This case involves a challenge under the corporate law of Indiana to a “poison pill,” which is a device used by corporate managers to fend off hostile tender offers. The case is a sequel to Dynamics Corp. of America v. CTS Corp., 794 F.2d 250 (7th Cir.1986). Although the Supreme Court has noted probable jurisdiction of the appeals by CTS and the State of Indiana from the part of the decision that invalidated on federal constitutional grounds an Indiana statute regulating takeovers, — U.S. -, 107 S.Ct. 258, 93 L.Ed.2d 17 (1986), that part is not pertinent to this sequel.
On March 10 of this year Dynamics Corporation of America, which already owned 9.6 percent of the common stock of CTS Corporation, made a tender offer that if accepted would bring its stock holdings up to 27.5 percent. CTS’s stock was trading at $36 a share (all dollar figures are rounded to the nearest dollar). The price in the tender offer was $43. On March 22, after very hurried consideration, CTS adopted a shareholders’ rights plan of the kind known as a “poison pill.” Under the plan, if and when one shareholder (namely Dynamics) obtained 15 percent or more of the company’s common stock, every other shareholder would be entitled to buy a package of stock and debentures for 25 percent of the market price of the package. In April the district court issued a preliminary injunction against the poison pill. 637 F.Supp. 406 (N.D.Ill.1986). We affirmed. The tender offer then went through, and Dynamics used its newly obtained 27.5 percent position in CTS to begin a proxy fight to oust CTS’s existing board of directors.
Meanwhile CTS had responded to the district court’s injunction by referring the question of further defensive measures to a committee of outside directors. After several days of intensive deliberation the committee returned to the full board with a recommendation that the company be sold and that pending sale a new poison pill be adopted under which, if any shareholder obtained 28 percent of the company’s common stock, all the other shareholders would be entitled to turn in their shares and receive in exchange for each share a $50 debenture (bond), payable after one year, with interest at 10 percent per annum. The plan would remain in effect for one year but could be cancelled by the board of directors at any time. It would be can-celled automatically if anyone made a cash tender offer for all outstanding shares at a price of $50 or more.
The board adopted the plan on the same day we upheld the district court’s preliminary injunction (April 23). At the time, CTS’s stock was selling at $38 a share, but it rose to $45 the next day, when CTS announced that it had adopted the plan as part of a strategy for selling the company — though this was also the first day after we affirmed the district court’s preliminary injunction. Dynamics moved immediately to enjoin this second “poison pill,” but on May 3 the district judge denied the motion, 635 F.Supp. 1174 (N.D.Ill.1986), and she adhered to this decision on reconsideration following issuance of our opinion affirming the preliminary injunction against the first poison pill, 638 F.Supp. 802 (N.D.Ill.1986). Dynamics has appealed.
The election for the board of directors was held on May 16. CTS’s board, which had campaigned on a platform of selling the company and had represented that the second poison pill was designed to maximize the price at which the company would be sold, won reelection by a narrow margin.
As noted in our previous opinion, citing earlier decisions of this court, “the task for a district judge asked to grant a preliminary injunction is to compare the irrepara*708ble harm to the plaintiff if the injunction is denied, weighted by the likelihood that the denial would be erroneous because the plaintiff will prevail in the plenary trial, with the irreparable harm to the defendant if the injunction is granted, weighted by the likelihood that the grant would be erroneous because the defendant, not the plaintiff, will prevail in the trial.” 794 F.2d at 252. This “sliding scale” approach implies that if, so far as it is possible to determine, the irreparable harm that the plaintiff is likely to suffer from denial of the preliminary injunction is equal to that which the defendant is likely to suffer if the injunction is granted, likelihood of success on the merits becomes decisive, and the question is simply: who is likelier to win the case when and if it is tried on the merits — the plaintiff or the defendant? Id.
Our previous opinion noted that the irreparable harms from granting or denying the preliminary injunction appeared to be in equipoise in this case, so that analysis would have to focus on the likelihood that Dynamics would prevail on the merits if its suit to invalidate the poison pill was tried. After pointing out that the controlling law was that of Indiana and that in matters of corporation law the Indiana courts normally take their cue from the Delaware courts, which are more experienced in such matters, we observed that while the board of directors of a corporation that is a target or potential target of hostile tender offers has the power to adopt a poison pill, the particular poison pill it adopts must be reasonably related to the goal of shareholder wealth maximization. Since, moreover, there is a potential conflict of interest between the managers and shareholders of the target, courts are not simply to rubber stamp the board’s judgment but must review it carefully to make sure that in adopting the poison pill the board really was acting in the best interests of the corporation. See, e.g., Unocal Corp. v. Mesa Petroleum Co., 498 A.2d 946, 954-55 (Del.1985); AC Acquisitions Corp. v. Anderson, Clayton & Co., 519 A.2d 103, 107-08 (Del.Ch.1986). In making this determination the court must consider the procedures leading up to the adoption of the poison pill and the terms of the poison pill. CTS’s counsel acknowledged at the argument of the present appeal that the action of a board of directors in adopting a poison pill is subject to “enhanced scrutiny” by the courts, because of the potential conflict of interest between the board and the shareholders. In nevertheless and inconsistently insisting that the standard of judicial review is whether the board acted with gross negligence, counsel was misstating the test set forth in the Delaware decisions and in our previous opinion to govern review of poison pills.
The first poison pill that CTS adopted clearly flunked the test. The circumstances of its adoption indicated that the board’s objective was to block Dynamics’ tender offer regardless of the consequences for the welfare of CTS’s shareholders. So did the terms of that poison pill. If Dynamics completed its tender offer the pill would be administered and Dynamics would find that its 27.5 percent ownership position had shrunk to 20.7 percent, the dilution being caused by the stock component of the pill. It would also find that it had incurred a capital loss in excess of $20 million, for the value of the company would be no greater as a result of the exercise of the rights conferred by the poison pill but Dynamics’ share of the company would be smaller. And CTS would be saddled with a heavy ($80 million) new long-term debt, for no value received.
This summary of our previous decision establishes the basic principles that guide our consideration of the present appeal. As we interpret the two opinions that the district judge wrote in explanation of her refusal to enjoin the second poison pill, the irreparable harms are again in equipoise, so that decision again must turn on an evaluation of the merits. In what appears to be the key passage in the opinions, the judge said that “while [Dynamics] has shown fair ground for litigation, its probability of success ... is decidedly lower than *709was true in connection with the earlier rights plan, and ... the balance of hardships does not weigh sufficiently in [Dynamics’] favor to justify injunctive relief.” 635 F.Supp. at 1177-78. Implicit in the “sliding scale” approach to deciding whether to grant a preliminary injunction is the proposition that even if the movant has a relatively small chance of prevailing at trial, he may still be entitled to the injunction if the irreparable harm to him from a denial would greatly exceed the irreparable harm to the defendant from a grant. The quoted language makes clear that the district judge did not think this such a case, and implies that she thought it a case where the irreparable harms were (as before) approximately equal. It is true that elsewhere she describes the irreparable harm to Dynamics as “slight.” Id. at 1182. But she never says the irreparable harm to CTS would be great. And her statement that the harm to Dynamics would be slight appears to have been based on her expectation that Dynamics would win the proxy fight (then two weeks away) — in which event, controlling the board as it would, it could waive the poison pill at will. In fact Dynamics lost the proxy fight, and the balance of harms is therefore as before. In any event, the ground on which the district judge denied the preliminary injunction had nothing to do with the balance of harms; it was simply that she thought Dynamics unlikely to prevail at trial — unlikely, that is, to succeed in invalidating the second poison pill.
The scope of appellate review of a decision granting or denying a preliminary injunction depends on the ground of that decision. If it is based on a purely factual determination, we can reverse only if the determination is clearly erroneous; if on a purely legal determination, we can reverse if our view of the law is different from the district court’s, for review of legal determinations is plenary. But if the decision is based on the balance of harms, or on that balance together with the likelihood of success on the merits (the likelihood operating to weight the balance in favor of plaintiff or defendant in accordance with the sliding scale approach), we can reverse only if persuaded that the district court abused its discretion, that is, was not merely incorrect but unreasonable. See, e.g., Lawson Products, Inc. v. Avnet, Inc., 782 F.2d 1429, 1436-39 (7th Cir.1986); Roland Machinery Co. v. Dresser Industries, Inc., 749 F.2d 380, 388-92 (7th Cir.1984).
Where the balance of harms is equal and therefore drops out of the analysis — where the only issue is the plaintiff’s likelihood of winning when the case is tried — the scope of appellate review is similar to that of final judgments. The two opinions the district judge wrote in connection with Dynamics’ request for a preliminary injunction against the second poison pill contain a mixture of factual and legal determinations all feeding into her ultimate determination regarding the lawfulness of the pill. That determination is more tentative than if the case had been tried, because it is based on a less complete development of evidence than possible in a regular trial, but its ingredients are the same and — with issues of irreparable harm cancelled out — is a matter of finding facts and applying law, not of exercising discretion. In reviewing the decision, therefore, we must examine the procedures leading up to the adoption of the second poison pill and the terms of the pill, and, accepting all of the district judge’s findings of fact that are not clearly erroneous, we must determine whether she was justified in concluding that the infirmities that had led her and us to invalidate the first poison pill have been cured. Of course, if we cannot determine from the judge’s findings and conclusions, read against the record of the case, whether her findings of fact are clearly erroneous, then we must remand for further analysis; and this, as we shall see in due course, is the case here.
The second poison pill was adopted in different circumstances and has different provisions from the first. The second was devised by the outside directors (constituted as a special committee) in consultation with the company’s investment advisor, Smith Barney. The district court found *710that unlike the board the first time around, the special committee did not start with the preconceived idea of preventing a takeover of CTS whether by Dynamics or any other person or firm that might make a tender offer; that after thoroughly and impartially considering a variety of alternative methods of maximizing the shareholders’ wealth the committee decided that the company should be sold; and that it adopted the poison pill (the second pill, the one challenged in this appeal) in order to maximize the price at which the company would be sold. The trigger was set at 28 percent to allow Dynamics to complete its tender offer but prevent it from acquiring any more shares. The committee’s main concern was that if Dynamics acquired enough shares to have a “blocking position” other potential acquirers would be deterred from trying to buy the company. A secondary concern was that Dynamics would be able to acquire the additional shares without having to pay a premium for thereby obtaining control of the company, thus depriving all shareholders of any participation in a “control premium.”
The district court also found that the committee had made an honest-to-goodness effort to value the company and had come up with a price of $50 per share. Anyone who thinks the company is worth at least that much can take it over simply by making a cash tender offer for all outstanding shares of the company, at a price of $50 or better per share; for if such an offer is made, the poison pill will self-destruct. The shareholders are unlikely to refuse such an offer, since $50 is the company’s own valuation (and not a modest one, either, as we shall see). Should a lower price become attractive to shareholders because the market price of CTS’s stock falls — for example, at a market price of $29 (the current market price) even a sale price as low as $42 would yield a higher premium than a sale price of $50 if the market price were still $38 — the board of directors has the power to cancel the pill, and would be under considerable pressure to exercise that power. During the proxy fight Dynamics said that if it got control of the board it would sell the company for $46 a share; and though this may just have been puffing, it is some evidence that the board’s valuation of the company was not too far off target. The stock market may have considered the second poison pill, unlike the first, a positive development, for the price of CTS stock rose (though only temporarily) after the announcement of the second pill; it had fallen after the announcement of the first. (But the cause may have been our affirmance of the preliminary injunction against the first poison pill, rather than the announcement of the second poison pill.) Finally, this poison pill is to remain in the medicine chest for only a year; and if at the end of that time the board has not sold the company as the board said the poison pill would help it do, the board will have a lot of explaining to do and may lose the next election, having barely squeaked by in this one after promising to sell the company.
If the facts we have recited were the only facts established by the record, we would affirm the district judge’s decision. But there are more. To begin with, regarding the procedural adequacy of CTS’s decision to adopt a second poison pill, the role of the investment advisor is troubling. As noted in our first opinion, Smith Barney’s compensation package for advising the board on how to respond to Dynamics’ original tender offer included a bonus if the offer was defeated (“the independence fee”). For the second round Smith Barney was again the advisor but this time it was retained by the committee of outside directors rather than by the entire board, and, according to an uncontradicted affidavit, its compensation package does not include an “independence fee.” What it does include is unknown, however, since Smith Barney’s agreement with the committee is not in the appellate record. But Judge Getzendanner said that Smith Barney will obtain an “incentive fee for the white knight strategy,” which we understand to mean that it will be paid more if CTS is sold to a buyer agreeable to the board than if it is acquired in a hostile takeover. In *711the record is a letter to Smith Barney from CTS’s treasurer describing the fee that Smith Barney is to receive under different scenarios and fixing a generous fee for sale to a “white knight” (a buyer acceptable to the board) and a negligible fee if Dynamics obtains control of the board in a proxy fight. The letter contains no mention of a fee if there is a hostile takeover. The letter was written before the special committee of outside directors retained Smith Barney, but we are unable to determine from the record whether it is still in effect. Besides providing for the independence fee, the original compensation agreement had included a large bonus for a sale to a white knight, and again we are unable to determine from the record whether that agreement (presumably minus the independence fee) is still in effect. The situation regarding Smith Barney’s compensation in the second round thus is thoroughly confused, suspicious, and not clarified by the district court’s opinions. The scheme of compensation may involve a bonus of some sort for sale to a white knight, even if a “black knight” offers the shareholders more money for their shares. More generally, there is no indication to what extent the amount of compensation is linked to the price at which the company was sold — though some such link might be necessary to align Smith Barney’s self-interest with the interests of the shareholders.
Standing alone, this impurity in the procedures leading up to the adoption of the second pill would not warrant reversal, or even a remand. Nor would the fact that the outside directors aren’t really disinterested, since they will probably lose their directorships (and directors’ fees) if CTS is acquired, unless the acquiror is a foreign company. A lack of disinterest cannot be equated to bad faith or procedural unfairness; otherwise defensive measures would be unlawful per se, which they are not. It is true that the special committee acted with some haste, but given its concern (the reasonableness of which we shall examine later) that Dynamics might, after completing its initial tender offer, seek to buy enough additional shares to acquire a blocking position, it had to move rapidly.
Dynamics cites Smith v. Van Gorkom, 488 A.2d 858 (Del.1985), where the Delaware Supreme Court held directors liable to their corporation for having accepted a friendly offer to buy the company at a price far above market price without extended deliberations or consultation with an investment advisor. Dynamics argues that Van Gorkom sets the standard of deliberation applicable to this case. However, so forcefully and cogently has the decision been criticized (see Fischel, The Business Judgment Rule and the Trans Union Case, 40 Bus.Law. 1437 (1985); Herzel & Katz, Smith v. Van Gorkom: The Business of Judging Business Judgment, 41 Bus.Law. 1187 (1986)) — essentially for having ignored the adage that a bird in the hand is worth two in the bush — that we hesitate to conclude that the Indiana courts would follow it. But even if they would, the decision would not control the present, very different, case. In Van Gorkom the deliberations leading up to the decision to accept the offer occupied only two hours, and it was perhaps arguable that the directors had made a snap decision, without adequate consideration of alternatives that might have yielded an even greater profit for the shareholders. Here the deliberations stretched over days, and (another deference from the situation in Van Gorkom) included consultation with the company’s investment advisor. True, the consultation may have been contaminated by the method of compensating the advisor. But we are not minded to hold directors to a standard of procedural scrupulousness that would transform their role from that of businessmen to that of Article III judges. Furthermore, if directors can be sued both for selling their company too hastily (Van Gorkom), that is, at too low a price, and for selling it too slowly (this case), that is, for setting too high a price, they are placed on a legal razor’s edge.
More serious than any procedural irregularities in the decision by CTS’s board to adopt the second poison pill are the substantial, and as yet unanswered, questions *712regarding the relationship between the terms of the second poison pill and the welfare of CTS’s shareholders — but of course these questions make the issue of Smith Barney’s compensation more serious than it would otherwise be. The first question is, why is the poison pill activated when one shareholder obtains 28 percent of the shares, rather than 50 percent? Despite much talk of “blocking positions,” the record contains no evidence as to how Dynamics could use 28 percent or any other fractional ownership short of 50 percent to block anything. Although Indiana law authorizes corporations to require a superma-jority vote for certain structural changes, such as a merger, CTS has not imposed such a requirement; it can do anything by a simple majority of the shares. See Ind. Code § 23 — 1—5—2(b). Dynamics can no more block majority decisions with 49.99 percent of the company than with 27.5 percent. If someone tenders for more than 50 percent of the shares (however slightly more), and succeeds in the tender, he can if he wants squeeze out Dynamics and become sole owner of CTS. Of course, squeezing out a minority shareholder is not completely painless, for the shareholder has appraisal rights that, depending on how a court values them, may make the transaction a costly one for the majority shareholder; recent developments in Delaware law have strengthened (some think too much) the position of the dissenting shareholder. See, e.g., Herzel & Colling, Establishing Procedural Fairness in Squeeze-Out Mergers After Weinberger v. UOP, 39 Bus.Law. 1525 (1984). And the larger the minority shareholder, the greater the pressure the majority shareholder may feel to squeeze him out rather than tolerate his continued presence. For even though the minority shareholder, whether large or small, can’t wrest control of the board of directors from the majority shareholder, he may by virtue of the size of his stake in the company have an incentive to play a more vigorous role as corporate monitor and gadfly than would a smaller shareholder (CTS in its brief calls Dynamics a “contentious” and “unruly” minority shareholder). So all things considered, CTS might indeed be a somewhat less attractive acquisition target with a very large minority shareholder. But whether the difference between a 27.5 percent shareholder and a 49 percent shareholder is material with respect to these concerns is an issue unexplored on this record.
We do not want to be understood as holding that ownership of 28 percent of CTS’s stock could not create a “blocking position” in some sense that might justify the use of this percentage to trigger a poison pill. Poison pills with lower triggering percentages have been upheld. But every poison pill is sui generis, and there must be evidence justifying setting the triggering percentage below the level that would give a minority shareholder an actual legal right to block decisions taken by the majority. We cannot find that evidence. It is true that the board obtained the opinion of a qualified financial expert (other than Smith Barney) that the 28 percent trigger was proper, but the opinion was offered as a conclusion; its basis was not explained. The district court’s opinions do not illuminate the issue.
The secondary ground for the 28 percent trigger, that it is necessary in order to assure that all shareholders will participate in any premium for conveying control cf the company, is also unexplained. It is true that the more shares Dynamics has of CTS’s stock, the fewer additional shares it need buy to take control of CTS; but it is hard to see how this can hurt the shareholders. If someone tenders for 51 percent of the company, Dynamics is in a good position to make a counteroffer, since it has a big base to start from. But of course the counteroffer will succeed only if it is at a higher price. Dynamics’ position as a large minority shareholder, far from discouraging the auction of the company— CTS’s avowed goal in adopting the second poison pill — should encourage it. Nor can we see how this conclusion would be altered if Dynamics had 28 percent instead of 27.5 percent of CTS’s stock — or for that matter 49 percent. A recent study finds *713that the existence of a large minority shareholder encourages takeovers, see Schleifer & Vishny, Large Shareholders and Corporate Control, 94 J.Pol.Econ. 461, 463 (1986), and while the study neither is conclusive nor shows that shareholders of CTS would be better off if Dynamics had more than 27.5 percent of CTS’s shares, it does underscore the absence from the record of persuasive or even plausible reasons for the “blocking position” or “stalking horse” or “control premium” grounds for the 28 percent trigger. In a recent case the presence of a large minority shareholder appears to have facilitated rather than impeded an advantageous takeover. See Hecco Ventures v. Sea-Land Corp., Civ. Action No. 8486 (Del.Ch. May 19, 1986) [Available on WESTLAW, DE-CS database]. We are pointed to no contrary examples.
A further consideration is that the larger Dynamics’ position is, the easier it will be for another firm to take over the company by making a deal with Dynamics, for acquiring Dynamics’ shares will bring it closer to owning a majority. The stated premise of CTS’s hostility to Dynamics’ increasing its position in CTS stock is that Dynamics doesn’t care for the long-run health of CTS, that it is just a short-run profit maximizer. But if so, this means that Dynamics should be willing and eager to sell its shares to anyone that will offer it a good profit. It is true that if Dynamics had 49 percent of the company and then sold out to the owner of another one percent plus, the remaining shareholders would not obtain a control premium. But that is an argument for a fair-price amendment, which would guarantee all shareholders the same price as paid for control, rather than for a poison pill. Again these are matters not illuminated by the record or the district judge’s opinions.
Dynamics’ principal argument against the second poison pill is simply that $50 a share is too high a valuation of CTS; no one will pay $50 a share for CTS, so the practical effect of the poison pill is to make the company invulnerable to being taken over. In principle Dynamics or anyone else can make a tender offer at a lower price but if it does so the poison pill will be activated and the tender offeror (assuming it succeeds in taking over the company) will end up owing every shareholder who refuses to tender his stock but instead asserts his right under the poison pill more money than the stock is worth. Indeed, by tendering for an additional half percent of the stock Dynamics could end up as the only shareholder in a company whose liabilities (which would now include the $50 debentures received by the other shareholders along with CTS’s other debts) might greatly exceed its assets: a bankrupt company. If owners of all 72 percent of the shares not owned by Dynamics after our hypothetical .5 percent tender (100 percent — 28 percent) asserted their rights under the poison pill, CTS would have acquired an additional debt of $200 million, compared to only $80 million of additional debt under the original poison pill. Additional debt of $200 million would be equal to three-quarters of the net worth of the company even if the company is valued (very generously, as it appears) at $50 a share. Any shareholder who tenders his stock to the tender offeror gives up his rights under the poison pill; but so long as the tender offer price is below $50, many and perhaps most shareholders will refuse.
The $50 figure is only a face value and of course securities frequently trade for less or more than their face value. In this case, if the company is worth less than $50, so that if all or most of the shareholders claimed their rights under the poison pill the company might be forced into bankruptcy, it might seem to follow that the shareholders would accept less than $50 to surrender their rights through sale of their stock to a tender offeror. But this is not correct. As soon as enough shareholders tender to give the tender offeror 28 percent, the poison pill will be triggered and some fraction of the remaining shareholders will try to make more than the tender offer price by swapping their stock for the debentures. If enough do so, the tenderer will find himself in control of a company with tremendous debt, and little value; he *714will have bought not a company but a mass of debt. The prospect will deter tender offers at prices below $50, whatever the debentures may really be worth.
But this only shows that the poison pill is indeed likely to be effective in its stated purpose of discouraging tender offers at less than $50. That purpose is not unlawful provided the board was trying reasonably and in good faith to sell the company at the highest possible price. The trigger price, however, is highly relevant to the issues of reasonableness and good faith, since if set too high it will prevent all tender offers, not just those that are below the corporation’s sale value. Since the entire premise of the second poison pill was that it was designed to facilitate the sale of the company, the pill cannot be upheld if the trigger price is an unreasonably high sale price.
The district judge’s determination that, at least on the basis of the evidence presented at the preliminary injunction hearing, the valuation was reasonable and in good faith is a finding of fact (see Mucha v. King, 792 F.2d 602, 604-06 (7th Cir.1986)), so we must uphold it unless it is clearly erroneous. And her finding that $50 was a reasonable and good-faith price for a security for which Dynamics had been willing to pay $43 (the price in its tender offer) is not on its face unreasonable, especially when we consider that once one departs from market values the process of valuing an asset becomes fraught with uncertainty. See Metlyn Realty Corp. v. Esmark, Inc., 763 F.2d 826, 835-36 (7th Cir.1985). Of course this might be a reason for forbidding corporate management (or even a committee of outside directors) to set a reservation price for the sale of the corporation rather than giving the market a free hand, but the law does not go so far.
Nevertheless the method by which the $50 figure was chosen is sufficiently troublesome to make this another issue on which further consideration by the district judge is necessary before we can determine whether her conclusion regarding reasonableness and good faith was clearly erroneous. Smith Barney determined CTS’s sale value primarily by multiplying CTS’s expected earnings for the coming year by a price-earnings multiple of 15.5. The multiple was based on the multiples commanded by other manufacturers of electronic components, but no effort was made to establish comparability. A price-earnings multiple reflects investor expectation concerning the company’s entire stream of future earnings (suitably discounted), not just a one-year projection. CTS is a troubled company and even if it were forecasting its next year’s earnings correctly, investors might not value the company at the same multiple of those earnings that other companies in its field are valued at. CTS has, moreover, a history of extravagant earnings projections. This history would tend to depress its price-earnings multiple in the stock market, at least when the earnings component is a projection, as distinct from actually realized earnings.
Even if the multiple of 15.5 was reasonable, the earnings figure that Smith Barney multiplied by 15.5-$3.23 per share for 1987-was a wildly extravagant forecast. As recently as April 1986 the company had reduced its estimate of expected earnings for 1987 from $1.50 to only 89$. The $3.23 figure was based on an expectation — for which no ground other than optimism has been suggested — that both CTS and its industry would show dramatic growth in the coming year. Smith Barney had estimated just a month before making its $50 valuation that while at a price of $40 a share the probability of selling the company would be 85 percent, at $43 it would be only 35 percent. By this method of estimation, at $50 a share the probability of selling the company would be close to zero. In a separate opinion issued on May 3, upholding Dynamics’ claim that CTS’s board had made misleading representations in the course of the proxy fight, the district judge had described $50 a share as “an outside high value based substantially on untested, optimistic management projections,” had noted CTS’s history of overly optimistic earnings projections, had said that “CTS *715really does not expect a triggering event and hence $50 notes very probably never will be issued,” and had pointed out that Smith Barney in estimating the company to be worth $50 a share had taken the board’s $3.23 projection at face value without attempting to verify it. It is true that Merrill Lynch advised the board of directors that Smith Barney’s methodology in valuing the company was sound, but Merrill Lynch, like Smith Barney, took for granted the earnings projection which the board had supplied Smith Barney. That projection, never independently validated, is the weakest link in the valuation chain. The affidavit of CTS’s other independent financial expert was prepared after the fact to bolster CTS’s litigating position.
When these facts are put together with the history recounted in our previous opinion, we are left with considerable uncertainty regarding the basis on which the district judge was able to reject the inference that the second poison pill was designed to keep CTS from being sold without the board’s consent and to keep Dynamics from expanding its base of stock holdings from which to wage another proxy fight for control of the board — so that, in the end, the current management and directors would keep their jobs. As the district judge herself said, “That the defendants are determined to stop [Dynamics] is evident, as they consider it better to sell CTS at a potentially depressed price than watch [Dynamics] take control.” 635 F.Supp. at 1181. About CTS’s determination to stop Dynamics, even at the price of selling the company for less than it was worth, the district judge said only that “this hostility towards [Dynamics], even if unreasonable, cannot be simply equated with a goal of entrenchment.” Id. We are not clear why not.
But it is possible that other facts outweigh those we have just reviewed and may justify the district judge’s conclusion that the $50 valuation was not either deliberately inflated or unreasonably high, though this we cannot determine from her opinions. In her separate opinion dealing with CTS’s alleged misrepresentations in the proxy fight, she faulted CTS for having failed to convey to the shareholders another valuation of CTS’s stock that Smith Barney had made — this one at $75 per share. If $75 was somehow in the ballpark (though how it could have been is unclear), her evident skepticism concerning $50 becomes difficult to understand; this is an inconsistency that can be cleared up on remand. Another point is that if the 15.5 multiple were applied to CTS’s earlier estimate of its 1987 earnings — 89c—it would mean that the stock was worth less than $14 a share, which no one thinks reasonable; remember that Dynamics’ tender was at $43. So though Smith Barney’s valuation procedure seems deeply flawed, the $50 valuation may not be invalidated by the flaws. Further considerations bearing on the reasonableness of the second poison pill are that it was to remain in effect for only a short period (one year), that it was before the shareholders (though not formally submitted to them for ratification) during the proxy fight, and that it is cancellable by the board. Concerning the last point, if the board turned down an objectively favorable tender offer for less than $50 by refusing to cancel the pill, it might have a hard time explaining this to the shareholders at the next proxy fight — a fight that Dynamics no doubt will wage with as much vigor and determination as the one just past, which it narrowly lost and perhaps only lost because of the promise by CTS’s board to sell the company. Unless that promise is redeemed or the company’s performance dramatically improved, the present board will probably lose the next proxy fight, just as it expected to lose the last one if Dynamics’ tender offer went through.
Yet these facts are not decisive either. The cancellable feature of the poison pill just permits the board to sell the company to a “white knight,” which it could do without a poison pill. And the authenticity of the stockholders’ indirect approval of the poison pill in the proxy campaign is undermined by the misrepresentations that CTS made concerning the value of the company, misrepresentations calculated to make the *716shareholders think (despite the absence of any grounds) that if they kept the old board it would sell the company within a year for at least $50 a share. It is true that Judge Getzendanner ordered corrective action — another mailing to the shareholders, to correct the misrepresentations — but the efficacy of this action, carried out as it was only a few days before the election, may be doubted.
Judge Getzendanner may not have clearly erred in finding that CTS had acted reasonably and in good faith in adopting the second poison pill, but there are enough doubts on this score — doubts concerning Smith Barney’s incentives, the choice of a 28 percent trigger in the particular circumstances, and the method by which a $50 valuation was arrived at — doubts not dispelled by the district court’s findings of facts — to require that the case be remanded for further consideration of Dynamics’ request for a preliminary injunction. And since nothing is to be gained by decision on an obsolete record, the parties should be allowed to present evidence of relevant developments since the last hearing, including evidence bearing on CTS’s efforts to find a buyer for the company and on the realism of its earnings projections back in May in light of the company’s experience since then. A forecast of earnings is not invalidated by failure to achieve it; even a realistic forecast is just an estimate, implying some probability that actual performance will deviate from it. But consistent failure to achieve or even approach forecasted performance is relevant evidence on the realism of a company’s methods of forecasting, an important issue in the present case.
We expressed in our last opinion and we repeat our skepticism concerning the propriety of poison pills. The present pill essentially prevents the shareholders from deciding to sell the company at less than $50 per share unless the directors approve. We grant that there is at least an arguable concern with two-tiered tender offers. The offeror offers a high price for enough shares to wrest control of the company from its present management in the expectation of being able to buy the remaining shares later at a lower price; knowing this, shareholders may fall over each other to tender their shares in response to the first offer; as a result, the time for holding an auction of the company with more than one bidder is drastically foreshortened. But this problem can be solved by a fair-price amendment to the corporation’s charter, which merely forbids different prices between the tiers. And any defensive provision that commands the support of a majority (and, a fortiori, if it commands a super-majority) of the corporation’s shareholders, as a charter amendment must do, is more likely to be in the shareholders’ best interests than a measure adopted by the board of directors alone. For supporting evidence see Linn & McConnell, An Empirical Investigation of the Impact of ‘Anti-takeover’ Amendments on Common Stock Prices, 11 J.Fin.Econ. 361 (1983). The poison pill, typically and here adopted by vote of the board rather than of the shareholders, eliminates the market for corporate control at any price below the value of the pill to the shareholder electing to take it (or rather, to administer it to the corporation). Eliminates it, period, when, as may be the case here, the value of the pill exceeds the maximum reasonable sale value of the company. The poison pill turns every tender offer into a two-tier offer with a higher rather than lower back-end: the offeror offers at one price and then is forced to buy out nontendering shareholders at a higher price. If it is enough higher, the incentive to make a tender offer is destroyed.
We noted in our previous opinion that the Delaware courts have emphatically refused to outlaw the poison pill. As a general statement this is accurate. But it was made before the CTS board decided to put the company on the block. The Delaware courts distinguish between defensive tactics adopted at a time when corporate independence is a plausible goal and those adopted later, when it is no longer plausible and the board is just trying to get the highest possible price for the company. Although the second poison pill in this case is *717said to be modeled on the one upheld in Revlon, Inc. v. MacAndrews & Forbes Holdings, Inc., 506 A.2d 173 (Del.1986), and indeed bears a family resemblance to it, Revlon had adopted its poison pill before the directors decided to sell the company. Once they decided to sell it “the whole question of defensive measures became moot. The directors’ role changed from defenders of the corporate bastion to auctioneers charged with getting the best price for the stockholders at a sale of the company.... [Ojbtaining the highest price for the benefit of the stockholders should have been the central theme guiding director action.” Id. at 182. And on this ground the court invalidated the defensive measure taken by Revlon after the decision to sell the company. To similar effect see Hanson Trust PLC v. ML SCM Acquisition, Inc., 781 F.2d 264, 282-83 (2d Cir.1986) (New York law). The second poison pill in the present case came after CTS’s board decided to sell the company.
We need not speculate on whether in some circumstances a poison pill might be a reasonable element of a plan for a corporate auction. The poison pill in this case operated, as we have said, as a reservation price: no one could bid lower than the value of the poison pill to the shareholders and hope to induce a significant number of them to sell their shares. Auctioneers frequently establish reservation prices, that is, refuse (or reserve the right to refuse) to sell the item to be auctioned unless a bid higher than the specified price is received. The analogy, however, is imprecise. The owner of the auctioned property is selling his own property; the management of a corporation is selling property that belongs to others, the shareholders, whose ability to ensure management’s faithful performance of its fiduciary duties to them is limited. Also, items usually are auctioned because there is no ready market for them, hence no market price. CTS’s stock is traded on the New York Stock Exchange, though it can be argued that the value fixed by the market is the value of the marginal share rather than the sale value of the entire company as a unit.
We need not pursue these issues further. It would be premature to decide in this case that, as a matter of Indiana law, a board of directors intent on selling the corporation for the highest possible price cannot set a reservation price and cannot use a poison pill as the means of doing so. But even if it can, the reservation price must be reasonably related to the value of the corporation, and the decision to establish such a price and do so by means of a poison pill must be made in good faith after proper consideration. The record before us creates a serious doubt whether CTS has satisfied this standard. But the record is stale and the district judge’s findings incomplete, and the best course in these circumstances is to return the matter to the district court for further consideration and evidence. No criticism of the district judge’s handling of this complex case is intended; the extreme time pressure under which she was forced to decide Dynamics’ motion for a preliminary injunction against the second poison pill is noted. But with the benefit of more leisurely consideration we discern flaws in her analysis which require not reversal but merely a remand for further consideration and evidence.
We must consider Dynamics’ alternative ground for appeal, which if sustained would require entry of the preliminary injunction without any further proceedings. Dynamics argues that the second poison pill runs afoul of Ind.Code § 23-l-25-l(a), which provides that “All shares of a class must have preferences, limitations, and relative rights identical with those of other shares of the same class.” CTS has only one class of common stock, yet while the poison pill is a right pertaining to every share of that stock, any shareholder who reaches the 28 percent trigger point is forbidden to exercise his poison pill rights. Dynamics cannot by acquiring another half percent turn its shares in for debentures, but every other shareholder can if Dynamics makes such an acquisition. That is why we said earlier that if Dynamics reached that point it would be the only shareholder, *718provided the others exercised their poison-pill rights and thereby exchanged their shares for debentures.
Although Dynamics’ argument is consistent with the language of the statute, Dynamics concedes that the Delaware courts have distinguished between discrimination among shares and discrimination among shareholders and have allowed corporations to engage in the latter form of discrimination as a concomitant to the deployment of defensive measures. See Providence & Worcester Co. v. Baker, 378 A.2d 121, 123 (Del.1977). Dynamics’ reply, that the Delaware Code does not expressly require identical rights across shares, as the Indiana Code does (see Del.Gen.Corp.Law § 151(a)), is wide of the mark. The Delaware Code forbids discrimination, and the Delaware courts have construed this to mean discrimination between shares, not shareholders. The Indiana Code forbids discrimination between shares but is silent on discrimination between shareholders.
If Dynamics’ argument were accepted, it would mean that the Indiana legislature, in adopting a statute designed to deter takeovers (for the provision in question is part of Indiana’s new anti-takeover statute, portions of which we held unconstitutional in our last opinion), unwittingly facilitated takeovers by outlawing poison pills. Nothing in the Delaware cases to which Indiana looks for guidance in corporate law matters would have alerted the legislature to this possibility, and we hesitate to conclude that the legislature inadvertently outlawed the poison pill by means of a statute designed to facilitate defensive measures against hostile tender offers.
If the statute is read in the manner suggested by Dynamics, no Indiana corporation will be able to adopt a poison pill. The essence of the poison pill is that the acquisition by one shareholder of a specified fraction of the company’s stock gives other shareholders-not the acquiring shareholder-certain rights, here the right to exchange one’s shares for $50 debentures. Allow the acquiring shareholder to exercise those rights, and the poison pill becomes a corporate suicide act. The ease would be different if the poison pill singled out Dynamics, and stripped its shares of rights pertaining to other shareholders’ shares. But the poison pill will strip rights from any shareholder who acquires 28 percent of the shares, and Dynamics is not the only shareholder who could do so. The statute was not designed to outlaw the poison pill yet that is the reading Dynamics seeks to impress on it.
Dynamics’ further argument that, read to allow poison pills, the Indiana statute violates the commerce clause of the federal constitution is frivolous. The commerce clause does not require states to outlaw private contracts that may impede interstate commerce. A state that, for example, fails to enact or enforce an antitrust law does not thereby violate the Constitution. The order of the district court denying Dynamics’ motion for a preliminary injunction is affirmed insofar as the statutory ground for the motion is concerned, but otherwise is vacated, and the case is remanded to the district court for further proceedings consistent with this opinion. No costs in this court.
Affirmed in Part, Vacated in Part, and Remanded with Directions.