EASTERBROOK, Circuit Judge.

Two manufacturers of car polish have moved their competition from the market to the courtroom. Reed-Union makes NU FINISH, "The Once a Year Car Polish(TM)", which it introduced in 1975. By 1990 NU FINISH was the No. 1 seller, which did not please Turtle Wax, a firm accustomed to success. Early in 1991 Turtle Wax introduced FINISH 2001, whose label boasts: "One Application Lasts One Year." Its ad campaign tracked that for NU FINISH, including a TV commercial that shows a beat-up car revived with new polish, which then lasts through a year of car washes. By 1992 FINISH 2001 was the No. 1 brand. Instead of making changes in its product, promotion (Turtle Wax wooed distributors to secure prominent displays for FINISH 2001), or price (FINISH 2001 costs about $1 per bottle less than NU FINISH), Reed-Union made a beeline for the courthouse.

The district court granted summary judgment to Turtle Wax on the copyright infringement claim, ruling that Reed-Union's commercial is a knockoff of a similar work by Star Brite Distributing, another producer of durable auto finishes, and that its central elements — laboratory test results, the revival of a weatherbeaten car, the imperviousness of the polish to many car washes — are scenes a faire outside the scope of copyright protection. 1993 U.S. Dist. LEXIS 16854 (N.D. Ill.). The judge also rejected contentions that the blue ribbon on bottles of FINISH 2001, and the one-year longevity claim, violate the Lanham Act; an argument based on the Illinois Anti-Dilution law bit the dust at the same time. That left Reed-Union's principal contention: that the shape and color of the FINISH 2001 bottle, and the name of the product, violate Reed-Union's trademarks and trade dress. After a bench trial the district judge concluded that the NU FINISH mark is descriptive — "new finish" is the appearance of a car treated with the product — and therefore does not prevent rivals from combining the word "finish" with a symbol such as "2001" that implies technological progress (or, perhaps, "newer"). 1994 U.S. Dist. LEXIS 16352. As for trade dress: the products' bottles have similar shapes, but NU FINISH comes in da-glo orange, while FINISH 2001 is a fluorescent lime green. These vibrant colors make the packages distinctive. The FINISH 2001 label is black with orange-red text, close enough to NU FINISH's matte orange label to trouble the judge. But the judge concluded that consumers could tell the difference and attribute each to its maker. The record contains no evidence of actual confusion — the court disbelieved anecdotes offered by Reed-Union — and little evidence that buyers are likely to be confused. Finally, in a brief supplemental order, the court declared Reed-Union's copyright in its own commercial unenforceable, and it directed the Commissioner of Patents and Trademarks to cancel the registration for "The Once a Year Car Polish". Turtle Wax thus has prevailed across the board.

Reed-Union's brief, a model of bad appellate advocacy, presents 12 issues for review — many with sub-parts, for a total of 21 principal contentions. Posing so many issues ensures that each is superficially argued. To weaken its position still further, Reed-Union concentrates its fire on the findings the district judge made at the conclusion of the bench trial. For example, issue No. 8 reads: "Did the lower court err by arbitrarily giving no credibility to testimony regarding actual confusion incidents?" But unless a judge believes physically impossible things, or disbelieves testimony supported by unrefuted documents, the "finding, if not inconsistent, can virtually never be clear error." Anderson v. Bessemer City, 470 U.S. 564, 575 (1985). The district court thought he had heard a series of tall tales. Here is one example:

[One witness related] the story of an elderly gentleman in Massachusetts who approached a Reed Union sales representative at a gas station to compliment him in the shine of his car. The Reed Union representative said, "Why I use NU FINISH," to which the old gentleman responded, "I do too, but I don't get the same shine." The old gentleman then went to the trunk of his car where he produced FINISH 2001. This story might be a fine commercial, but it is inherently unbelievable to this Court.

The principal issue in a case such as this is whether consumers are likely to be confused about the identity or source of the products in the marketplace. The district court did not commit clear error in finding that there is no credible evidence of actual confusion. What of future buyers? Likelihood of confusion, too, is an issue of fact, with deferential appellate review. Scandia Down Corp. v. Euroquilt, Inc., 772 F.2d 1423, 1428-29 (7th Cir. 1985). Confusion is possible: the products' names are similar, the claims on bottles and in ads are all but identical, a bottle sells for less than $10 (so consumers may not devote much time to selection), and so on. One can reply that the colors of the bottles differ and that consumers pay extra attention because the product is used to preserve the value of an expensive auto (so there may be very large consumer surplus, which concentrates the mind). Which hypothesis is correct? Because confusion is a factual matter, the plaintiff must produce proof; a theory about how consumers might be confused will not do, unless evidence supports the theory. Libman Co. v. Vining Industries, Inc., No. 95-1905 (7th Cir. Nov. 16, 1995), slip op. 5; August Storck K.G. v. Nabisco, Inc., 59 F.3d 616, 618-19 (7th Cir. 1995). Because the party complaining about infringement bears this factual burden, scepticism or disbelief by the trier of fact is fatal. And the district judge, as trier of fact, disbelieved Reed-Union's proof — not only its anecdotal evidence of actual confusion but also its expert evidence of the potential for confusion in the future.

Each side engaged a marketing expert, who conducted a survey. Each survey concluded that some consumers who set out to buy one product would end up with the other, because of confusion; each survey concluded that some consumers who wanted to buy a product made by one firm would end up with a product made by the other. Each, indeed, measured the error rate at approximately 25 percent. Is 25 percent large or small? Compared to what? Befuddlement is part of the human condition. No matter how clear the markings, no matter how different the names, no matter how distinctive the bottles, some confusion is inevitable. Each expert therefore set out to measure the amount of "natural" confusion among consumers. Turtle Wax would be in trouble only if the confusion between NU FINISH and FINISH 2001 exceeded that base, and by a significant amount — significant both in the statistical sense (so that we are confident that a rerun of the tests would show confusion) and in the legal sense (because the trademark laws do not forbid minuscule effects on confusion, even if they can be measured reliably). Cf. Gammon v. GC Services Limited Partnership, 27 F.3d 1254 (7th Cir. 1994) (concurring opinion).

To estimate the "normal" rates of confusion, each expert used a different pair of products. Reed-Union's expert used ARMOR ALL car wax as the control product and found little natural confusion. Turtle Wax's expert selected PRISM wax as the control and found confusion exceeding 20 percent, implying that the name and packaging of FINISH 2001 could not be faulted. The district judge preferred the latter study, because ARMOR ALL bore the name of the manufacturer prominently on the front of the bottle, while both NU FINISH and FINISH 2001 carry the manufacturer's name inconspicuously on the back, as does PRISM. ARMOR ALL is a well known brand name, which reduces confusion. In other words, the court concluded that Reed-Union's expert introduced a bias into the study, producing an artificially low estimate of the normal degree of confusion affecting the purchase of products of this kind. Reed-Union argues that the district judge should have preferred its study, whose author, Reed-Union believes, had better credentials. But the reasons the judge gave are sound — the question is what the author did in this litigation, not where he got his degree or where he teaches — and it is impossible to describe his conclusion as a clear error.

Given this judgment about the ultimate question in the case, most remaining issues pale in significance. We bypass all other objections to the conduct of the trial and the district court's conclusions — not only because they do not affect the outcome but also because we are not persuaded that the district judge erred at any of these intermediate steps. So, too, we trip lightly past Reed-Union's argument that the district court should not have removed "The Once a Year Car Polish" from the register of trademarks under 15 U.S.C. § 1119. The district judge concluded that this "phrase defines a central characteristic of [one class of] car waxes in general" and therefore cannot be a trademark. This is a problematic conclusion. There are thousands of ways to express the proposition that a polish lasts a long time, and the use of one of them as a trademark does not prevent other firms from expressing the same thought; Reed-Union does not object to "One Application Lasts One Year" on the bottle of FINISH 2001. See Sunmark, Inc. v. Ocean Spray Cranberries, Inc., 64 F.3d 1055, 1058-59 (7th Cir. 1995). Cf. Qualitex Co. v. Jacobson Products Co., ___ U.S. ___, 115 S.Ct. 1300 (1995) (holding that a color may be a trademark of a particular product, in part because there are so many other options in the spectrum). But Reed-Union devotes only one page in its brief to this subject. Its presentation is so sketchy that we deem the matter waived. See Tyler v. Runyon, 70 F.3d 458, 464-65 (7th Cir. 1995).

Copyright is the most difficult part of this case. Reed-Union's "Junkyard" commercial shows the application of NU FINISH to restore the finish of a battered car in a junkyard. The announcer claims that the product is not a wax, that it lasts a year, that it is easy to apply, and that it can be used on boats. The commercial concludes with the passage of the car through 52 washes, the finish seemingly unaffected, and a claim that independent laboratory tests show its durability. Turtle Wax promoted FINISH 2001 with its "Revolutionary" commercial showing an old car being restored to luster. The car is not in a junkyard, but spare parts are strewn about. The announcer trumpets that FINISH 2001 does not require rubbing or buffing and that an independent lab has verified that it lasts a year. Like the Junkyard commercial, the Revolutionary commercial makes a money-back guarantee. Another Turtle Wax commercial, called "The Mission," shows before-and-after scenes of cars polished with FINISH 2001; one car then goes through 104 washes, with no discernible effect on the gleaming surface.

In concluding that Turtle Wax was entitled to use the Revolutionary and Mission commercials, the district court advanced two reasons. The first, which the judge expressed when ruling on summary judgment, is that Turtle Wax used only hackneyed ideas — polishing up old cars, washing the polished cars, claiming the support of lab tests. The copyright laws do not permit anyone to claim ownership of ideas or common situations. Holmes v. Hurst, 174 U.S. 82 (1899); Nash v. CBS, Inc., 899 F.2d 1537 (7th Cir. 1990). And situations sometimes imply the rudiments of expression; this is the idea behind the scenes a faire doctrine. Olson v. National Broadcasting Co., 855 F.2d 1446, 1453 (9th Cir. 1988). The second, which the judge gave when specifying the terms of the final judgment, is that the Junkyard commercial is not even copyrightable — perhaps because no work assembled from scenes a faire can be copyrightable, perhaps because of copyright misuse. The district court's brief order suggests both things, but it does not explain why one cannot build an original work out of public-domain parts, see Rockford Map Publishers, Inc. v. Directory Service Co., 768 F.2d 145 (7th Cir. 1985) (maps are copyrightable), what the "misuse" consists in, or why "copyright misuse" invalidates the copyright itself. Misuse of copyright in pursuit of an anticompetitive end may be a defense to a suit for infringement, along the lines of the patent-misuse doctrine in antitrust. See Paul Goldstein, 2 Copyright § 9.6.1 (1989). We do not say that it is (an open issue in this court); copyrights do not exclude independent expression and therefore create less market power than patents. At all events, general allegations of "misuse" — here, apparently, Reed-Union's attempt to invoke the copyright law against Turtle Wax for doing no more than Reed-Union had done to Star Brite — does not make the copyright vanish. The difference is important. If, as the district court's final order says, "Reed-Union's copyright registrations on the junkyard commercial are declared unenforceable against Turtle Wax", then Turtle Wax can show the Junkyard commercial itself, dubbing FINISH 2001 in place of NU FINISH whenever the product appears. Application of the scenes a faire doctrine, or a copyright misuse defense, enables Turtle Wax to show its own commercials but does not authorize it (or anyone else) to appropriate the Junkyard commercial.

The difference between the scenes a faire doctrine and copyright invalidity is vital to maintain. Take Shakespeare's Romeo and Juliet, the inspiration of many plays, musicals, and operas, including Bellini's I Capuletti e i Montecchi and Bernstein's West Side Story. The core scenes of Romeo and Juliet — the war between rival clans, developing love between children of the clans, death and reconciliation after the rejection of revenge — are in the public domain (and not only because Shakespeare borrowed freely himself). Bernstein could incorporate them into his own work. But this did not place West Side Story itself in the public domain, for it has exposition, language, and music different from Shakespeare's. Bernstein could (and did) copyright his own contributions; no one can copy or perform West Side Story without permission, although anyone can use similar scenes. So too with performance. Romeo and Juliet has been produced with and without music, in ancient and modern costumes, with voice and as ballet, and with many different casts. If James Cameron today made a movie of Romeo and Juliet in a reconstruction of the Globe Theatre, with Arnold Schwarzenegger as Romeo and Jamie Lee Curtis as Juliet, he could get a copyright. No one could copy or exhibit the film without permission, for the characterizations, movements, and vocal inflections would be original matter protected by copyright; but this would not curtail anyone's rights to stage or film a different production of the play. Situation comedies, soap operas, romance novels built on love triangles, these and more are cobbled together from staple themes; although the themes are not copyrightable, a unique combination of them is. Just so with maps, which can be copyrighted even though they depict a street layout that the mapmaker did not create, and with photographs. Ansel Adams published multiple views of El Capitan in Yosemite National Park, in different seasons and lighting. He did not create the mountain, the park, the seasons, or the lighting, but his expression of those conditions is an artistic achievement. Anyone can take and sell a photo of El Capitan, but Ansel Adams' photos are protected by copyright.

If, as the district court held, the Junkyard commercial replows ground cleared by the Star Brite commercial, then Turtle Wax is free to use the same scenes and incidents. Turtle Wax can copy Star Brite's work without protest from Reed-Union; and the common use of similar scenes shows that the fundamental ideas of the production are not copyrightable. But each expression of a theme can claim protection to the extent it differs from its predecessors. "[T]here will typically be expression in the treatment of stock scenes, and . . . it will be a rare work that consists entirely of scenes a faire." William F. Patry, 1 Copyright Law Practice 338 (1994). The Junkyard commercial is not that "rare work"; the expression of the stock scenes it contains is copyrightable, although the themes are not. We therefore think that the district judge was right the first time and wrong the second: the Revolutionary and Mission commercials do not infringe Reed-Union's copyright in the Junkyard commercial, but Reed-Union retains a valid copyright, which it may enforce against aping of the expression that distinguishes the Junkyard commercial from other examples of the genre.

The judgment on the copyright claim is vacated, and the case is remanded for the entry of a new judgment consistent with this opinion. The remainder of the judgment is affirmed.