POSNER, Circuit Judge.

Jack Rardin, the plaintiff, bought for use in his printing business a used printing press from Whitacre-Sunbelt, Inc. for $47,700. The price included an allowance of $1,200 to cover the cost of dismantling the press for shipment and loading it on a truck at Whitacre's premises in Georgia for transportation to Rardin in Illinois. The contract of sale provided that the press was to be "Sold As Is, Where Is," that payment was to be made before the removal of the press from Whitacre's premises, and that Whitacre was to be responsible only for such damage to the press as might be "incurred by reason of the fault or negligence of [Whitacre's] employees, agents, contractors or representatives." To dismantle and load the press, Whitacre hired T D Machine Handling, Inc., which performed these tasks carelessly; as a result the press was damaged. Not only did Rardin incur costs to repair the press; he also lost profits in his printing business during the time it took to put the press into operating order. He brought this suit against Whitacre, T D, and others; settled with Whitacre; dismissed all the other defendants except T D; and now appeals from the dismissal of his case against T D for failure to state a claim. (The facts we recited are all taken from the complaint.) The only issue is whether Rardin stated a claim against T D under Illinois law, which the parties agree controls this diversity suit.

The contract indemnified Rardin against physical damage to the press caused by the negligence of Whitacre's contractor, T D, and the settlement with Whitacre extinguished Rardin's claim for the cost of repairing the damage. The damages that Rardin seeks from T D are the profits that he lost as a result of the delay in putting the press into operation in his business, a delay caused by T D's negligence in damaging the press. Rardin could not have sought these damages from Whitacre under the warranty, because consequential damages (of which a loss of profits that is due to delay is the classic example) are not recoverable in a breach of contract suit, with exceptions not applicable here. Rardin had no contract with T D, and his claim against T D is a tort claim; consequential damages are the norm in tort law.

We agree with the district judge that Illinois law does not provide a tort remedy in a case such as this. We may put a simpler version of the case, as follows: A takes his watch to a retail store, B, for repair. B sends it out to a watchmaker, C. Through negligence, C damages the watch, and when it is returned to A via B it does not tell time accurately. As a result, A misses an important meeting with his creditors. They petition him into bankruptcy. He loses everything. Can he obtain damages from C, the watchmaker, for the consequences of C's negligence? There is no issue of causation in our hypothetical case; there is none in Rardin's. We may assume that but for C's negligence A would have made the meeting and averted the bankruptcy, just as but for T D's negligence the press would have arrived in working condition. The issue is not causation; it is duty.

The basic reason why no court (we believe) would impose liability on C in a suit by A is that C could not estimate the consequences of his carelessness, ignorant as he was of the circumstances of A, who is B's customer. In principle, it is true, merely to conclude that C was negligent is to affirm that the costs of care to him were less than the costs of his carelessness to all who might be hurt by it; that, essentially, is what negligence means, in Illinois as elsewhere. See McCarty v. Pheasant Run, Inc., 826 F.2d 1554, 1556-57 (7th Cir. 1987). So in a perfect world of rational actors and complete information, and with damages set equal to the plaintiff's injury, there would be no negligence: the costs of negligence would be greater to the defendant than the costs of care and therefore it would never pay to be negligent. And if there were no negligence, the scope of liability for negligence would have no practical significance. But all this is a matter of abstract principle, and it is not realistic to assume that every responsible citizen can and will avoid ever being negligent. In fact, all that taking care does is make it less likely that one will commit a careless act. In deciding how much effort to expend on being careful — and therefore how far to reduce the probability of a careless accident — the potential injurer must have at least a rough idea of the extent of liability. C in our example could not form such an idea. He does not know the circumstances of the myriad owners of watches sent him to repair. He cannot know what costs he will impose if through momentary inattention he negligently damages one of the watches in his charge.

Two further points argue against liability. The first is that A could by his contract with B have protected himself against the consequences of C's negligence. He could have insisted that B guarantee him against all untoward consequences, however remote or difficult to foresee, of a failure to redeliver the watch in working order. The fact that B would in all likelihood refuse to give such a guaranty for a consideration acceptable to A is evidence that liability for all the consequences of every negligent act is not in fact optimal. Second, A could have protected himself not through guarantees but simply by reducing his dependence on his watch. Knowing how important the meeting was he could have left himself a margin for error or consulted another timepiece. Why impose liability for a harm that the victim could easily have prevented himself?

The present case is essentially the same as our hypothetical example. T D is in the business of dismantling and loading printing presses. It is not privy to the circumstances of the owners of those presses. It did not deal directly with the owner, that is, with Rardin. It knew nothing about his business and could not without an inquiry that Rardin would have considered intrusive (indeed bizarre) have determined the financial consequences to Rardin if the press arrived in damaged condition.

The spirit of Hadley v. Baxendale, 9 Ex. 341, 156 Eng.Rep. 145 (1854), still the leading case on the nonrecoverability of consequential damages in breach of contract suits, broods over this case although not cited by either party or by the district court and although the present case is a tort case rather than a contract case. The plaintiffs in Hadley v. Baxendale owned a mill, and the defendants were in business as a common carrier. The defendants agreed to carry the plaintiffs' broken mill shaft to its original manufacturer, who was to make a new shaft using the broken one as a model. The defendants failed to deliver the broken shaft within the time required by the contract. Meanwhile, the plaintiffs, having no spare shaft, had been forced to shut down the mill. The plaintiffs sued the defendants for the profits lost during the additional period the mill remained closed as a result of the defendants' delay in delivering the shaft to the manufacturer. The plaintiffs lost the case. The defendants were not privy to the mill's finances and hence could not form an accurate estimate of how costly delay would be and therefore how much care to take to prevent it. The plaintiffs, however, as the court noted, could have protected themselves from the consequences of a delay by keeping a spare shaft on hand. See 9 Ex. at 355-56, 156 Eng.Rep. at 151. Indeed, simple prudence dictated such a precaution, both because a replacement shaft could not be obtained immediately in any event (it had to be manufactured), and because conditions beyond the defendants' control could easily cause delay in the delivery of a broken shaft to the manufacturer should the shaft ever break. See also EVRA Corp. v. Swiss Bank Corp., 673 F.2d 951, 957 (7th Cir. 1982); Afram Export Corp. v. Metallurgiki Halyps, S.A., 772 F.2d 1358, 1368-69 (7th Cir. 1985). Rardin, too, could have taken measures to protect himself against the financial consequences of unexpected delay. He could have arranged in advance to contract out some of his printing work, he could have bought business insurance, or he could have negotiated for a liquidated-damages clause in his contract with Whitacre that would have compensated him for delay in putting the press into working condition after it arrived.

As we noted in EVRA Corp. v. Swiss Bank Corp., supra, 673 F.2d at 956, Illinois follows Hadley v. Baxendale. So if this were a contract case, Rardin would lose — and this regardless of whether the breach of contract were involuntary or, as he alleges, due to the promisor's negligence. See 673 F.2d at 957; Siegel v. Western Union Tel. Co., 312 Ill.App. 86, 37 N.E.2d 868 (1941). It is a tort case, but so was EVRA, where, applying Illinois law, we concluded that the plaintiff could not recover consequential damages. The plaintiff had instructed its bank to deposit a payment in the bank account of a firm with which the plaintiff had a contract. The bank telexed its correspondent bank in Geneva — which happened to be Swiss Bank Corporation — to make the transaction. As a result of negligence by Swiss Bank, the transaction was not completed, whereupon the plaintiff lost its contract because the other party to it declared a default. The plaintiff sued Swiss Bank for the lost contract profits, and lost. We held that the principle of Hadley v. Baxendale is not limited to cases in which there is privity of contract between the plaintiff and the defendant. Swiss Bank could not have estimated the consequences of its negligence and the plaintiff, like the plaintiffs in Hadley, could have averted disaster by simple precautions. See 673 F.2d at 955-59. This case differs from both Hadley and EVRA in that there is no suggestion that Rardin was imprudent in failing to take precautions against damage or delay. But as in those cases the defendant was not in a position to assess the consequences of its negligence. In this respect the present case and EVRA are actually stronger for defendants even though these are tort rather than contract cases since neither case involves a defendant who is dealing face-to-face with the plaintiff. While it is generally true that consequential damages are recoverable in tort law although not in contract law, EVRA shows that the classification of a case as a tort case or a contract case is not decisive on this question.

We are reinforced in our conclusion that T D is not liable to Rardin by a series of cases — beginning with Moorman Mfg. Co. v. National Tank Co., 91 Ill.2d 69, 61 Ill.Dec. 746, 435 N.E.2d 443 (1982) (decided only a few weeks before EVRA, and not cited to or by us in that case), continuing in cases like Redarowicz v. Ohlendorf, 92 Ill.2d 171, 65 Ill.Dec. 411, 441 N.E.2d 324 (1982), and discussed by this court in Chicago Heights Venture v. Dynamit Nobel of America, Inc., 782 F.2d 723, 726-29 (7th Cir. 1986), and Dundee Cement Co. v. Chemical Laboratories, Inc., 712 F.2d 1166 (7th Cir. 1983) — in which the Supreme Court of Illinois has held that damages for "purely economic loss" cannot be recovered in tort cases. See also Kishwaukee Community Health Services Center v. Hospital Building Equipment Co., 638 F. Supp. 1492, 1495-1504 (N.D.Ill. 1986). The doctrine is not unique to Illinois. Originating in Chief Justice Traynor's opinion in Seely v. White Motor Co., 63 Cal.2d 9, 45 Cal.Rptr. 17, 403 P.2d 145 (1965), it has become the majority rule (see the thorough discussion of the case law in Spring Motors Distributors, Inc. v. Ford Motor Co., 98 N.J. 555, 489 A.2d 660 (1985)), and was adopted as the rule for admiralty as well in East River S.S. Corp. v. Transamerica Delaval Inc., 476 U.S. 858, 106 S.Ct. 2295, 90 L.Ed.2d 865 (1986). We need not consider the outer boundaries of the doctrine; it is enough that it bars liability in a suit for lost profits resulting from negligence in carrying out a commercial undertaking.

The doctrine (called in Illinois the Moorman doctrine) rests on the insight, which is consistent with the analysis in EVRA, that contractual-type limitations on liability may make sense in many tort cases that are not contract cases only because there is no privity of contract between the parties. The contractual linkage between Rardin and T D was indirect but unmistakable, and Rardin could as we have said have protected himself through his contractual arrangements with Whitacre, while there was little that T D could do to shield itself from liability to Whitacre's customer except be more careful — and we have explained why a finding of negligence alone should not expose a defendant to unlimited liability.

The Moorman doctrine goes further than is necessary to resolve this case. Once a case is held to fall within it, the plaintiff has no tort remedy. In our hypothetical case about the watch, the plaintiff could not sue the repairer even for property damage. Anderson Electric, Inc. v. Ledbetter Erection Corp., 115 Ill.2d 146, 104 Ill.Dec. 689, 503 N.E.2d 246 (1986). Moorman itself was a case in which there was a contract between the parties, so there was no reason to allow a tort remedy. The present case, like Anderson, is one where, although there is no contract, the policies that animate the principle which denies recovery of consequential damages in contract cases apply fully and forbid a tort end-run around that principle.

The "economic loss" doctrine of Moorman and of its counterpart cases in other jurisdiction is not the only tort doctrine that limits for-want-of-a-nail-the-kingdom-was-lost liability. It is closely related to the doctrine, thoroughly discussed in Barber Lines A/S v. M/V Donau Maru, 764 F.2d 50 (1st Cir. 1985), that bars recovery for economic loss even if the loss does not arise from a commercial relationship between the parties — even if for example a negligent accident in the Holland Tunnel backs up traffic for hours, imposing cumulatively enormous and readily monetizable costs of delay. See Petition of Kinsman Transit Co., 388 F.2d 821, 825 n. 8 (2d Cir. 1968). Admittedly these doctrines are in tension with other doctrines of tort law that appear to expose the tortfeasor to unlimited liability. One is the principle that allows recovery of full tort law that appear to expose the tortfeasor to unlimited liability. One is the principle that allows recovery of full tort damages in a personal-injury suit for injury resulting from a defective or unreasonably dangerous product — a form of legal action that arises in a contractual setting and indeed originated in suits for breach of warranty. Another is the principle, also of personal-injury law, that the injurer takes his victim as he finds him and is therefore liable for the full extent of the injury even if unforeseeable — even if the person he runs down is Henry Ford and sustains a huge earnings loss, or because of a preexisting injury sustains a much greater loss than the average victim would have done. See, e.g., Stoleson v. United States, 708 F.2d 1217, 1221 (7th Cir. 1983). Both are doctrines of personal-injury law, however, and there are at least three differences between the personal-injury case and the economic-loss case, whether in a stranger or in a contractual setting. The first difference is that the potential variance in liability is larger when the victim of a tort is a business, because businesses vary in their financial magnitude more than individuals do; more precisely, physical capital is more variable than human capital. The second is that many business losses are offset elsewhere in the system: Rardin's competitors undoubtedly picked up much or all of the business he lost as a result of the delay in putting the press into operation, so that his loss overstates the social loss caused by T D's negligence. See Grip-Pak, Inc. v. Illinois Tool Works, Inc., 694 F.2d 466, 473-74 (7th Cir. 1982). Third, tort law is a field largely shaped by the special considerations involved in personal-injury cases, as contract law is not. Tort doctrines are, therefore, prima facie more suitable for the governance of such cases than contract doctrines are.

True, the "thin skull" principle illustrated by Stoleson is sometimes invoked to allow recovery of lost profits in cases where there is physical damage to property. See, e.g., Consolidated Aluminum Corp. v. C.F. Bean Corp., 772 F.2d 1217, 1222-24 (5th Cir. 1985). The thinking here seems to be that the requirement of physical damage at least limits the number of plaintiffs. Rardin could appeal to that principle here — since title had passed to him before T D began dismantling and loading the press — were it not that the Moorman line rejects it, as we explained in Chicago Heights Venture, a case in which there was some property damage. There is also liability (under an exception discussed in Chicago Heights Venture, see 782 F.2d at 726-29) when the injury is the consequence of a sudden, calamitous accident as distinct from a mere failure to perform up to commercial expectations. Consolidated Aluminum illustrates this exception, which is related to the principle that allows consumers injured by a defective product to sue the supplier in tort whether or not the consumer has a contract with him. There are other exceptions. The largest perhaps is the familiar principle that allows a suit for fraud against a person with whom the plaintiff has a contract, while the closest to the present case is the principle that allows a suit against an attorney or accountant for professional malpractice or negligent misrepresentation that causes business losses to the plaintiff. See Rozny v. Marnul, 43 Ill.2d 54, 250 N.E.2d 656 (1969); Pelham v. Griesheimer, 92 Ill.2d 13, 64 Ill.Dec. 544, 440 N.E.2d 96 (1982); Greycas, Inc. v. Proud, 826 F.2d 1560, 1563-65 (7th Cir. 1987). These cases are distinguishable, however, as ones in which the role of the defendant is, precisely, to guarantee the performance of the other party to the plaintiff's contract, usually a seller. The guaranty would be worth little without a remedy, necessarily in tort (or in an expansive interpretation of the doctrine of third-party beneficiaries) against the guarantor.

Although cases barring the recovery, whether under tort or contract law, of consequential damages in contractual settings ordinarily involve smaller potential losses than pure stranger cases do (such as the Lincoln Tunnel hypothetical discussed in Kinsman), this is not always so. In our watch hypothetical, in EVRA, and for all we know in Hadley and in the present case, the financial consequences of a seemingly trivial slip might be enormous. And it is in contractual settings that the potential victim ordinarily is best able to work out alternative protective arrangements and need not rely on tort law. Our conclusion that there is no tort liability in this case does not, therefore, leave buyers in the plaintiff's position remediless. Rardin could have sought guarantees from Whitacre (at a price, of course), but what he could not do was require the tort system to compensate him for business losses occasioned by negligent damage to his property.

A final example will nail the point down. The defendant in H.R. Moch Co. v. Rensselaer Water Co., 247 N.Y. 160, 159 N.E. 896 (1928), had agreed to supply the City of Rensselaer with water of specified pressure for the city's mains. There was a fire, the company was notified but failed to keep up the pressure, and as a result the fire department could not extinguish the fire, which destroyed the plaintiff's building. In a famous opinion denying liability, Chief Judge Cardozo stated that even if the failure of pressure was due to negligence on the defendant's part, the plaintiff could not obtain damages. The city was acting as the agent of its residents in negotiating with the water company, and the water company was entitled to assume that, if it was to be the fire insurer for the city's property, the city would compensate it accordingly. Similarly, in dealing with T D, Whitacre was acting in effect as Rardin's agent, and T D was entitled to assume that, if it was to be an insurer of Rardin's business losses, Whitacre on behalf of Rardin would compensate it accordingly. Rardin in short could protect itself against T D's negligence by negotiating appropriate terms with Whitacre.

The protracted analysis that we have thought necessary to address the parties' contentions underscores the desirability — perhaps urgency — of harmonizing the entire complex and confusing pattern of liability and nonliability for tortious conduct in contractual settings. But that is a task for the Supreme Court of Illinois rather than for us in this diversity case governed by Illinois law. It is enough for us that Illinois law does not permit a tort suit for profits lost as the result of the failure to complete a commercial undertaking.

AFFIRMED.