EASTERBROOK, Circuit Judge.

During the early nineteenth century anyone shorter than 1.57 meters was immune from conscription into the French army. Adolphe Quetelet discovered that by the army's measurements there was a statistical deficit of males just over that height and a corresponding surplus of persons just under it. Quetelet inferred that 2,200 of each 100,000 persons either had shrunk themselves or had bribed the measurer to report their height falsely. Letters a S.A.R. le Duc Regnant de Saxe-Cobourg et Gotha sur la Theorie des Probabilites, appliquee aux sciences morales et politiques 144-45, 401-03 (1846). See Stephen M. Stigler, The History of Statistics 215-18 (1986). This is one of the best examples of detecting crime by statistical techniques — although statistics did not reveal which 2.2% of the population was culpable and who was really shorter than 1.57 meters.

More than a hundred years later Amoco Oil Company used a similar statistical method to conclude that Terry J. Baker was cheating on the price of gasoline — although, like Quetelet, Amoco used pencil and paper rather than computers to work it out. Amoco's retail dealers can buy gas by the tank truck, paying the price on the day of delivery. Dealers seeking to conserve on cash may elect a different method: Amoco fills the station's underground tanks with gasoline, which dealers buy as it passes through the pump to the customer's car. This latter method poses a problem if Amoco changes the wholesale price after it has delivered the gasoline to the underground tank. To learn how much gas is withdrawn while each price is in effect, Amoco requires the dealers on its "Meter Marketing Plan" to take daily inventories. The difference between that day's inventory and the prior day's — reflected in the change of the meter in each pump — is the amount the dealer has bought and must pay for at the price in force that day. Dealers agree by contract to read the meter on each pump every day and to report these readings (and pay for the gas) on a schedule Amoco sets, usually two or three times a week.

You don't have to be a genius to see how to manipulate these reports. If Amoco raises the price effective Wednesday, then the dealer reports meter readings as of the close of business Tuesday greater than those actually on the pump. The report of increased sales means that the dealer buys additional gas on Tuesday, at the lower price. This is called "reading ahead." (The dealer reports a figure "ahead" of the meter.) If Amoco reduces the price effective Saturday, then at the close of business Friday the dealer reports readings less than those shown on the pump. This tactic, called "reading back" (reporting less than the figure on the meter), cuts down the volume of gas paid for at the higher price. Reading ahead on price increases combined with reading behind on price reductions shaves a little off the average price the dealer pays Amoco for each gallon.

During the afternoon of December 18, 1989, Amoco called Baker to announce a price increase of 4 ¢ per gallon for regular unleaded gas effective at 12:01 AM on December 19. Dianna Baker, Terry Baker's wife and the person responsible for sending reports on the volume of gas sold, filled out on December 19 forms purporting to state the readings as of that time. On December 21 Amoco conducted an audit of the sales at one of Baker's four stations. Terry Baker refused to admit the auditor to the office or show him the sheets containing the meter readings for earlier days (although Dianna Baker did copy some of the numbers onto a separate page, by her account "correcting an error" in the records and by Amoco's lying about sales; we need not resolve the conflict). The auditor recorded the numbers on the pumps' meters. Later comparison with the reports Dianna Baker had made showed that five of the readings for 12:01 AM on December 19 exceeded the readings at mid-day of December 21.

Amoco suspected that the Bakers were "reading ahead," buying 4,000 gallons of extra gas before the price increase. The Bakers told Amoco that the discrepancy was attributable to bad weather, which led Mrs. Baker to start from readings early on December 18 and estimate later sales rather than read the meters directly at the appropriate time. According to the Bakers, the same bad weather depressed sales December 18-20, producing the embarrassment on December 21. As Dianna Baker's reports had not suggested that the readings were estimated rather than exact, and the Bakers had not submitted precise readings when the weather improved, Amoco decided to take a closer look at the Bakers' other reports. Peter Feeney, one of Amoco's territory managers, pulled the reports for October through December 1989 from a warehouse. (Apparently Amoco does not enter the daily pump readings into its computer system, although that would assist in catching arithmetical errors as well as in detecting evasion.)

Feeney discovered a striking pattern: the volume at all of the Bakers' stations experienced startling fluctuations around the time of price changes. During November 1989 one of the Bakers' stations sold an average of 3,250 gallons of regular leadfree gas per day. On November 11 the wholesale price rose from 60.8 ¢ to 62.8 ¢; the Bakers reported sales of 6,375 gallons per day in the two preceding days, and 1,100 gallons per day in the three following days. On November 21 the price dropped from 60.3 ¢ per gallon to 59.3 ¢ per gallon; the Bakers reported sales of 2,600 gallons per day in the three preceding days and 6,100 gallons per day in the two following days. The same pattern appears at all four stations for all three months. Although the Bakers insist that these fluctuations are attributable to chance (for example, one of the stations is near a sports stadium, so sales fluctuate with the timing of contests), luck produces a normal distribution of errors. That is, bad weather or a big game is as likely to occur before a price change as after. Yet the Bakers' reports routinely show higher sales before a price increase than after, and higher sales after a price reduction than before — by impressive amounts. Desktop computers and statistical packages enabled us to perform a quick analysis on the numbers. The probability that the observed fluctuations are attributable to random events such as the weather is minuscule. If Amoco had its dealers' daily pump readings in a computer database, a statistical program could catch any pattern of evasion in a few moments of computing time. Instead of sending auditors with clipboards, Amoco could examine the behavior of every dealer in its nationwide network overnight.

Amoco wanted nothing more to do with the Bakers and gave notice under their contracts. Terry Baker filed this suit, contending that the termination violated the Petroleum Marketing Practices Act, 15 U.S.C. § 2801-06. After first holding that litigants are entitled to jury trial in actions under the Act, 751 F. Supp. 1357 (E.D.Wis. 1990), and that the federal statute preempts any state rules concerning franchises in the petroleum business, the court granted summary judgment for Amoco, 761 F. Supp. 1386 (1991). Baker conceded that the report for December 19 overstated actual sales but contended that inclement weather and Amoco's practice of allowing dealers to select the time of day at which they read the pumps accounted for the error. The district judge thought these explanations sufficient to create a material dispute about whether the Bakers committed "fraud," which under both contract and § 2802(b)(2)(C) and (c)(1) is a ground for termination. But the court held that there was no disputed factual issue material to § 2802(b)(2)(A), which provides that the oil company may break off dealings for "failure by the franchisee to comply with any provision of the franchise, which provision is both reasonable and of material significance to the franchise relationship". Baker had not reported accurately the volume of gas sold, which violated the contract directly and as "misrepresentation"; both allow termination, the court concluded. See also § 2802(c)(8), listing failure to pay as the sort of violation for which termination is appropriate. Cf. Thompson v. Amoco Oil Co., 903 F.2d 1118, 1121-22 (7th Cir. 1990); Moody v. Amoco Oil Co., 734 F.2d 1200, 1217 (7th Cir. 1984); Brach v. Amoco Oil Co., 677 F.2d 1213, 1221 (7th Cir. 1982).

Baker insists that there is no difference between "fraud" and "misrepresentation"; both have a mental element, which must be resolved at trial. Perhaps so, if we like the district court were to limit attention to the events of December 18-21. Yet this was only the first of six specifications Amoco laid against the Bakers. The second reads: "You have failed to properly report gasoline sales required by your agreements with Amoco." No material dispute impedes immediate judgment on this charge. Terry Baker conceded in his deposition that for years he had reported meter readings within 24 hours one way or the other of price changes rather than at the time of the change, and Amoco's charts are smashing evidence that his decisions invariably worked to his financial advantage. The contract between Baker and Amoco requires accurate daily readings. Whatever dispute there may be about the time of day at which the reading must be taken (Amoco insists that it is the close of business on the day before a price change, or 12:01 AM for stations open 24 hours, while Baker says that the contract allows him to choose the time within a 24-hour grace period) does not explain even a small part of the discrepancies Amoco uncovered. Baker's explanation boils down to (a) "everyone does it", and (b) Amoco condones it. Why pick on me?, Baker asks. He surmises that Amoco is settling scores for an earlier dispute that the parties compromised.

To show that "everybody does it" Baker offered the deposition testimony of two out of the eighty-five Amoco dealers in the Milwaukee area. These two stated that they misstate the meter readings in order to be more competitive with dealers of rival brands, and that they believe Amoco accepts this practice. To show that Amoco indeed condones it, Baker offered the deposition testimony of John Sobczak, formerly an assistant to one of Amoco's district managers, who related that he told dealers they could read ahead or back when competition was fierce.

Holes aplenty dot Baker's presentation. Among the missing: any language in the contracts creating ambiguity on this score and testimony that Amoco ever told Baker that reading ahead or back was OK (or even that there was a 24-hour leeway in the time to report readings). In 1983 Robert Koblewski, one of Amoco's territory managers, sent Baker a letter repeating advice given orally: that Baker's reporting methods violated the contract and he was at risk of termination. Sobczak left Amoco in April 1980 and conceded that he has no knowledge of Amoco's practices since then. All in all, Baker's evidence sounds like the plea of a cashier that "everyone knows" it is OK to take small amounts from the till, and that hourly wages have been reduced so that everything comes out right after the "withdrawals." Such a plea would not stop the employer from firing one caught in the act. One employee's "understanding," even the "understanding" of all employees collectively, cannot prevent the employer from making and enforcing rules. So here — and this even if the price of gasoline has been increased a little to make up for the loss created by dealers who read ahead and back. Higher prices fall on the dealers who report accurately, and the Amoco network will be more competitive if all dealers adhere to the rules. Amoco sets the rules, which the dealers cannot vary by establishing an underground network of contrary beliefs and self-serving "understandings."

Nothing in this record prevents Amoco from enforcing its contracts. Baker has conceded making inaccurate reports of gasoline sales. In the terms of the Act, the requirement of accurate reporting is "reasonable", and repeated violations are "of material significance to the franchise relationship". Cheating by even a small amount on a contract is significant. Nothing in the Act requires a wholesaler to show a pattern as a condition to cessation of dealings. Amoco did show a pattern with respect to Baker, and it establishes that the termination reflected an ongoing violation of contract rather than the spillover of some earlier contretemps. That one reason for action (the earlier dispute) is off limits does not imply that other, valid, reasons must be disregarded. Gruber v. Mobil Oil Corp., 570 F. Supp. 1088, 1096 (E.D.Mich. 1983). It is enough if the valid reason standing alone would support the decision. Cf. Mt. Healthy City School District Board of Education v. Doyle, 429 U.S. 274, 285-87, 97 S.Ct. 568, 575-76, 50 L.Ed.2d 471 (1977). Amoco need not furnish gasoline to a person who cannot be relied on to account for the sales accurately.

AFFIRMED.