A manufacturer of boat engines, Outboard Marine, brought this suit for breach of warranty (a suit governed by Illinois law) against a supplier of cables for its engines, the ACCO division of FKI Industries, formerly known as Babcock Industries. The jury awarded Outboard Marine $4,403,000. Discontented with this verdict, Outboard Marine asked the judge to enter judgment for $12,-384,700, or in the alternative to grant a new trial on damages. The judge refused to do either, and Outboard Marine appeals.
The cables were installed in a new type of Outboard Marine engine, which failed at an alarming rate and had to be replaced at substantial cost to Outboard Marine. The evidence at trial established that ACCO had failed to make the cables in accordance with Outboard Marine’s specs and that this failure had contributed to the failure of the new engines. There was also evidence of engine defects unrelated to the defective cables— defects that were the sole responsibility of Outboard Marine and that would have caused many of the engine failures even if the cables had been manufactured properly. Outboard Marine submitted to the jury a damages calculation of $19,132,302 which assumed that all the engine failures and resulting recalls and replacements had been due to ACCO’s bad cables. ACCO countered with evidence that even if this assumption were correct, Outboard Marine’s damages would be only $12,384,700, because of accounting and other errors that Outboard Marine had made in computing the $19.13 million figure. ACCO went much further, however, arguing to the jury that the plaintiff was actually entitled to only a paltry $303,000 in damages (the cost of the defective cables) because everything above that was due not to the alleged breach of warranty by ACCO but to the defective character of the plaintiff’s design of its new engine. Neither party explained to the jury how to calculate the damages if both a breach of warranty and the plaintiff’s own design defects had contributed to the costs incurred by the plaintiff as a result of the engine failures. Between $303,000 and $12.38 million, the jury was on its own.
Outboard Marine’s case for the entry of judgment of $12,384,700 depends entirely, as its lawyer acknowledged at argument, on the proposition that ACCO had conceded at trial that if there was a breach of warranty the $12.38 million figure was a minimum estimate of the damages for the breach. Such a concession would be tantamount to a stipulation *184as to damages, leaving only the issue of breach to be decided by the jury. Taylor v. Green, 868 F.2d 162, 165 (5th Cir.1989); see also Davis v. United States, 716 F.2d 418, 430 (7th Cir.1983). Only there was no concession. That $12.38 million figure was premised on the assumption, which ACCO did not accept—which it vigorously challenged—that Outboard Marine’s own design defects had not contributed to the engine failures. It was a worst-case analysis: if the jury rejected ACCO’s arguments about the design defects, the maximum damages it should award would be the $12.38 million figure, not the $19.13 million urged by Outboard Marine; but ACCO asked the jury not to reject its arguments about the design defects.
Outboard Marine’s argument for a new trial on damages is stronger than its argument for a judgment, and indeed at first glance compelling. The only estimates tendered to the jury were $303,000, $12.38 million, and $19.13 million. No one knows how the jury came up with $4.4 million. It looks as if the jury started with the $303,000' of conceded damages and then added one-third of $12.38 million (roughly $4.1 million), but this is just a guess; and if this is what the jury was doing, it was wrong, because it should have subtracted the $303,000 from the $12.38 million (which included the $803,000) before taking one-third. Thus, a rational path is not readily discernible between the evidence and argument of the parties and the jury’s verdict. Desperately ACCO argues that the evidence showed that there were three causes of the engine failures and that two of them (two forms of corrosion damage) were the fault of Outboard Marine and only one (the defective cables) was the fault of ACCO, so the jury assigned two-thirds of the liability to Outboard Marine. This wholly speculative reconstruction (replaced at argument by the even more speculative suggestion that the jury actually calculated the damages due to the respective parties’ mistakes) set up ACCO for Outboard Marine’s riposte that, if ACCO is right, the jury was importing tort- notions of comparative fault into a contract case, where such notions do not belong. Outboard Marine goes further. It argues that any attempt to apportion damages between the breach of warranty and its (that is, Outboard Marine’s) own mistakes improperly makes comparative negligence a defense in a contract case; and let us begin our analysis of the issue of a new trial on damages with that question.
Tort and contract law have similar aims, and their doctrines tend therefore to be isomorphic, Evra Corp. v. Swiss Bank Corp., 673 F.2d 951, 957-58 (7th Cir.1982); Wisconsin Knife Works v. National Metal Crafters, 781 F.2d 1280, 1289 (7th Cir.1986), though strict liability plays a bigger role in contract than in tort. Mitigation of damages in contract law, for example, corresponds to avoidable consequences in tort law. Barron v. Ford Motor Co. of Canada Ltd., 965 F.2d 195, 199 (7th Cir.1992); compare Restatement (Second) of Torts § 918 (1979), with Restatement (Second) of Contracts § 350 (1981). Both these doctrines, however, are primarily concerned with efforts by victims of a breach of contract or a tort, exerted after the breach or tort has occurred, to minimize the harm. Failing to fasten your seatbelt—a pre-accident measure to avoid not the aceident but injury if the accident occurs—is an exception in terms of sequence, but it is not, strictly speaking, a form of contributory or comparative negligence. Those are defenses (complete in the case of contributory negligence, partial in the case of comparative negligence) based on the accident victim’s failure to take steps that would have prevented the accident from occurring. A seatbelt defense is based on a failure to prevent an injury resulting from an accident that may have occurred without the slightest fault on the part of the plaintiff. The difference is riot fundamental, and in regard to liability for a defective product that causes personal injury has been pretty much erased—an automobile manufacturer, for example, is equally culpable whether the defect caused the accident or, by making the car less than optimally crashworthy, magnified the consequence of an accident that may have occurred without any fault on the part of the manufacturer. Outboard Marine is correct, however, that there is no contract doctrine of contributory or comparative negligence.
*185There are good-reasons for this absence. The standard tort results from the interaction of two activities (driving and walking, for example), and a change in either one could reduce the probability of the accident giving rise to the tort. The pedestrian can cross the street at the crosswalk, as well as the driver being able to drive more slowly. The typical contract, in contrast, has a performing party and a paying party; these activities proceed on separate tracks; it is rare that a breach comes about as a joint consequence of these qualitatively entirely different activities. Such interdependencies as there are between the duties of the parties to a contract are governed by the doctrine of conditions. Casio, Inc. v. S.M. & R. Co., 755 F.2d 528, 532 (7th Cir.1985); 2 E. Allan Farnsworth, Farnsworth on Contracts § 8.2 (1990). If the payor stops paying, the performing party can stop performing, and vice versa. Given the protection that the doctrine gives each party against a failure by the other, it is difficult to see what room is left for a doctrine of contributory or comparative fault. And a promisor can always condition his promise on the observance by the promisee of precautions designed to head off disaster if the promise is broken—can, that is, impose a duty of care on the promisee beyond what is implicit in the doctrine of conditions. There is no need for the law to read such a duty into every contract—nor can it be assumed that most promisees would willingly agree to such conditions. Often a promisee wants and contracts for protection against his own mistakes. Contracts have an insurance function. People insure themselves against the consequences of their own negligence. Of course, when they do this, there is what is called the “moral hazard” problem; being insured, they don’t have as much incentive to be careful. This problem is controlled in the contract setting by the doctrine of mitigation of damages.
Warranty is a mixed case. It is often an alternative remedy to tort, most clearly in products liability cases. The Uniform Comparative Fault Act (1977) § 1(b), 12.U.L.A. 123,127 (1996), applies to warranty suits, but it excludes purely economic loss and “actions that are fully contractual in their gravamen,” id. at 127-28 (Comment), thus excluding this case. Ethyl Corp. v. BP Performance Polymers, Inc., 33 F.3d 23 (8th Cir.1994); Manning v. International Harvester Co., 381 N.W.2d 376, 378-79 (Iowa App.1985). And anyway the uniform act has not been adopted in Illinois and we are given no basis for thinking that the Illinois courts would adopt a principle of comparative fault in purely contractual warranty cases as a matter of common law.
But it does not follow from the absence of a general duty of care on the part of contract promisees that a promisee’s fault is irrelevant in computing damages. The promisor who breaks his promise is liable only for the harm that he causes, which is to say the harm that would have been avoided had he not broken his promise. Collins v. Reynard, 154 Ill.2d 48, 180 Ill.Dec. 672, 673, 607 N.E.2d 1185, 1186 (1992); O’Neil v. Continental Bank, N.A., 278 Ill.App.3d 327, 214 Ill.Dec. 923, 933, 662 N.E.2d 489, 499 (1996); 3 Farnsworth, supra, § 12.1, p. 148. He is therefore not hable for the harm that occurred solely as a result of the plaintiffs acts. Royal Business Machines, Inc. v. Lorraine Corp., 633 F.2d 34, 46 (7th Cir.1980); Jones v. Melrose Park Nat’l Bank, 228 Ill.App.3d 249, 170 Ill.Dec. 126, 133, 592 N.E.2d 562, 569 (1992); Vesper Construction Co. v. Rain for Rent, Inc., 602 F.2d 238, 242-44 (10th Cir.1979); Signal Oil & Gas Co. v. Universal Oil Products, 572 S.W.2d 320, 328 (Tex.1978); 1 James J. White & Robert S. Summers, Uniform Commercial Code § 6-5, pp. 309-10 (3d ed.1988). Because of Outboard Marine’s own design defects, some of its new engines would have failed (or so the jury could reasonably have found) even if the cables had been manufactured to the contractual specifications, which is to say that some of the harm that Outboard Marine suffered as a result of the engine failures would have occurred even if there had been no breach of warranty. That harm would have to be subtracted from the $12.38 million or $19.13 million in order to determine Outboard Marine’s recoverable damages.
This leaves the question of how the jury arrived at $4.4 million. Outboard Marine insists that it was ACCO’s responsibility *186to furnish the jury with the necessary guidance. This insistence is very odd. The plaintiff rather than the defendant is responsible for quantifying the plaintiffs damages. It is true that a defendant who fails to present a plausible estimate of damages takes a risk, as we have pointed out a number of times. E.g., Kemezy v. Peters, 79 F.3d 33, 36 (7th Cir.1996); Gorlikowski v. Tolbert, 52 F.3d 1439, 1447-48 (7th Cir.1995); cf. East Hampton Dewitt Corp. v. State Farm Mutual Automobile Ins. Co., 490 F.2d 1234, 1245-46 (2d Cir.1973) (Friendly, J.). If, as happened here, the defendant gives the jury one very small damages estimate ($303,000) and one very large one ($12.38 million), he may be worse off than if he had suggested a third, intermediate alternative. But never before have we heard the argument that if the defendant fails to specify an intermediate figure, the jury is debarred from coming up with one on its own unless it is a figure to the plaintiffs liking. Such a rule would reward plaintiffs for adopting the same risky strategy of withholding from the jury an intermediate figure. Outboard Marine could have argued to the jury as easily as ACCO could have done that if the jury thought that both parties’ mistakes had contributed to the harm it should disentangle their respective contributions, using some formula that the lawyer or expert witness for Outboard Marine would explain. We cannot think of any reason why, when both plaintiff and defendant adopt all-or-nothing positions on damages, the plaintiff, who has the burden of persuasion on damages as on liability, should be favored in appellate review of the jury’s stab at an intermediate figure.
That leaves us with the problem that the jury was not given sufficient guidance on how to compute damages if it didn’t want just to pick out one of the three figures the parties had put before it. But we do not think that this problem warrants reversal. The law neither requires juries to state reasons for their verdicts nor permits courts to inquire into the reasoning process of the jurors for the purpose of impeaching their verdict. Fed.R.Evid. 606(b); McDonald v. Pless, 238 U.S. 264, 267, 35 S.Ct. 783, 784, 59 L.Ed. 1300 (1915); Needham v. White Laboratories, Inc., 847 F.2d 355, 360 (7th Cir. 1988). The reasons are various, but among them are that jurors are not experts in articulation and that inarticulate Gestalt judgments arrived at by deliberation may be as reliable as those made algorithmically or otherwise “logically.” If these are bad reasons, which they may be, the civil jury is a bad institution, something that most federal judges do not believe and that few if any are prepared to acknowledge.
It follows, we think, that if a jury verdict is reasonable, the fact that the jury was not given the materials for constructing a rational path from the evidence to the verdict is not fatal. The law applies a bottom-line test: the verdict is evaluated against the evidence, rather than against the likelihood that the jury arrived at the verdict by the “rational” process that a judge would employ. E.g., Wassell v. Adams, 865 F.2d 849, 856 (7th Cir.1989); Dopp v. Pritzker, 38 F.3d 1239, 1249 (1st Cir.1994); East Hampton Dewitt Corp. v. State Farm Mutual Automobile Ins. Co., supra, 490 F.2d at 1246. It is enough (this is implicit in the bottom-fine test) that a rational jury could have come to such a verdict, however improbable that this jury used a rational process; the test is thus like the “rational basis” test in equal protection cases; it is a test of the reasonableness of the result rather than the reasonableness of the process that led to the result. The trial in this case lasted a month, and immersed in the ease as they were the jurors may have come to a reasonable judgment that about two-thirds of Outboard Marine’s damages were due to its own mistakes, rather than being caused by ACCO’s mistakes. Nothing in Outboard Marine’s briefs on appeal suggests that such a conclusion would have been unreasonable in light of the evi-dencé. The trial judge did not consider the verdict unreasonable, and he had the advantage over us of having heard the evidence at first hand and observed the jury. His judgment is therefore entitled to considerable deference. E.g., Gasperini v. Center for Humanities, Inc., — U.S. —, —, 116 S.Ct. 2211, 2225, 135 L.Ed.2d 659 (1996); Webb v. City of Chester, 813 F.2d 824, 836-37. (7th Cir.1987); Ressler v. States Marine Lines, Inc., 517 F.2d 579, 583 (2d Cir.1975). The *187verdict was neither absurdly high nor absurdly low, and the verdict will therefore stand.
AFFIRMED.