POSNER, Circuit Judge.

In 1977 Bennie Lenard brought this civil rights suit against an Illinois municipality and four of its police officers, seeking damages for events growing out of an arrest. He claimed that the officers had beaten him; and that the officers and the municipality had conspired to deprive him of his right to equal protection of the laws (because he is black), had prosecuted him maliciously, and had conspired to prosecute him maliciously (and thereby to “obstruct justice”). A jury trial resulted in a judgment for $360,000 in damages, and the district judge awarded Lenard $181,000 in attorney’s fees under 42 U.S.C. § 1988. (All dollar figures in this opinion are rounded off to the nearest $1,000.) The defendants appealed from both the judgment and the fee award. We reversed the judgment, ordering all but the equal protection claim and all defendants except officers Argento and Sansone dismissed, and remanding for a new trial to determine the damages due Lenard on that claim. 699 F.2d 874 (7th Cir.1983). We vacated the award of attorney’s fees and suggested that on remand the district court should reduce it since Lenard had not prevailed in his suit to the extent the court had thought when it awarded the fees. Id. at 899.

On remand, with the case now assigned to a different district judge, Judge McMillen, who has since retired from the bench, there was a four-day damage trial to a jury, resulting in a judgment for Lenard of $267,000 in damages, from which the defendants have not appealed. The judge awarded Lenard $327,000 in attorney’s fees, from which the defendants have appealed. They also challenge an item of costs, but the challenge is frivolous and requires no discussion.

The defendants resist the award of attorney’s fees on several grounds: because it is larger than the first award, even though this court in effect told the district court to cut down that award; because (a related point) the district court cut only 251 of the 3,399 hours requested by Lenard’s lawyers for the entire time put in on the case, including time put in on the claims that failed; because the award exceeds the judgment; because it exceeds the fee to which Lenard’s lawyers are entitled by their contingent-fee contract with Lenard (the contract provides for a 50 percent fee in the event of a retrial, and 50 percent of *1245 $267,000 is only $134,000, which is little more than a third of the fee award); because Lenard turned down a settlement offer of $250,000; because Lenard was represented at trial by three lawyers; and because the district judge did not make adequate findings to sustain such a large award of fees. The defendants say the highest award to which Lenard conceivably is entitled is $90,000.

Their argument about Lenard’s refusing a settlement offer of $250,000 is frivolous. (There was also an offer of a “structured” settlement, i.e., one to be paid over time, of $425,000, but we are told neither the details of the offer nor its discounted present value.) The offer, made on the eve of the retrial, was inclusive of fees. Lenard went on to obtain a judgment for damages and fees that, combined, is more than twice as large as the offer. He would have been foolish to accept the offer — the judgment alone is larger. The defendants should not have wasted our time with this argument, nor with complaining that Lenard was represented by three lawyers. The defendants were also represented by three lawyers, and while we do not lay down a flat rule that it is always reasonable for one side to have as many lawyers as the other side, neither shall we lay down a flat rule of one plaintiff’s lawyer per case. The number of Lenard’s lawyers is relevant only to whether the district judge made adequate findings to justify the large fee that he awarded. In addition, the defendants’ claims that a contingent fee contract places a ceiling on the fee that can be awarded under section 1988 and that the fee award cannot exceed the award of damages were both scotched by the Supreme Court in City of Riverside v. Rivera, — U.S. —, 106 S.Ct. 2686, 2694-97, 2700, 91 L.Ed.2d 466 (1986), though, as will appear, the existence of such a contract and the relationship between the fee award and the damages award are relevant to the reasonableness of the former.

The defendants’ complaint that the district court disobeyed this court by refusing to cut the original award (instead the district court nearly doubled it) has merit, however — though less than the defendants think, because they ignore a change in law since the first appeal. At that time this circuit required a plaintiff to allocate lawyer time among his different theories and reduce his fee request if one or more theory failed. See 699 F.2d at 899 (“attorney’s fees should be awarded only for the preparation and presentation of claims on which the plaintiff has prevailed”), and cases cited there. The defendants argue that since this court rejected three out of Lenard’s four theories of liability, the fees allocable to the liability phase of the case should be cut by three-fourths. The results would be dramatic. Of more than 3,000 hours in lawyers’ time for which the district judge awarded fees, only 250 were incurred in connection with the retrial, when only the viable theory remained in the case; all the rest related to the first trial, the post-trial motions, and the first appeal.

But the mechanical claim-chopping approach advocated by the defendants was rejected by Hensley v. Eckerhart, 461 U.S. 424, 436 n. 11, 103 S.Ct. 1933, 1941 n. 11, 76 L.Ed.2d 40 (1983). The Supreme Court changed the axis of inquiry from successful-unsuccessful to related-unrelated. “[Ujnrelated claims [must] be treated as if they had been raised in separate lawsuits, and therefore no fee may be awarded for services on the unsuccessful claim.” Id. at 435, 103 S.Ct. at 1940 (footnote omitted). But in many cases “the plaintiff’s claims for relief will involve a common core of facts or will be based on related legal theories. Much of counsel’s time will be devoted generally to the litigation as a whole, making it difficult to divide the hours expended on a claim-by-claim basis. Such a lawsuit cannot be viewed as a series of discrete claims. Instead the district court should focus on the significance of the overall relief obtained by the plaintiff in relation to the hours reasonably expended on the litigation.” Id. at 435, 103 S.Ct. at 1940. For tactical reasons and out of caution lawyers often try to state their client’s claim in a number of different ways, some *1246 of which may fall by the wayside as the litigation proceeds. The lawyer has no right to advance a theory that is completely groundless or has no factual basis, but if he presents a congeries of theories each legally and factually plausible, he is not to be penalized just because some, or even all but one, are rejected, provided that the one or ones that succeed give him all that he reasonably could have asked for.

The standard set forth in Hensley supplants that of our previous opinion in this case, and a change of law is a reason not to apply the doctrine of law of the case. We shall give Lenard the benefit of Hensley, and therefore not reduce the award merely because most of his theories were rejected — if the theories on which he won and those on which he lost were related and if his overall success in the lawsuit justified the hours his lawyers spent on the case.

Lenard passes the second test. Even after losing his beating claim in this court, his conspiracy-to-obstruct-justice (by prosecuting him maliciously) claim, and his malicious prosecution claim, so that he was down to just his equal protection claim, Lenard managed to obtain on retrial a damages judgment three-fourths the size of the original judgment. The difference between the judgments, moreover, may be due entirely to the vagaries of trial by jury rather than to the loss of three claims.

But we must also consider the first criterion in Hensley, which is whether the claims are related. The beating claim and the claim of a conspiracy to deny Lenard the equal protection of the laws are related to each other, because they both grow out of the beating. The claims of malicious prosecution and of conspiracy to prosecute maliciously are similarly related to each other, but are not related to either the beating claim or the equal protection claim. The beating and equal protection claims relate to the conduct of the police after they arrested Lenard (both immediately after the arrest and later on, at the police station). The malicious prosecution claims relate to the question whether the police had probable cause to arrest him. The former pair essentially states a claim for excessive force; the latter for false arrest. False arrest and excessive force are unrelated except in forming a.sequence. Arresting a person on probable cause does not justify beating him up, and the beating does not invalidate the arrest. These forms of misconduct thus are independent conceptually as well as in time. Compare Garrity v. Sununu, 752 F.2d 727, 734-35 (1st Cir.1984). And, unlike the situation in Abraham v. Pekarski, 728 F.2d 167, 175 (3d Cir.1984), evidence showing one form of misconduct is not relevant to establishing the other.

While in Mary Beth G. v. City of Chicago, 723 F.2d 1263, 1279-80 (7th Cir.1983) , we described the plaintiffs false arrest and excessive force claims as seeking “damages flowing from the conduct surrounding the arrest itself,” we did so only to contrast those claims with the claim, which we held to be unrelated, for damages from a strip search conducted at the police station after the arrest. Since both the false arrest and excessive force claims had failed, there was no question of whether to award attorney’s fees for one but not the other. But they were (and in this case are) as separate from each other as both are separate from a claim of police misconduct occurring after the arrested person has been brought to the police station (where the strip search in Mary Beth G. took place). They are as different as a claim of discrimination and a claim of retaliation for opposing that discrimination. See Sisco v J.S. Alberici Construction Co., 733 F.2d 55, 58-59 (8th Cir.1984); see also Martin v. Lauer, 740 F.2d 36, 47-48 (D.C.Cir.1984); Mims v. Shapp, 744 F.2d 946, 954 (3d Cir.1984). Sequential claims—claims that are causally related (Lenard wouldn’t have been beaten if he hadn’t been arrested) but do not depend on the same facts—are not related for purposes of Hensley. A plaintiff should not be rewarded for a failed attempt to báse liability on conduct that did not result in an actionable wrong to him, even if the conduct closely precedes, or *1247 follows close on the heels of, conduct that was wrongful. The case must be remanded to redetermine the fees to which Lenard is entitled.

The growing volume of attorney’s fees litigation in this circuit prompts us to make some observations for the guidance of the parties and district judge on remand and of bench and bar in other, similar cases.

1. We cannot sustain a substantial award of attorney’s fees on the basis of so skimpy an opinion as the district court wrote. The opinion devotes only a page to the question of attorney’s fees, and uses the page to announce its conclusion rather than to state reasons for the conclusion. Many district judges treat the fee-award phase of a case as a detail, and this is natural; it is also appropriate — when the plaintiff’s request is for only a few hundred or a few thousand dollars. See Henry v. Webermeier, 738 F.2d 188, 193 (7th Cir.1984). But here the request was for more than $500,000 in attorney’s fees and the award was for $327,000. Such an award is not chicken feed. It is considerably larger than the average judgment in a federal case; it is more than 30 times the minimum amount in controversy required in a federal diversity case. It would not be chicken feed even if the question were (as it may become on remand) not $327,000 versus $500,000 but $90,000 versus $100,000.

2. Although the Supreme Court has now made clear both that a contingent fee contract does not place a ceiling on the award of attorney’s fees and that the award may exceed the judgment (in Rivera it was almost 10 times greater than the judgment), the existence and terms of such a contract, and the relationship between the fees sought by counsel and the relief obtained, continue to be relevant to the district judge’s decision on how large a fee to award. The statute allows only a reasonable fee. This means a fee large enough to induce competent counsel to handle the plaintiff’s case, but no larger. H.R.Rep. No. 1558, 94th Cong., 2d Sess. 8-9 (1976); S.Rep. No. 1011, 94th Cong., 2d Sess. 5-6 (1976), U.S.Code Cong. & Admin. News 1976, p. 5909; City of Riverside v. Rivera, supra, 106 S.Ct. at 2697. The existence of a contingent fee contract cannot be conclusive evidence of what a reasonable fee is, because the plaintiff’s lawyer might not have been willing to take the case on the terms set forth in the contract except for the hope of some additional award under section 1988. The contract in this case gives Lenard’s lawyers the larger of the contingent fee and the statutory fee. Thus, should the statutory fee exceed $134,000 (half the damage award), the lawyers would get more than the contingent fee and Lenard would get to keep the whole damages award. Maybe without the inducement provided by the expectation of an award under section 1988 Lenard’s lawyers would not have persisted with this case, now in its tenth year. The ultimate damages judgment of $267,000, which took two trials and one appeal to win, is one of the largest unreversed judgments in a police-brutality case in this circuit; perhaps Lenard’s lawyers could not realistically have hoped for so much. In that event the prospect of a contingent fee (and no more) might not have been sufficiently attractive to induce them to take and press this case with the vigor, and the investment of time, necessary to maximize the chances of winning big.

The district judge made no findings with respect to the bearing of the contingent fee contract or (what is closely related) the relationship between the damages and the fees sought. Had he focused on these things he might have concluded that this was the kind of civil rights case where it is evident from the start that the plaintiff has a solid claim for substantial damages — the kind of civil rights case that resembles an ordinary tort case, conventionally and satisfactorily handled on a contingent fee basis. In that event (though not otherwise) the fee set in the contingent fee contract would have been presumptively adequate to attract competent counsel. We said in Kirchoffv. Flynn, 786 F.2d 320, 328 (7th Cir.1986), that “the use. of contingent fees is appropriate in cases that en *1248 force old precedents and allow effective compensation as a percentage of the total recovery.” Although this was said before Rivera, there is no inconsistency. While all the Justices agreed that a contingent fee contract does not place a ceiling on the amount that may be awarded under the civil rights attorney’s fees statute, see 106 S.Ct. at 2696-97, 2700, 2705, none suggested that the terms of such a contract were irrelevant, see id. at 2696 n. 9. We had said no more in Kirchoff. See also Chalmers v. City of Los Angeles, 796 F.2d 1205, 1211 (9th Cir.1986); Cooper v. Singer, 719 F.2d 1496, 1503 n. 12 (10th Cir.1983) (en banc).

It will not do to romanticize civil rights litigation unduly. While some of it is marginal financially, some of it is adequately remunerative, and it is no part of civil rights law to overcompensate successful plaintiffs. Where the market in legal services can work, it should be allowed to work. See Merritt v. Faulkner, 697 F.2d 761, 770 (7th Cir.1983) (separate opinion). This may — or may not — be such a case. That is an issue for the district judge to resolve in the first (and we hope last) instance.

If the district judge had considered the terms of the contingent fee contract he might have gone on to question whether all of those 3,000-plus hours — a number resulting largely from the presence of three lawyers on Lenard’s side — were really necessary. When a personal-injury lawyer takes on a tort case that he reasonably thinks worth $250,000, he doesn’t put in 3,000 hours on the case; a 33 or 40 or even 50 percent contingent fee will not remunerate him for that amount of time. Yet no one thinks that persons with such claims cannot obtain adequate representation on a contingent fee basis. Civil rights tort suits are not identical to conventional personal injury suits; but where the civil rights suit involves substantial stakes, settled precedent, and no defense of immunity, the resemblance may be close enough to give the terms of the contingent fee arrangement considerable evidentiary significance — a proposition that we do not interpret Rivera to have rejected. Five Justices in Rivera thought the award of attorney’s fees too high, though Justice Powell, the swing Justice, concluded that the district court had not abused its discretion and he therefore voted to affirm. 106 S.Ct. 2698, 2700.

3. After the district judge on remand finishes carving out the time that Lenard’s lawyers spent on unrelated claims that failed, the fee award calculated by multiplying hours spent on the related claims by the appropriate hourly fee may be less than the contingent fee; in that event Lenard rather than the defendants may want to use the terms of the contingent fee agreement as a basis for arguing for an upward adjustment, and we do not rule out such arguments. Furthermore, if as Lenard’s counsel contended at oral argument this case has yielded benefits in civil rights enforcement beyond the damages awarded — because of the publicity that the suit has received, which may make the police of Melrose Park, Illinois more sensitive to the rights of black citizens — those benefits can be considered in determining whether the time put in by Lenard’s lawyers was reasonable in light of the success (intangible as well as tangible) obtained. That is, if a civil rights suit generates what an economist calls an “external benefit” (a benefit not reaped by Lenard himself), there should be compensation for having created it.

We vacate the fee award and remand to the district court to recalculate the award in accordance with the principles laid down in this opinion. We order the defendants to interpose no more frivolous objections to Lenard’s fee request, on pain of sanctions if they do, and we encourage the parties to compromise their differences. The $90,000 figure suggested by the defendants is probably too low; the $327,000 awarded in the order that we are vacating is clearly too high. The parties should be able to negotiate a satisfactory compromise, failing which the district court will have to rede *1249 termine the fee. We hope not to see this case a third time.

Vacated and Remanded.