In a diversity suit for breach of contract governed by Illinois law, the district judge after a bench trial awarded Shannon Has-lund $537,634.41 in damages, plus prejudgment interest, against Simon Property Group (SPG). 284 F.Supp.2d 1102 (N.D.Ill.2003). SPG’s appeal argues that the provision of the contract that it was found to have violated was too indefinite to be enforceable, that no injury was proved, and that in any event no prejudgment interest should have been awarded.

During the dot-com boom of the late 1990s, SPG, a real estate company that operates hundreds of shopping malls, decided to form a subsidiary, “clixnmor-tar.com,” to create Internet-related services ancillary to its mall business. It appointed its chief information officer, Melanie Alshab, to be the president of the new subsidiary. She approached Haslund, a management consultant who had done work for SPG in the past and was employed by Ernst & Young, to be clixnmor-tar’s vice president for operations. Has-lund was interested, but told Alshab that she wanted not only a substantial raise (from $125,000, her salary at Ernst & Young, to $175,000), but also equity in elixnmortar. She was taking a chance by leaving an established firm for a startup, and so she wanted upside potential. She made clear that unless she was given equity she wouldn’t sign on with the new company. Alshab got authorization from her superiors to offer Haslund not only the salary increase that she requested but also one percent of clixnmortar’s equity. The deal was confirmed in a letter to Haslund from SPG’s. director of human resources that under the caption “Annual Salary” recited “$175,000 plus 1% equity in clixn-mortar.com, structure to be determined.”

Haslund started her new job at the end of 1999 shortly after receiving this letter. No stock was issued to her, however, either then or later. She kept badgering SPG for the stock to no avail, and 10 months after starting work she was fired, having denounced SPG’s boss in an email to a firm that was in the process of acquiring an interest in elixnmortar. The start-up never turned a profit — in fact never[*655] had any significant income — and was soon moribund, though it wasn’t dissolved until last year.

The fact that a contract is incomplete, presents interpretive questions, bristles with unresolved contingencies, and in short has as many holes as a Swiss cheese does not make it unenforceable for indefiniteness. Otherwise there would be few enforceable contracts. Complete contingent contracts are impossible. The future, over which contractual performance evolves, is too uncertain. We once decided a case in which the contract exceeded 2000 pages yet the dispute that gave rise to the suit had not been anticipated (or, if anticipated, provided for). S.A. Healy Co. v. Milwaukee Metropolitan Sewerage District, 50 F.3d 476 (7th Cir.1995). If contracting parties had to provide for every contingency that might arise, contract negotiations would be interminable. Contracts can be shorter and simpler and cheaper when courts stand ready to fill gaps and resolve ambiguities in the minority of contracts that get drawn into litigation. Jinwoong, Inc. v. Jinwoong, Inc., 310 F.3d 962, 965 (7th Cir.2002).

But that is in general and not in every case. A contract is rightly deemed unenforceable for indefiniteness when it leaves out (1) a crucial term that (2) a court cannot reasonably be asked to supply in the name of interpretation. Academy Chicago Publishers v. Cheever, 144 Ill.2d 24, 161 Ill.Dec. 335, 578 N.E.2d 981, 984 (1991); Hintz v. Lazams, 58 Ill.App.3d 64, 15 Ill.Dec. 546, 373 N.E.2d 1018, 1020 (1978); Goldstick v. ICM Realty, 788 F.2d 456, 461-62 (7th Cir.1986) (Illinois law); Olympia Equipment Leasing Co. v. Western Union Telegraph Co., 797 F.2d 370, 381 (7th Cir.1986). An example is the contract price. E.g., Tranzact Technologies, Ltd. v. Evergreen Partners, Ltd., 366 F.3d 542, 546 (7th Cir.2004) (Illinois law); Cloud Corp. v. Hasbro, Inc., 314 F.3d 289, 292 (7th Cir.2002) (ditto); Feldman v. Allegheny Int’l, Inc., 850 F.2d 1217, 1223-24 (7th Cir.1988) (ditto); Goldstick v. ICM Realty, supra, 788 F.2d at 461-62. Not only is price central, so that if the choice of price could be delegated to a court it would be the court and not the parties that was the contract maker, but there is no interpretive path that leads from the terms the parties agreed on to the price they would have agreed on. In the division of functions between parties and the judiciary in the joint enterprise of fixing contractual meaning, the selection of the contract price falls clearly on the parties’ side. What is more, the omission of crucial terms is powerful evidence that no contract was intended.

So if the employment agreement had said that Haslund would receive equity but hadn’t indicated how much, a court could not supply the missing percentage. Cf. Architectural Metal Systems, Inc. v. Consolidated Systems, Inc., 58 F.3d 1227, 1229 (7th Cir.1995) (Illinois law); Ray Dancer, Inc. v. DMC Corp., 175 Ill.App.3d 997, 125 Ill.Dec. 447, 530 N.E.2d 605, 610 (1988). No interpretive technique would enable the court to build a bridge between what the parties had agreed to and the percentage of the equity in the new firm that she would receive. But the contract did specify the percentage. What it omitted was a number of details, such as the form of the equity — would it be voting stock or nonvoting stock? — and whether there would be restrictions on vesting: could Haslund show up for work on December 27, 1999, and the next day announce she was quitting and demand her shares? Important details, to be sure; but their absence did not necessarily make the contract indefinite. A court might be able to fill them in without shouldering an inordinate burden of inquiry or creating an inordinate risk of[*656] error. There might for example be a custom in the industry with respect to whether stock in a startup issued to a new employee carries voting rights, whether the right to the stock vests immediately or only after the employee has been on the job for a reasonable period of time, whether the right is forfeited if the employee is fired for cause, and whether and when and to whom the employee can sell the stock once it is issued to him.

Evidence of trade usage is admissible to supply the answers to such questions. E.g., Chicago Bridge & Iron Co. v. Reliance Ins. Co., 46 Ill.2d 522, 264 N.E.2d 134, 139 (1970); Merchants Environmental Industries, Inc. v. SLT Realty Limited Partnership, 314 Ill.App.3d 848, 246 Ill.Dec. 866, 731 N.E.2d 394, 405 (2000); Architectural Metal Systems, Inc. v. Consolidated Systems, Inc., supra, 58 F.3d at 1229. Neither party presented such evidence, but Haslund presented evidence that, if not nearly as good, was nevertheless sufficient. This was Alshab’s testimony that there were no restrictions on the equity interest that Haslund was to receive — that all that “structure to be determined” meant was that SPG had not yet decided on the form that equity in clixnmortar would take, for example the number of shares that would be issued and whether there would be classes of stock, such as voting and nonvoting. Alshab’s testimony was implausible — it seems almost unthinkable that SPG would have agreed to give Haslund an unrestricted one percent interest in a company that SPG hoped not wholly without reason would be highly valuable; for if taken literally the absence of any restrictions would mean that she could quit and sell the stock the day she received it even if she’d been on the job for only five minutes. And Alshab had quit SPG before the trial and there was no love lost between her and her former employer; she may have been inclined to shade her testimony in favor of Haslund, whom after all she had recruited. Nevertheless, her testimony was not so implausible that the district judge’s crediting it over the evasive and ambiguous denials of SPG’s principal, David Simon, could be found to be reversible error.

SPG might have presented evidence, presumably in the form of expert testimony by management consultants knowledgeable about dot-com startups during the boom, that equity in a startup would never be given to a new employee without restrictions, specifically a restriction on vesting and sale. SPG presented no such evidence. It is common knowledge that such restrictions are the norm, but that doesn’t mean that they are universal.

So the contract was enforceable, and was broken, and the next question is whether the district judge was right to award Haslund some half million dollars in damages (or, indeed, as we shall see, whether he was right to award her any amount of damages). The amount he awarded was one percent of his estimate that clixnmortar had a net worth when Haslund was fired of $54 million. The principal evidence on which he relied for this extravagant estimate — for remember that when Haslund was fired (indeed at all times) clixnmortar had not shown a profit, on the basis of which a capital value could be estimated by conventional methods, and apparently had no assets whose value could be appraised — consisted of two transactions. In the first a company named CPG Partners bought a 9.3 percent equity interest in clixnmortar from SPG for $5 million, which implies (arithmetically) a total value of clixnmortar of $54 million. In the second transaction a consulting company named Found.com (the investor to which Haslund had sent the[*657] email that precipitated her being fired) forgave $8.7 million in consulting fees owed it by SPG in exchange for a 6.9 percent equity stake in clixnmortar, implying, by the same arithmetical procedure of dividing the amount of consideration by the percentage of equity received, a similar valuation of the company.

These transactions — though the one with CPG, at least, was rich in the structure missing from Haslund’s one percent interest, since the number of shares and the type of stock were specified, along with restrictions on transfer — constitute but poor evidence of clixnmortar’s value. Compare Ashe v. Sunshine Broadcasting Corp., 90 Ill.App.3d 97, 45 Ill.Dec. 560, 412 N.E.2d 1142, 1145 (1980). The transaction with CPG was devoid of economic substance, because at the same time that CPG paid $5 million for 50,000 shares in clixn-mortar’s common stock SPG paid $5 million for 50,000 shares of the common stock of a subsidiary of CPG — which means that no money changed hands. Probably the aim of the transaction was to make each company seem to be worth $5 million more than it was. But what is certain is that the swap provided no information with regard to the value of either company. And Haslund does not argue that SPG’s accounting maneuver should estop SPG to deny that 9.3 percent of clixnmortar was worth $54 million when the transaction was made.

As for the “investment” by Found, its $3.7 million receivable in uncollected consulting fees was the softest of soft numbers. For, having essentially no assets, liquid or otherwise, clixnmortar could not be expected to cough up $3.7 million in cash for consulting services. Found would have expected to have to write off much of the bill; and so clixnmortar would not have given Found, or Found insisted on receiving, equity actually worth $3.7 million. The stock Found received was worth only the collectable portion of the debt, not its face value, and no evidence concerning that amount was presented.

Even if, as we do not for a moment believe, clixnmortar was worth $54 million when Haslund was fired, and even if we accept, as we must (though reluctantly), Alshab’s testimony that there would have been no restrictions whatever on Has-lund’s selling the stock had she been issued it, and even though Haslund might well have wanted to sell her stock then had it been issued to her, it is beyond unlikely that she could have persuaded anyone to pay her $537,000 for the stock. Even if we accept, as we should not, that CPG and Found really did invest in a meaningful sense in clixnmortar, there is a big difference between putting money into a startup and buying stock in a startup from another investor. In the first case the investor is strengthening the startup by his investment and thus increasing the likelihood of its succeeding. In the second case he is contributing nothing to the company unless the purchase is big enough to affect the price at which the company might issue new stock. The district judge speculated that CPG or Found might have bought Haslund’s stock. But this is hopelessly speculative and utterly implausible — CPG, remember, had not contributed a nickel to clixnmortar, while Found had contributed only a wooden nickel. Neither company had indicated any willingness actually to invest cash in the startup.

All that Haslund has going for her on the proof-of-damages front is SPG’s failure to propose an alternative damages figure other than zero. As we have remarked in the past, this is a risky strategy for defendants. Avitia v. Metropolitan Club of Chicago, Inc., 49 F.3d 1219, 1230 (7th Cir.1995); AMPAT/Midwest, Inc. v. Illinois Tool Works Inc., 896 F.2d 1035,[*658] 1046 (7th Cir.1990); Lancaster v. Norfolk & Western Ry., 773 F.2d 807, 823 (7th Cir.1985). It turns the damages phase of a litigation into a simulacrum of final-offer arbitration, where each party submits a figure to the arbitrator and he must choose one of the figures — he cannot pick a number in between. When a plaintiff asks for damages in a specific amount and the defendant ripostes that the plaintiffs damages are zero, period, and no evidence is presented that would support an intermediate figure, the trier of fact has to decide whether it is more likely that the plaintiffs figure (or the plaintiffs minimum figure, if the plaintiff has proposed a minimum figure — for the strategy of proposing only a maximum figure is as risky for a plaintiff as the strategy of proposing only zero is for a defendant) is correct or that zero is correct, because he is given no basis for picking an intermediate figure. If the plaintiff testifies that his damages were $100 million and the defendant that they were $0, the trier of fact cannot just throw a dart to determine the amount of damages to award within that range; yet that is what he would be doing were there no basis in the evidence for estimating the damages. Assessing damages is often and permissibly speculative, but only within limits. A “plaintiff has the burden of proving damages to a reasonable degree of certainty.” Williams v. Board of Education, 52 Ill.App.3d 328, 10 Ill.Dec. 161, 367 N.E.2d 549, 553 (1977); see also Bigelow v. RKO Radio Pictures, Inc., 327 U.S. 251, 264-66, 66 S.Ct. 574, 90 L.Ed. 652 (1946); BE & K Construction Co. v. Will & Grundy Counties Building Trades Council, 156 F.3d 756, 770 (7th Cir.1998); Sir Speedy, Inc. v. L & P Graphics, Inc., 957 F.2d 1033, 1038 (2d Cir.1992).

On this record, we have to say that it is more likely that zero is correct than that $537,000 is, and we are sufficiently confident about this conclusion to pronounce the district judge’s contrary finding clearly erroneous and therefore not binding on us. There is no evidence that there would have been a market for Has-lund’s equity interest, had it been given to her in accordance with the contract, during the limited period between when she was fired and when it became apparent that clixnmortar would never leave the starting gate. The “investments” by CPG and Found are no evidence at all that there would have been a willing buyer for Has-lund’s stock even in the unlikely event that the stock would have been completely unrestricted.

She argues that by breaking the contract and never issuing her the stock, SPG prevented her from testing the market; and it is true that when a defendant by violating the plaintiffs rights makes it difficult for her to prove her damages, all reasonable doubts about the amount of damages are resolved in her favor. Bigelow v. RKO Radio Pictures, Inc., supra, 327 U.S. at 264-66, 66 S.Ct. 574; BE & K Construction Co. v. Will & Grundy Counties Building Trades Council, supra, 156 F.3d at 770; Bailey v. Meister Brau, Inc., 535 F.2d 982, 991 (7th Cir.1976). But this rule does not apply to the threshold issue of injury. Once the plaintiff proves injury, broad latitude is allowed in quantifying damages, especially when the defendant’s own conduct impedes quantification. But the injury itself must be proved in the usual way, without speculation or burden shifting. Story Parchment Co. v. Paterson Parchment Paper Co., 282 U.S. 555, 562-63, 51 S.Ct. 248, 75 L.Ed. 544 (1931); Fishman v. Estate of Wirtz, 807 F.2d 520, 550-51 (7th Cir.1986); In re Lower Lake Erie Iron Ore Antitrust Litigation, 998 F.2d 1144, 1176 (3d Cir.1993). Haslund failed to prove that she was injured by the breach. So far as appears, the stock to[*659] which she was entitled would have been worthless in her hands.

But, surely, it will be replied, she could have gotten something for her equity interest — if only a few hundred dollars. No doubt. But if that is all she could have gotten, she would have held on to the stock in the hope that clixnmortar would succeed — in which event she would have been holding fairy dust when the company failed. Although SPG’s breach of her employment contract appears to have been deliberate and indeed reprehensible, without proof of actual loss Haslund is entitled only to nominal damages, Kleinworb Benson North America, Inc. v. Quantum Financial Services, Inc., 285 Ill.App.3d 201, 220 Ill.Dec. 457, 673 N.E.2d 369, 378 (1996); Movitz v. First Nat’l Bank of Chicago, 148 F.3d 760, 765 (7th Cir.1998) (Illinois law) — which eliminates any entitlement to prejudgment interest, as well.

The judgment is reversed with instructions to enter judgment for the plaintiff for nominal damages only.

REVERSED AND REMANDED, WlTH DIRECTIONS.