1991, was the federal period of limitations. Accord, Welch v. Cadre Capital, 946 F.2d 185 (2d Cir. 1991); Boudreau v. Deloitte, Haskins Sells, 942 F.2d 497 (8th Cir. 1991); Anixter v. Home-Stake Production Co., 947 F.2d 897 (10th Cir. 1991) (all holding Lampf Pleva retroactive).

Perhaps, however, the new § 27A prevents all consideration of Beam. If so, then the question is whether Beam altered the law. For if Beam restates established law, then the jig was up in this circuit when, in July 1990, Short applied the federal period to the parties in that case. Beam is an outgrowth of American Trucking Associations, Inc. v. Smith, 496 U.S. 167, 110 S.Ct. 2323, 110 L.Ed.2d 148 (1990), in which the question whether a civil decision applied to the parties in the law-changing case must be applied retroactively divided the Justices evenly. The status of the law of retroactivity on June 19, 1991, the day before Beam, was accordingly obscure. Our court might have decided the question either way — just as the two circuits that addressed the subject before Beam did not agree on the proper treatment. Compare Hill v. Equitable Trust Co., 851 F.2d 691, 695-99 (3d Cir. 1988) (applying new federal period of limitations to pending cases), with Welch v. Cadre Capital, 923 F.2d 989 (2d Cir.) (applying new federal period prospectively only), vacated and remanded for reconsideration in light of Beam and Lampf Pleva, ___ U.S. ___, 111 S.Ct. 2882, 115 L.Ed.2d 1048 (1991), on remand, 946 F.2d 185 (1991) (applying federal period retroactively).

The parties filed their briefs several months before Congress enacted § 27A. We have had no guidance on its meaning from other circuits. So far as we can tell, the legislation poses an interesting issue only for this one. As of June 19, 1991, three courts of appeals had applied a federal limitations period to litigation under § 10(b). The others will continue to apply periods derived from state law to suits filed on or before June 19, 1991. The second circuit had adopted a federal period of limitations but before June 19 had held that the new rule was not retroactive; it will therefore join the other circuits in applying state law. The third circuit had adopted a federal rule and applied it retroactively; § 27A thus compels the third circuit to continue using a federal rule. Only in this court was the retroactivity of the federal rule an open question on June 19; only we will have to decide whether Beam applies or changes the law. If we can expect no help from our judicial colleagues, it is all the more vital that we receive aid from the parties. Counsel have not addressed the subject in this court, and we invite them to do so on remand — and, of course, in this court should the subject return here.

5. The jury awarded as damages the full price of the stock, plus interest. This is a rescissionary measure, and although §§ 11 and 12 of the '33 Act, 15 U.S.C. § 77k, l, use such a starting point, those sections also allow the defendant to reduce the award by demonstrating that the misstatement did not cause the decline in value. Damages under § 10(b), by contrast, usually are the difference between the price of the stock and its value on the date of the transaction if the full truth were known. E.g., Goldberg v. Household Bank, f.s.b., 890 F.2d 965, 966-67 (7th Cir. 1989); Jordan, 815 F.2d at 441-43. Sometimes this principle comes under the name "loss causation": the plaintiff must establish that the misstatement caused him to incur the loss of which he complains; it is not enough to establish that the misrepresentation caused him to buy or sell the securities. E.g., Bastian v. Petren Resources Corp., 892 F.2d 680, 683-84 (7th Cir. 1990).

A rescissionary measure of damages implies that the Medtest stock was worthless in late 1982. Yet the corporation is still in business, apparently close to putting its process on the market (or selling it to a larger firm). Although a patent application is worth (materially) less than a patent, it is worth something, and some patent applications are worth a great deal. Apparently neither the Pommers nor defendants attempted to enlighten the jury about the value of Medtest's stock in 1982. They might believe it prudent to do so if there is a second trial.

REVERSED AND REMANDED.