Judicial approval of a settlement involving the valuation of securities is a discretionary act, and an application to revisit that approval asks a court a question about its own discretionary processes. If the judge who approved the settlement concludes that misstatements in the hearings leading to the approval did not affect his earlier decision, that brings the litigation to an end. We affirm the district judge's decision not to reopen his approval of the settlement in this case.

* In 1974 Esmark, Inc., which owned 53.5% of the stock of TransOcean Oil, Inc., decided to acquire the rest. TransOcean's stock was selling for about $7 per share. On September 30, 1974, Esmark made a tender offer through its subsidiary Vickers Energy Corp., at $12 per share, for all of the outstanding stock. The offer fetched more than four million shares, raising Esmark's ownership to 87.5%. Three groups of shareholders brought class action suits, contending that the tender offer circular withheld information tending to show that Esmark believed TransOcean to be worth a good deal more than $12 per share. A large institutional holder brought a fourth suit.

The suit filed in the Chancery Court of Delaware moved forward while the others languished. The Supreme Court of Delaware concluded that Esmark had violated its duty to disclose its internal valuations showing that TransOcean was worth more than $12 per share. Lynch v. Vickers Energy Corp., 383 A.2d 278 (Del.1977). In proceedings on remand in Delaware, the Chancellor determined that TransOcean's stock was worth less than $12 in 1974 and less than $15 in 1978. The Chancellor therefore held that there should be no remedy. While an appeal from this decision was pending, Esmark and the class representatives in the action filed in the Northern District of Illinois reached a settlement. The terms of the settlement, as finally amended, called for Esmark to pay an additional $2.80 for each of the shares it obtained in the tender offer and for TransOcean to be merged into Esmark, with 0.90 shares of Esmark's stock replacing each outstanding share of TransOcean's stock. TransOcean's board approved the merger (over one dissent) on condition that the Illinois court find the price "fair" within the meaning of Delaware law.

Investors who had tendered their stock to Vickers in 1974 were free to opt out of this settlement; those who did so ultimately received $8.50 per share in litigation that followed a second decision of the Supreme Court of Delaware. Lynch v. Vickers Energy Corp., 429 A.2d 497 (Del.1981) (damages formula based on rescissionary measures), overruled in part by Weinberger v. UOP, Inc., 457 A.2d 701 (Del.1983). But those who still held shares of TransOcean at the time of the settlement could not opt out. The settlement called for a merger, which would convert all shares of TransOcean to shares of Esmark by operation of Delaware law. Thus investors who thought they could obtain more than the value of 0.90 Esmark shares by litigation (or simply by holding the stock) could do this only by persuading the district court in Illinois to reject the settlement. Investors holding approximately half of the 1.5 million outstanding shares objected, and the district court held a hearing in August 1979 on the fairness of the settlement. (This procedure gave the objectors an automatic prior review they would not have had under Delaware law. Esmark, holding 87.5% of the stock, could have rammed through a merger by force of its own votes. The dissatisfied investors could have stopped the merger only by convincing the Chancery Court of Delaware that the merger lacked "entire fairness" and a "business purpose," the legal standards before Weinberger abolished the "business purpose" requirement and made the appraisal remedy exclusive in the absence of fraud.)

The proponents of the settlement introduced three kinds of testimony. A professor of financial economics testified that the price of TransOcean in the stock market represented the collective wisdom of informed investors about the value of the stock, and that the price offered in the settlement (the value of 0.90 Esmark shares in August 1979 was about $25) was well in excess of the value as measured by the market. See Mills v. Electric Auto-Lite Co., 552 F.2d 1239, 1244-48 (7th Cir.), cert. denied, 434 U.S. 922, 98 S. Ct. 398, 54 L. Ed. 2d 279 (1977) (endorsing this method of determining value for stocks of corporations trading in liquid public markets).1 An investment banker testified that TransOcean should be valued by capitalizing its average annual earnings over the past five years; the value derived in this way also was less than the one offered in the settlement. Irwin L. Levy, an oil and gas consultant, offered a third method: he computed the value of TransOcean's oil and gas reserves and its costs of production on existing and projected fields; he predicted changes in the price of oil and gas, and derived TransOcean's after-tax profits for future years. Then he discounted these future profits to present value at a rate of 18% per year. Levy came up with a value of $23.90 per share, again less than the settlement.

The objectors cross-examined Levy vigorously about his assumptions and methods. They also attacked Levy's approach through the testimony of other participants in the industry who thought that Levy had overstated TransOcean's costs, understated the potential of its fields, and used an excessive rate of discount, which substantially reduced the present value of the future cash flows. They challenged him to name a transaction in which the 18% rate had been used; he did not. Other witnesses for the objectors--including the dissenting director of TransOcean and a financial analyst who established value by looking at prices paid in recent transactions concerning other oil and gas company stocks--maintained that TransOcean was worth between $30 and $48 per share.

The district court approved the settlement. TransOcean was merged into Esmark. In June 1980 Esmark put TransOcean back on the market. It received bids ranging from $381 million (or approximately $31 per former share of TransOcean) to some $740 million ($60 per share). It accepted the high bid, from Mobil Corp., in August 1980. Several of the objectors at the 1979 hearing (the appellants here), who held about 450,000 shares, immediately sought to reopen the judgment under Fed.R.Civ.P. 60(b), arguing that Esmark had withheld from the court the fact that it planned to resell TransOcean at a profit. The objectors later added the contention that Levy had lied to the court about his experience in oil and gas transactions and had made numerous errors in his assumptions and calculations, thus undercutting his testimony, on which, the objectors maintained, the court had relied in approving the settlement. Moreover, they contended, counsel for Esmark knew or should have known of these problems and concealed them from the court. (Though Rule 60(b)(3) limits to one year the time within which to make a motion based on misrepresentation or fraud, no one has made anything of the fact that the objectors took more than one year after the 1979 judgment to raise the contention that Levy testified falsely, which is now the central issue in the case. We shall be similarly silent, though without implying that this would not have been worthy of attention had the parties raised it.)

The district court held another hearing in 1982, this time concentrating on Levy's experience and computations, on counsel's knowledge of these, and on Esmark's intentions in August 1979. The court concluded that Levy had exaggerated his experience in oil and gas transactions, representing that he was involved in "deals" in which the parties had used cash flow calculations and an 18% rate of discount. The court found this untrue; Levy had not been involved in sales of assets ("deals") involving such calculations, though his firm may have used the same methodology to give advice about deals that fell through and about lending transactions. The judge stated, however, that Levy's misrepresentation had not affected the judge's decision to approve the settlement. He concluded that Levy's experience in sales of oil and gas assets had been challenged in 1979; "he was unable [when challenged in 1979] to name a single, completed transfer in which the 18% rate was applied, a fact of which we then made careful note." The objectors attacked Levy vigorously and introduced evidence that 13% was a more realistic rate. The evidence produced in 1982 did not shed harsher light on Levy's computations than the evidence in 1979. As a result, the court concluded, "we were not misled into relying on Levy's testimony for more than it was worth."The court also stated: "We have found no evidence of fraud or deliberate deceit in this testimony. There is evidence that Levy's evaluation of TransOcean's net asset worth was conservative. There is also evidence that some of Levy's statements were not well supported. Finally, there is some evidence that Levy misstated or exaggerated the extent of his experience in actual transactions.... To the extent Levy made positive misstatements about his experience, he was challenged in 1979." The court observed that Esmark's managers may have thought that TransOcean was worth more than $25 per share but that they were entitled to present an expert with a "conservative" assessment. The process of valuation is inexact and people will be found at both ends of a range. Esmark presented the facts necessary to compute value; the court thought this enough. Finally, the court rejected the assertion that Esmark harbored a secret intention to resell TransOcean. The objectors had presented no more than a chain of inferences based on Esmark's need for cash and its ultimate sale of TransOcean. This was not enough, the court concluded, to show "even more probably than not" that Esmark had concealed in August 1979 a current intention to resell.

The objectors launch a fusillade of charges against the district court's failure to reopen the judgment. They maintain that the court misunderstood the depth of Levy's deceit, that Esmark's lawyers recklessly vouched for Levy's expertise, that Esmark withheld an internal valuation exceeding Levy's and hid from the court its intent to resell TransOcean, and that Levy erred in his calculations. We first consider the standard under which to review the district court's decision and then apply that standard to the facts developed at the 1982 hearing.

Judgments in civil cases fix the rights of parties and entitle them to go about their lives. They may be reopened only for extraordinary reasons. C.K.S. Engineers, Inc. v. White Mountain Gypsum Co., 726 F.2d 1202, 1205 (7th Cir.1984); Smith v. Widman Trucking & Excavating, Inc., 627 F.2d 792, 795 (7th Cir.1980); DeFilippis v. United States, 567 F.2d 341, 342 (7th Cir.1977); Brennan v. Midwestern United Life Insurance Co., 450 F.2d 999, 1003 (7th Cir.1971), cert. denied, 405 U.S. 921, 92 S. Ct. 957, 30 L. Ed. 2d 792 (1972).

There are good reasons for the stringent limits on reopening a final judgment. The rule that the final judgment governs the future induces the parties to focus all of their efforts in a single proceeding. If they know that the judgment will establish their rights once and for all, they will bring to bear the information and energies necessary to produce an informed decision. If they believe that they can have a second try, they are more likely to skimp the first time around, and as a result the judicial process will become less accurate.

True, mistakes are bound to occur. There are few perfect trials. But the inevitable shortfalls of the legal system do not call for revisiting judgments. Second proceedings have problems of their own. The passage of time dims memories and makes evidence harder to find and assess. The knowledge that comes with hindsight is a mixed blessing at best; new knowledge may color the evaluations of what people knew (or should have known) at an earlier time, though the judicial process must focus on the information available at the time of the underlying events or the first trial. And of course there may be new mistakes in the second proceeding even as the parties and the court cure the blunders of the first. So even if the first proceeding was flawed, there is no assurance that the second proceeding will come closer to the truth. These considerations inform the development of the law under Rule 60(b), as they inform principles of issue and claim preclusion, the interpretation of statutes of limitations, and other doctrines of repose in civil and criminal litigation. See Federated Department Stores, Inc. v. Moitie, 452 U.S. 394, 401-03, 101 S. Ct. 2424, 2429-30, 69 L. Ed. 2d 103 (1981); United States v. Kubrick, 444 U.S. 111, 117-18, 100 S. Ct. 352, 356-57, 62 L. Ed. 2d 259 (1980); Wainwright v. Sykes, 433 U.S. 72, 88-90, 97 S. Ct. 2497, 2507-08, 53 L. Ed. 2d 594 (1977).

These considerations have led us to say, in the cases we have cited and in many others, that judgments may not be reopened under Rule 60(b) except in compelling and extraordinary circumstances. The objectors remind us that we have also said that the provisions of Rule 60(b) should be liberally construed to avoid injustice. E.g., Brennan, supra, 450 F.2d at 1003; Ervin v. Wilkinson, 701 F.2d 59, 60-61 (7th Cir.1983). But as we observed in C.K.S. Engineers, supra, 726 F.2d at 1205-06, when reviewing some cases using the "liberal construction" language, this simply is a reminder that the purpose of Rule 60(b) is to authorize some limited review of judgments. A court ought not let the interests of finality lead it to decline to afford the sort of review Rule 60(b) authorizes. It does not mean that relief from a final judgment should be freely given; to the contrary, our review of the cases in C.K.S. Engineers ineluctably led to the conclusion that Rule 60(b) relief is an "extraordinary remedy and granted only in exceptional circumstances." Id. at 1205.

This conclusion is reinforced by the standard of review. The decision of a district court declining to reopen a judgment under Rule 60(b) may be reviewed "only for abuse of discretion," Browder v. Director, Department of Corrections, 434 U.S. 257, 263 n. 7, 98 S. Ct. 556, 560 n. 7, 54 L. Ed. 2d 521 (1978), and we have held that to find an "abuse of discretion" the court of appeals must conclude that "no reasonable man could agree with the district court." Smith v. Widman Trucking, supra, 627 F.2d at 795-96; Simons v. Gorsuch, 715 F.2d 1248 (7th Cir.1983). When the district judge reviewing the Rule 60(b) motion also rendered the underlying judgment and concludes that any error complained of did not affect the earlier decision, the movant's task on appeal is more difficult still. International Nikoh Corp. v. H.K. Porter Co., 374 F.2d 82, 84 (7th Cir.1967). It is perhaps conceivable that a moving party could establish that "no reasonable man" could find that an error did not affect the result when the judge rendering the original decision has found no effect, but there cannot be many such cases.

There is discretion and then there is discretion. The scope of the district court's discretion is greatest when there is no strictly legal rule of decision, when there is no standard against which to compare the district court's decision. Roland Machinery Co. v. Dresser Industries, Inc., 749 F.2d 380, 388-90 (7th Cir.1984); Piper Aircraft Corp. v. Wag-Aero, Inc., 741 F.2d 925, 937-40 (7th Cir.1984) (Posner, J., concurring); United States v. McCoy, 517 F.2d 41, 44 (7th Cir.) (Stevens, J.), cert. denied, 423 U.S. 895, 96 S. Ct. 195, 46 L. Ed. 2d 127 (1975); Henry J. Friendly, Indiscretion About Discretion, 31 Emory L.J. 747 (1982). When there is no right answer, it becomes difficult to speak of "error," and the scope of review shrinks accordingly. See In re Sinadinos, 760 F.2d 167, 170 (7th Cir.1985).

A motion under Rule 60(b) often puts to a court a question without a right answer. The district judge must weigh incommensurables--the value of finality, the probability that an error affected the outcome of the proceeding, the probability that a second go-round would produce a "better" outcome, the costs of that second proceeding to the parties (and ultimately to society as the finality of judgments is undercut). Dealing with these intersecting planes of legal argument is a task of great subtlety, calling on all the skills of the district judges. It is not, however, a task that gives rise to "error" (as opposed to questionable exercise of judgment), unless the judge leaves something important out of his analysis. This is why we have been so deferential in Rule 60(b) cases to decisions not to reopen.

The deference is never greater than when the underlying judgment is a settlement. The settlement itself is an exercise in compromise rather than resolution of legal issues. The parties compromise on legal and factual matters, and there could be a very substantial range within which the compromise would be reasonable. There is no "right" settlement. When the underlying judgment cannot be "wrong" in the strict sense--though an error could influence the range within which the settlement takes place--a judge's decision to leave well enough alone cannot readily be "wrong" either.2

The objectors offer two reasons for more searching appellate review here. First, they contend that there is enhanced review for perjury or misstatements. They say the review should be "plenary." Second, they argue that because this is a securities case and any false statements concern the value of securities, we should apply the standards of the securities acts rather than Rule 60(b). Neither argument is availing.

Rule 60(b)(3) allows a court to relieve a party from a final judgment on account of "fraud (whether heretofore deemed intrinsic or extrinsic), misrepresentation, or other misconduct of an adverse party" provided the motion is filed within a year after entry of the judgment. We have indicated that an adverse party's fraud or subornation of perjury permits relatively free reopening of the judgment when the perjury goes to the heart of the issue. Peacock Records, Inc. v. Checker Records, Inc., 365 F.2d 145, 147 (7th Cir.1966); see also Harre v. A.H. Robins Co., 750 F.2d 1501, 1503 (11th Cir.1985) (false testimony on ultimate issue requires vacation of judgment where the defense attorneys know or should have known of the falsity).3

But not all falsehoods satisfy the standard of the Rule. "[O]f an adverse party" modifies "fraud" and "misrepresentation" as well as "other misconduct." See Fackelman v. Bell, 564 F.2d 734, 737 (5th Cir.1977); Moore's Federal Practice p 60.24 at p. 60-216 (1983). A movant must establish fraud by "clear and convincing evidence," Ervin v. Wilkinson, supra, 701 F.2d at 61, and this includes the burden of showing that the adverse party procured or knew of any false testimony. It is the rare trial in which no one speaks falsely (if only by accident) in any particular, and a heavy burden is necessary here (as we explained above) to protect the finality of judgments against efforts to turn the vicissitudes of litigation into grounds for more litigation still. Unless the false testimony can be traced to the adverse party, the case must be decided under the residual category of Rule 60(b)(6), which permits review only when "the violation created a substantial danger of an unjust result." Merit Insurance Co. v. Leatherby Insurance Co., 714 F.2d 673, 682-83 (7th Cir.1983), cert. denied, --- U.S. ----, 105 S. Ct. 297, 83 L. Ed. 2d 232 (1984). It must be unmistakable that a new proceeding would come much closer to the truth.

The district court concluded that neither Esmark nor Esmark's counsel knew of Levy's misstatements. The objectors do not now challenge this conclusion. They say only that in the exercise of due diligence counsel should have investigated Levy's background further and found out. This smacks of the proposition that every party "vouches" for its witnesses, a view long departed and little missed in federal practice. See Fed.R.Evid. 607 and the Advisory Committee's note. Expert witnesses such as Levy are free agents. Parties and counsel have an obligation not to deceive the court about the witness and to correct statements they know to be false, but they are not responsible for the details of the witness's testimony. Rule 60(b)(3) therefore does not apply to the attack on Levy, and the objectors must establish that his testimony "created a substantial danger of an unjust result."

The arguments based on the securities laws do no better. The securities laws doubtless require truthful statements and penalize fraud, as the objectors emphasize, but this does not imply that they authorize the reopening of judgments or override the standards of Rule 60(b). The securities laws were designed to handle transactions in markets, in which the parties might have too little incentive to reveal information and might find it too easy to tell lies or half-truths. They do not apply expressly to misstatements in the course of litigation, and they were not designed to regulate the conduct of trials. To the contrary, Sec. 3(a)(10) of the Securities Act of 1933, 15 U.S.C. Sec. 77c(a)(10), exempts from ordinary registration procedures securities issued in an exchange transaction approved by a court.

Litigation contains its own safeguards--including discovery, cross-examination, the supervision of a judge, and exposure to prosecution for perjury--designed to regulate the flow of information. Although courts have established private rights of action to enforce Rule 10b-5, 17 C.F.R. 240.10b-5, the catch-all anti-fraud rule based on Sec. 10(b) of the Securities Exchange Act of 1934, 15 U.S.C. Sec. 78j(b), see Herman & MacLean v. Huddleston, 459 U.S. 375, 103 S. Ct. 683, 74 L. Ed. 2d 548 (1983), courts have never believed that private actions based on Rule 10b-5 replace all other remedies for fraud. The existence of a private right of action depends on the meaning of the legislation, and it is quite farfetched to suppose that Congress meant to change the rules of civil litigation when it enacted the Securities Act of 1933 and authorized the SEC to promulgate Rule 10b-5 under the '34 Act. This is not a case involving an implied private right of action, but many of the same principles for deducing the meaning of silent legislation apply here and counsel rejection of the objectors' argument. Cf. Blue Chip Stamps v. Manor Drug Stores, 421 U.S. 723, 748-55, 95 S. Ct. 1917, 1931-34, 44 L. Ed. 2d 539 (1975) (judicially-created private rights of action under Rule 10b-5 must be modified or abrogated if they interfere with other more important policies); Santa Fe Industries, Inc. v. Green, 430 U.S. 462, 477-80, 97 S. Ct. 1292, 1302-04, 51 L. Ed. 2d 480 (1977) (same).

The principles developed in the application of Rule 60(b) are tuned to the benefits and burdens of reopening judgments. We conclude that these principles are the sole means of attack on final civil judgments. The fact that there are rules for securities cases on the merits--and, we suppose, particular rules for just about every other substantive category of litigation in federal courts--is irrelevant. The legal system has devised ways of channeling attacks on final judgments back through a single device specially designed for the purpose of reopening. The principles of claim and issue preclusion cut off one kind of attack; the rule that witnesses are absolutely immune from civil liability for their testimony, Briscoe v. LaHue, 460 U.S. 325, 103 S. Ct. 1108, 75 L. Ed. 2d 96 (1983), cuts off another; we conclude that Congress did not mean to inject the securities laws into a branch of judicial administration from which other competing principles have been carefully removed.

The objectors allege a fraud that could be $400 million, the difference between the value of TransOcean implied by the settlement and the price paid by Mobil. The current objectors' former holdings of some 450,000 shares would entitle them to "only" $15 million or so (the $45 per share they say TransOcean was worth in 1979, before increases in the price of petroleum products that account for part of Mobil's bid, less the $25 they did get, all plus interest from 1979 that would substantially increase the recovery). That is still a tolerably large number. Yet although the stakes are large and the record thick, it is easier to apply the standards of Rule 60(b) than it was to state them.

The objectors' principal argument is that Levy lied in 1979 by claiming that he had participated in "deals" in which cash flow analysis was used in conjunction with an 18% rate of discount. The lie provided the only basis, they say, for his valuation; but for the misrepresentation the 18% rate would not have been sustainable. Levy admitted at the hearing in 1982 that he did not participate in any such deals in which oil and gas properties changed hands. Levy's former testimony, the objectors maintain, unduly influenced the judge and caused him to approve the settlement. This must be so, the objectors conclude, because the district judge indicated that he agreed to the settlement because it was within the range established by the many disagreeing experts, and he did not discard Levy's view as worthless. But the district court rejected this line of argument at a factual level, concluding that cross-examination at the 1979 hearing exposed the weakness of Levy's background (in 1979 Levy "was unable to name ... a single, completed transfer in which the 18% rate was applied, a fact of which we then made careful note"), that Levy's misstatement was not wilfully false ("we have found no evidence of fraud or deliberate deceit in this testimony"), and that any misstatements did not substantially affect the judgment ("we were not misled"). Under the principles we have discussed, these findings--which are supported by the record--establish that the district court did not abuse its discretion in declining to reopen the judgment. International Nikoh Corp., supra.

We recognized above the possibility that a judge's conclusion that he had not been misled at the earlier proceedings might be set aside as clearly erroneous. The objectors ask us here not to take the judge at his word. They also argue that revelation of Levy's lack of experience would have ensured an appeal in 1979, so that the district court's word should not be permitted to be the last one even if the judge has accurately recounted his own mental state. We conclude, however, that the judge's word is sounder than the objectors' speculations and miles from an abuse of discretion under the appropriate, highly deferential, standard.

Much other evidence in the record supported Levy's valuation. The district judge pointed out that "there was substantial 'real world' justification for the 8% real rate of return on equity assumed by Levy--the real rate of return was supported by studies of Walsh and Deacon and of the Federal Petroleum [sic; should be Power] Commission as well as industry reports--and for Levy's 10% inflationary expectations." (Levy built 9-11% inflation into both TransOcean's future earnings and the discount rate. This is a permissible way to discount a cash flow to present value. Jones & Laughlin Steel Corp. v. Pfeifer, 462 U.S. 523, 538-47, 103 S. Ct. 2541, 2551-55, 76 L. Ed. 2d 768 (1983); O'Shea v. Riverway Towing Co., 677 F.2d 1194 (7th Cir.1982). Much therefore turned on the assertion that 8% was close to the real rate of return on equity demanded in the oil and gas industry in 1979.)

Levy's estimate was not the sole support of the settlement. The stock market also offered valuable information, because TransOcean stock was trading at $20 or less in 1979. The price of an actively traded stock reflects the value placed on it by the professional investors who follow a firm closely. It is a "valuation" that includes the information of many professionals, not just one. It is therefore an unusually reliable source of information when the essential conditions (liquid markets, public information, and a following by professional investors) are met. The price at which people actually buy and sell, putting their money where their mouths are, is apt to be more accurate than the conclusions of any one analyst. See Richard Brealey, An Introduction to Risk and Return from Common Stocks (2d ed. 1983); Ronald J. Gilson & Reinier H. Kraakman, The Mechanisms of Market Efficiency, 70 Va.L.Rev. 549 (1984). This is why we recognized in Mills, supra, that the price of stock in a liquid market is presumptively the one to use in judicial proceedings. See also, e.g., Seaboard World Airlines v. Tiger International, 600 F.2d 355, 361-62 (2d Cir.1979); In re LTV Securities Litigation, 88 F.R.D. 134, 144 (N.D.Tex.1980). The information most important to the valuation of TransOcean, from the price of oil and the affairs of OPEC and the state of federal regulation of oil and gas to the holdings of the firm and its historical costs of production and profits, was a matter of public record. Levy's analysis was based on the same sort of information available to market professionals. He testified that he used only information provided by TransOcean and did no independent investigation of TransOcean's reserves and costs.

Doubtless the market in TransOcean stock was not as liquid after the Esmark tender offer as before, and the price was influenced by what traders thought they would get in any acquisition by Esmark. So it is hazardous to draw firm inferences; the district judge said in 1979 that because of such circumstances he was not persuaded that the market price of TransOcean was reliable. Nonetheless, the market price was not worthless. It reflected not only the value of a merger but also the value investors would have had in the absence of a merger, which was not a certainty. It reflected as well the value of Esmark's shares to be exchanged in a merger--shares the value of which would be influenced by the value of the TransOcean assets Esmark would acquire. The price therefore offered some support for the valuation expressed in the settlement. So too do the facts that (a) the great majority of the TransOcean shareholders leaped at the chance to get $12 in the tender offer, less than half what the objectors received in this settlement, (b) the class representatives in this suit, with their own money on the line, were willing to settle for a payment worth about $25, (c) an investment banker, using a capitalization method, derived a value less than that of the settlement, and (d) the Chancellor of Delaware had recently valued the TransOcean shares at less than $15. Levy was not off in a corner by himself.

Indeed, the district judge recognized that Levy's testimony was of limited utility no matter what one makes of his honesty and experience. Levy offered a valuation based on a discounted cash flow. Such valuations are highly sensitive to assumptions about the firm's costs and rate of growth, and about the discount rate. It is a simple matter to increase or reduce the outcome of a cash flow analysis by 200% by making changes in assumptions that appear by themselves to be insignificant. The hazards of "valuation studies" in litigation are well known. In condemnation cases under the Takings Clause it is common to find opposing experts, nominally using the same methodology, testify to values an order of magnitude apart just because they use slightly different assumptions and standards of comparison.

These things happen not only because of the hazards of cash flow analysis but also because there is no "true" or "intrinsic" value for any asset. The value of a firm is just what it, or its assets, will fetch in the market. It is not written that water (on which life depends) will be cheap while gold (which is valued chiefly as a decoration) will be dear; all depends on the forces of supply and demand, not on intrinsic worth. Thus in condemnation cases courts try to approximate the results of market transactions, not to generate other measures of worth. United States v. 50 Acres of Land, --- U.S. ----, 105 S. Ct. 451, 83 L. Ed. 2d 376 (1984); United States v. Miller, 317 U.S. 369, 374, 63 S. Ct. 276, 280, 87 L. Ed. 336 (1943). Cash flow studies may be necessary when there is no other way to find value, but they are not the best way. Cf. E.I. duPont de Nemours & Co. v. Collins, 432 U.S. 46, 97 S. Ct. 2229, 53 L. Ed. 2d 100 (1977) (discussing the problem of valuing an investment company).

People estimate value differently; every distribution has an upper range and a lower range; one side will find an expert from the high range and the other an expert from the low range. "It has long been known that there is no proposition so outrageous that some well-qualified economist will not support it. Equally there are no propositions so well founded and in keeping with common sense that some economists will nevertheless not deny them." Samuel Brittan, Financial Times (Nov. 30, 1981). The experienced district judge was well aware of this. The process of valuation by testimony may obscure rather than illuminate the value under study, while market processes overcome personal biases, quirks and errors. Drawing on the test of self-interest (will people back up their views with cash?), market prices tend toward the value that the most astute investors place on the assets. Thus the strong preference of courts and the SEC for market-based values rather than ones manufactured by "experts." Collins, supra, 432 U.S. at 54-57, 97 S. Ct. at 2234-35 (the "net asset value" of the investment company, based on the market value of its holdings, is a proper indicator of value despite contrary methods of computation). The record here contained several methods of estimating market values, some of which were higher and some lower than Levy's. The district court considered them, and thus it was well warranted in concluding that a misstatement about Levy's experience in "deals" did not unduly influence the decision back in 1979.

We do not deny that the district judge "relied on" Levy in 1979; he considered all of the evidence for what it was worth. The district court did not write an opinion in 1979, and he did not say in his opinion in 1984 that he gave Levy no weight. He said disparaging things at the 1979 hearing about all of the testimony offered to support the settlement, and he was harsher on the other witnesses than on Levy. But neither, apparently, did he give Levy as much weight as the objectors think he did. He simply thought less ill of Levy's performance. He made up his mind, as he should, in light of the many methods of valuation contained in the record and the protection afforded by the self-interest of the class representatives who were willing to take $25 for their own shares. The result was not an abuse of discretion. We cannot say that a rerun of the valuation hearing--using in 1985 only the facts available in 1979--almost certainly would come much closer to the "true" value a settlement should have taken.

The objectors' challenges to Levy's other testimony about TransOcean's costs and future operations fall with the challenge to Levy's testimony about his deal-making experience. The objectors do not contend that Levy lied about the facts or his methodology, or that he withheld material information in his possession; they simply object to what he did with the data. This is nothing other than an effort to reargue what was spread on the record and decided in 1979.

We have a similar reaction to the objectors' contention that counsel for Esmark should have found and revealed the problems in Levy's testimony. Counsel need not do the work of their opponents. The district court here found that counsel for Esmark were in fact unaware of the errors in the testimony and had no reason to be aware of these problems. This finding is not clearly erroneous.

Although the objectors aver that they would have appealed in 1979--and thus exposed Levy's methods to full review had they but known of the defect in his qualifications, this line of argument is unavailing. Because of the discretionary nature of the settlement and valuation process, and the fact that at least some of the flaws in Levy's testimony were known in 1979, the appeal then would have followed the path we have just traversed. We have gone deeper than the scrutiny ordinarily required of the denial of a Rule 60(b) motion, and this additional tour through the record and its support for the settlement convinces us that the belated disclosure of Levy's nonexperience with "deals" did not affect the outcome of this litigation, let alone create "a substantial danger of an unjust result" (Merit Insurance, supra, 714 F.2d at 683).

The next argument is that Esmark in fact disagreed with Levy's valuation even back in 1979. This has two prongs: first that Esmark believed that Levy had miscalculated or misstated TransOcean's costs of developing incremental barrels of oil at more than $14 per barrel of oil, while TransOcean's management believed them to be $5.30, a figure Esmark's managers knew because it was the industry's average; second, that Esmark's management developed an internal valuation of TransOcean exceeding Levy's. The first of these compares apples and oranges. Levy was trying to estimate costs of producing future barrels of oil, while the $5.30 figure concerned historical costs. As development companies bring marginal fields into production--a decision justified by the rising cost of oil--the costs of production rise too. At all events, the district court concluded that Levy's economic assumptions were challenged fully in 1979 and that "[t]here has been absolutely no evidence presented that Esmark withheld any information concerning TransOcean's worth from us or from the ... objectors ... Nor is there any evidence that Esmark disbelieved Levy's testimony at the time it was offered...." These findings are not clearly erroneous.

The second prong--that Esmark had a different assessment of TransOcean's worth--goes to the evaluation of the facts rather than the information on which that evaluation was based. We certainly would be surprised if, in 1979, Esmark's managers did not think the stock worth more than Levy testified. Why else would they have gone ahead with the merger? It is always the case that an acquiring firm thinks the stock or other assets to be acquired worth more than the current price (which here exceeded Levy's estimate). The question is not whether the acquiring firm thinks the stock worth more, but whether the acquiring firm has withheld facts that would establish the value of the stock or assets in the absence of the acquisition. The district court found that Esmark withheld no facts.

The "valuation study" that the objectors say establishes Esmark's private beliefs contained no facts at all. It was an effort by one manager of Esmark, not skilled in oil and gas production, to assess TransOcean's worth in connection with a possible request to a German bank for financing. The manager used a system we had thought confined to scoring diving contests. He took all the values given in the August 1979 hearing, cut off the high and low ones, and averaged the rest. The district court correctly thought that keeping this "valuation" a secret was not exactly hiding the ball. The valuation was a cross between a guess and business puffery. See also 8 Del.Code Sec. 262(h) (objecting investors entitled to "fair value exclusive of any element of value arising from the accomplishment or expectation of the merger of consolidation"); Weinberger, supra, 457 A.2d at 713 ("speculative elements of value that may arise from the 'accomplishment or expectation' of merger are excluded"). Evidence amounting to no more than the subjective conclusions of the acquiring firm's managers falls in the "speculative" category; facts, not speculations, are what matter.

If any beliefs of Esmark's managers stemmed from plans to change TransOcean's operations or management, or to engage in other transactions with the assets, this too would be excluded from the process of computing value. Weinberger held that "elements of future value" are part of the entitlement of the old investors when "known or susceptible of proof as of the date of the merger" (ibid.; emphasis added). Corporate law, like the law of property and contract, allows acquirors to keep increments of value they produce in a firm, as well as those that later arise by luck, in order to give the acquirors incentives to find and improve undervalued or poorly employed properties. Here the subsequent events--including any changes Esmark may have made in TransOcean's management or operations, and its ability to set off a bidding contest for the assets--were no more than speculation as of the time of the merger. Esmark was not required to disclose its subjective wishes and beliefs. These might have been pertinent in 1974, when investors had to decide whether to tender. In 1979, though, there was no investment decision, no prediction, to make. Only the objective value of TransOcean mattered.

We, like the district court, therefore exclude from consideration the fact that Mobil paid in 1980 more than twice the value implied by the merger in 1979. Only facts known in 1979 count. After the decision developments in Iran influenced the world's markets in oil and gas. Mobil's higher bid may have been based on increases in the price of oil and gas (Levy had assumed they would increase only at the rate of inflation). The bid also may have been based on plans to change TransOcean's managers or operation. Or perhaps it showed that Mobil would accept a lower rate of return on investment, which implies a higher cash price (Mobil's internal computations may have used a rate as low as 4% rather than Levy's 8%). None of this matters. Any increment of value attributable to changes after August 1979 in the market for oil and gas, or to Mobil's willingness to make changes or bear special risks, belongs to Mobil's and Esmark's shareholders rather than TransOcean's. The investors in a firm are entitled only to what it is worth as it exists, not as it could become in other hands.

The objectors's final argument is that the district court's finding that Esmark did not have a plan in August 1979 to sell TransOcean is clearly erroneous. The objectors say that Esmark had a cash flow problem, and that the sale of TransOcean was a logical way to handle the problem; Esmark concedes that it had such a problem. The objectors point out that Esmark had some dealings in November 1979 with a German bank concerning possible dispositions of assets; they say these dealings support an inference that Esmark planned all along to sell TransOcean. We agree with the objectors that the district court was required to consider the sale to Mobil for a very limited purpose--not because the value of TransOcean rose quickly, but because Esmark had represented in 1979 that it did not plan to sell, and this may have misled the court. But this is the end of the agreement.

The district court considered very carefully whether Esmark had planned in 1979 to sell TransOcean. It was troubled by this possibility, because it believed that an impending sale would have made the valuation exercise in which it engaged unnecessary. (We need not decide whether, in light of Weinberger, this is an accurate understanding.) The court reviewed the evidence with painstaking care. It was aware that the business situation "suggests a climate in which the sale of TransOcean was a logical solution. The defendants cannot dispute this business truism." Yet it also concluded that the record shows only that Esmark pondered many ways of raising cash and did not contemplate until April 1980 a substantial probability of selling TransOcean. It would extend an already tiresome opinion to rehearse the support for this conclusion. We are satisfied, though, that it is not clearly erroneous.

AFFIRMED.