Nellcor sells the "Clinivision" line of medical software. Clinivision includes a database system copyrighted by Micro Data Base Systems (MDBS), which licensed Nellcor to use the software in applications written for Microsoft's DOS and Windows operating systems. The license has a most-favored-nations clause: MDBS promised Nellcor that it would charge Nellcor the lowest price any similar reseller pays, but it did not promise to charge any particular price in the future. Nellcor promised to affix stickers (which the parties call "tokens") to every package containing MDBS software. Tokens serve as a metering device, since MDBS otherwise can't tell how many copies of the software Nellcor has distributed. At the outset of their arrangement, MDBS sold three kinds of tokens, with decreasing prices: network-installation tokens, single-user tokens, and upgrade tokens. An upgrade token was to be used only when a customer who had paid for a full license upgraded to a newer version.

In 1996 MDBS filed this suit (which Nellcor removed to federal court under the diversity jurisdiction) seeking to collect unpaid royalties, and late in 1997 MDBS notified Nellcor that its license would be terminated unless it cured all deficiencies. Nellcor has not cured--not, at least, to MDBS's satisfaction--so MDBS deems their relationship over. Nellcor views this as a calamity, because (it informs us) it will take more than a year to incorporate into Clinivision a database manager written by someone other than MDBS. Fearing that the hiatus would put it out of business, Nellcor asked the district court to enjoin the termination. (Actually it asked the court to compel MDBS to accept, as sufficient under the agreement, the royalties Nellcor proposed to tender. The district judge understood this as a request for an anti-termination injunction.) The district court denied the motion, ruling that Nellcor is unlikely to prevail on the merits.

MDBS finds fault with three aspects of Nellcor's performance. First, although MDBS has raised the price of its software, Nellcor has declined to pay more than the price MDBS set in 1993. Second, Nellcor distributed some software to new customers using upgrade tokens, shortchanging MDBS on royalties. Third, without MDBS's permission, Nellcor has incorporated MDBS's software into packages designed for use under IBM's OS/2 operating system. MDBS furnished Nellcor with a development version of the OS/2 product so that Nellcor could begin work on a solution for customers who preferred IBM's operating system over Microsoft's, but Nellcor and MDBS have never signed an agreement permitting Nellcor to distribute a retail version for the OS612 platform. Nellcor contends that a license should be implied from the parties' course of dealing, but even if this is so Nellcor faces a formidable obstacle: it shipped the finished software to its customers without tokens of any kind--indeed without approaching MDBS to purchase tokens--yet denied to MDBS that it was shipping an OS/2 version. Now Nellcor concedes that it should have paid for the software. It wants a judicial order entitling it to continue distributing both the Microsoft-compatible and IBM-compatible versions. The district court held that Nellcor's three failings mean that MDBS is likely to prevail on the merits.

Nellcor weakly argues that even if it has disregarded the terms of its license and distributed software without the consent of the copyright owner, it still cannot be dropped as a distributor until 45 days after a judge finds that it has broken its promises. The contract might conceivably bear that reading. It provides that termination can occur after "any breach which is determined to be the responsibility" of Nellcor, if within 45 days it does not cure the defaults. Written in the passive, this clause does not say who does the determining. Reading it to require a court to make the determination is linguistically possible but would thrust the judge into the role of business manager and saddle firms with uncooperative (even dishonest) business partners for extended periods given the ordinary delay of litigation. Indiana law supplies the rule of decision in this suit, and Nellcor has not cited any Indiana case holding that a termination clause written in the passive transfers to the judicial branch the right and duty to make the initial decision that a party has broken its promise. One case says that a judge's decision is sufficient under such a clause, Medley v. Milwaukee, 969 F.2d 312, 315 (7th Cir.1992) (Wisconsin law), but so far as we can tell no judge in any state has held that a pre-termination judicial determination is necessary when a clause reads this way.

Nellcor contends that this does not matter--that its injury from termination is so great that it is entitled to interim relief even if it is likely to lose in the end. To obtain interlocutory relief "the applicant must show that the probability of success on the merits is sufficiently high--or the injury from the enforcement of the order sufficiently great--to warrant a conclusion that the balance of error costs tilts in favor of relief. In other words, the applicant must show that the costs of false negatives (the costs of denying interlocutory relief, when the applicant ultimately prevails on the merits) exceed the costs of false positives (the costs of granting interim relief, when the applicant eventually loses on the merits)." Illinois Bell Telephone Co. v. WorldCom Technologies, Inc., 157 F.3d 500, 503 (7th Cir.1998). See American Hospital Supply Corp. v. Hospital Products Ltd., 780 F.2d 589, 593-94 (7th Cir.1986); Lawson Products, Inc. v. Avnet, Inc., 782 F.2d 1429, 1433-34 (7th Cir.1986). When irreparable injury is great, the costs of false negatives can exceed the costs of false positives even if the probability of success is low, provided the costs an injunction imposes on the adverse party also are low. An example makes this clear. Suppose the applicant suffers $100,000 in irreparable injury per month while the case awaits decision, the expected time to decision is a year, the party opposed to the injunction would suffer a $5,000 one-time loss if relief were granted, and the applicant's probability of success on the merits is 5%. Then the anticipated error costs of denying relief are $60,000 (the $1.2 million loss times the 0.05 probability that the applicant is entitled to victory), while the error costs of granting relief are only $4,750 (the $5,000 loss times the 0.95 probability that the party opposing relief will succeed). Thus Nellcor is right in theory: a judge sometimes should grant interim relief even to a party that is unlikely to prevail at trial. But theory is one thing and Nellcor's claim another; the theory does it no good.

Although the district judge wrote that Nellcor will suffer substantial and irreparable injury if it desists from using MDBS's software while the case proceeds, there is no concrete evidence that it will do so. After the termination became effective in January 1998, Nellcor went right on including MDBS's software with its Clinivision products. After the district court denied injunctive relief in May 1998, Nellcor continued with this practice. We were informed at oral argument that even today, a year after the termination, Nellcor includes MDBS modules in the Clinivision product. Under the circumstances, the only function of an injunction would be to curtail the damages (and potential criminal liability) Nellcor faces for this unauthorized copying and distribution of MDBS's intellectual property. It would not alleviate potential injury to Nellcor's hospital clients, or alter Nellcor's ability to sell Clinivision products. Nellcor wanted the injunction as a security blanket. By ignoring the district court's denial, Nellcor has effectively announced a heads-I-win-tails-you-lose posture: it is going to distribute MDBS's software no matter what happens in court. We are not about to compel a district court to lend aid to a litigant that seeks judicial assistance while evincing unwillingness to respect an adverse decision.

Nellcor's wilful distribution of MDBS's product after the license ended may boost the damages to which MDBS is entitled, but to say that the stakes can be liquidated is to say that the injury is not irreparable. Indeed, we doubt that business losses from termination of a distribution license are irreparable. Only money is at issue. Courts routinely tote up and award as damages the loss inflicted by wrongful terminations of distributorships. E.g., To-Am Equipment Co. v. Mitsubishi Caterpillar Forklift America, Inc., 152 F.3d 658 (7th Cir.1998). Nellcor has not explained why the usual tools of damages calculation would fail to capture its injury, if MDBS has acted wrongfully.

Suppose nonetheless that injury were impossible to quantify and hence irreparable. This is a contract case, and within broad limits the parties to a contract may specify their remedies. The approach to interlocutory relief sketched above is just the norm courts apply when statutes, rules, and contracts are silent. The Nellcor-MDBS contract is not silent. Nellcor tells us that it needs two years to replace MDBS's database module, and it wants the court to give it this time so that it can sell Clinivision without interruption. But the MDBS-Nellcor contract does not specify a multi-year notice or replacement period. It gives Nellcor 45 days, no more, to cure deficiencies. An injunction affording Nellcor a longer period would not serve as a remedy pending trial; it would instead rewrite the contract to give Nellcor (much) greater rights than those for which it bargained. See generally Industrial Representatives, Inc. v. CP Clare Corp., 74 F.3d 128 (7th Cir.1996). Nellcor might have negotiated for a transition period under which, say, it could distribute Clinivision to established customers until it had incorporated a replacement database package. But the contract says nothing of the kind, and we are unwilling to give one party a boon for which it neither negotiated nor paid. One reason the parties may have settled on days rather than years is that the shorter period gave Nellcor a greater reason to cure deficiencies, and a shorter leash enabled MDBS to trust it more (or charge lower prices) secure in the knowledge that if something went wrong it would not be saddled with an unreliable business partner. That function would be defeated if a judge gave Nellcor two years anyway. Contracts serve commerce best when their terms are enforced rather than twisted after the fact. Nellcor made a business deal that allowed termination for cause on short notice, and it must live with that choice.

AFFIRMED.