In the six years since they filed this antitrust suit, the Chicago Bulls have won four National Basketball Association titles and an equal number of legal victories. Suit and titles are connected. The Bulls want to broadcast more of their games over WGN television, a “superstation” carried on cable systems nationwide. The Bulls’ popularity makes WGN attractive to these cable systems; the large audience makes WGN attractive to the Bulls. Since 1991 the Bulls and WGN have been authorized by injunction to broadcast 25 or 30 games per year. 754 F.Supp. 1336 (1991). We affirmed that injunction in 1992, see 961 F.2d 667, and the district court proceeded to determine whether WGN could carry even more games — and whether the NBA could impose a “tax” on the games broadcast to a national audience, for which other superstations have paid a pretty penny to the league. After holding a nine-week trial and receiving 512 stipulations of fact, the district court made a 30-game allowance permanent, 874 F.Supp. 844 (1995), and held the NBA’s fee excessive, 1995-2 Trade Cas. para. 71,253. Both sides appeal. The Bulls want to broadcast 41 games per year over WGN; the NBA contends that the antitrust laws allow it to fix a lower number (15 or 20) and to collect the tax it proposed. With apologies to both sides, we conclude that they must suffer through still more litigation.
Our 1992 opinion rejected the league’s defense based on the Sports Broadcasting Act, 15 U.S.C. §§ 1291-95, but our rationale implied that the NBA could restructure its contracts to take advantage of that statute. 961 F.2d at 670-72. In 1993 the league tried to do so, signing a contract that transfers all broadcast rights to the National Broadcasting Company. NBC shows only 26 games during the regular season, however, and the network contract allows the league and its teams to permit telecasts at other times. Every team received the right to broadcast all 82 of its regular-season games (41 over the air, 41 on cable), unless NBC telecasts a given contest. The NBA-NBC contract permits the league to exhibit 85 games per year on superstations. Seventy were licensed to the Turner stations (TBS and TNT), leaving 15 potentially available for WGN to license from the league. It disdained the opportunity. The Bulls sold 30 games directly to WGN, treating these as over-the-air broadcasts authorized by the NBC contract — not to mention the district court’s injunction. The Bulls’ only concession (perhaps more to *596the market’ than to the league) is that WGN does not broadcast a Bulls game at the same time as a basketball telecast on a Turner superstation.
Back in 1991 and 1992, the parties were debating whether the NBA’s television arrangements satisfied § 1 of the Sports Broadcasting Act, 15 U.S.C. § 1291. We held not, because the Act addresses the effects of “transfers” by a “league of clubs,” and the NBA had prescribed rather than “transferred” broadcast rights. The 1993 contract was written with that distinction in mind. The league asserted title to the copyright interests arising from the games and transferred all broadcast rights to NBC; it received some back, subject to contractual restrictions. Section 1 has been satisfied. But the league did not pay enough attention to § 2, 15 U.S.C. § 1292, which reads:
Section 1291 of this title shall not apply to any joint agreement described in the first sentence in such section which prohibits any person to whom such rights are sold or transferred from televising any games within any area, except within the home territory of a member club of the league on a day when such club is playing at home.
The NBA-NBC contract permits each club to license the broadcast of its games, and then, through the restriction on superstation broadcasts, attempts to limit telecasts to the teams’ home markets. Section 2 provides that this makes § 1 inapplicable, so the Sports Broadcasting Act leaves the antitrust laws in force.
Our prior opinion observed that the Sports Broadcasting Act, as a special-interest exception to the antitrust laws, receives a beady-eyed reading. A league has to jump through every hoop; partial compliance doesn’t do the trick. The NBA could have availed itself of the Sports Broadcasting Act by taking over licensing and by selling broadcast rights in the Bulls’ games to one of the many local stations in Chicago, rather than to WGN. The statute offered other options as well. Apparently the league did not want to use them, in part for tax reasons and in part because it sought to avoid responsibilities that come from being a licensor, rather than a regulator, of telecasts. Such business decisions are understandable and proper, but they have consequences under the Sports Broadcasting Act. By signing a contract with NBC that left the Bulls, rather than the league, with the authority to select the TV station that would broadcast the games, the NBA made its position under the Sports Broadcasting Act untenable. For as soon as the Bulls picked WGN, any effort to control cable system retransmission of the WGN signal tripped over § 2. The antitrust laws therefore apply, and we must decide what they have to say about the league’s effort to curtail superstation transmissions.
Three issues were left unresolved in 1992. One was whether the Bulls and WGN, as producers, suffer antitrust injury. 961 F.2d at 669-70. The NBA has not pursued this possibility, and as it is not jurisdictional (plaintiffs suffer injury in fact), we let the question pass. The other two issues are related. We concluded in 1992 that the district court properly condemned the NBA’s superstation rule under the quick-look version of the Rule of Reason, see National Collegiate Athletic Association v. Board of Regents of the University of Oklahoma, 468 U.S. 85, 104 S.Ct. 2948, 82 L.Ed.2d 70 (1984), because (a) the league did not argue that it should be treated as a single entity, and (b) the anti-free-riding justification for the su-perstation rule failed because a fee collected on nationally telecast games would compensate other teams (and the league as a whole) for the value of their contributions to the athletic contests being broadcast. 961 F.2d at 672-76. Back in the district court, the NBA argued that it is entitled to be treated as a single firm and therefore should possess the same options as other licensors of entertainment products; outside of court, the league’s Board of Governors adopted a rule requiring any club that licenses broadcast rights to superstations to pay a fee based on the amount the two Turner stations pay for games they license directly from the league.
Plaintiffs say that the single-entity argument was forfeited by its omission from the first appeal, but we think not. As our 1992 opinion observed, the case went to initial trial and decision within seven weeks, 961 F.2d at 676, a salutary development made possible in *597part by judicial willingness to entertain in subsequent rounds of the ease arguments that could not be fully developed in such short compass. If defendants in complex cases feared that any arguments omitted from the first phase of the case would be lost forever, they would drag their heels in order to ensure that nothing was overlooked, a step that would benefit no one. Cf. Schering Corp. v. Illinois Antibiotics Co., 89 F.3d 357 (7th Cir.1996). That is why we noted that the argument would be available in the ensuing stages of the case, 961 F.2d at 672-73, and why the district court properly entertained and resolved it on the merits.
The district court was unimpressed by the NBA’s latest arguments. It held that a sports league should not be treated as a single firm unless the teams have a “complete unity of interest” — which they don’t. The court also held the fee to be invalid. Our opinion compelled the judge to concede that a fee is proper in principle. 961 F.2d at 675-76. But the judge thought the NBA’s fee excessive. Instead of starting with the price per game it had negotiated with Turner (some $450,000), and reducing to account for WGN’s smaller number of cable outlets, as it did, the judge concluded that the league should have started with the advertising revenues WGN generated from retransmission on cable (the “outer market revenues”). Then it should have cut this figure in half, the judge held, so that the Bulls could retain “their share” of these revenues. The upshot: the judge cut the per game fee from roughly $138,000 to $39,400.
The district court’s opinion concerning the fee reads like the ruling of an agency exercising a power to regulate rates. Yet the antitrust laws do not deputize district judges as one-man regulatory agencies. The core question in antitrust is output. Unless a contract reduces output in some market, to the detriment of consumers, there is no antitrust problem. A high price is not itself a violation of the Sherman Act. See Broadcast Music, Inc. v. CBS, Inc., 441 U.S. 1, 9-10, 19-20, 22 n. 40, 99 S.Ct. 1551, 1557-58, 1562-63, 1564 n. 40, 60 L.Ed.2d 1 (1979); Buffalo Broadcasting Co. v. ASCAP, 744 F.2d 917 (2d Cir.1984). WGN and the Bulls argue that the league’s fee is excessive, unfair, and the like. But they do not say that it will reduce output. They plan to go on broadcasting 30 games, more if the court will let them, even if they must pay $138,000 per telecast. Although the fee exceeds WGN’s outer-market revenues, the station evidently obtains other benefits — for example, (i) the presence of Bulls games may increase the number of cable systems that carry the station, augmenting its revenues ’round the clock; (ii) WGN slots into Bulls games ads for its other programming; and (iii) many viewers will keep WGN on after the game and watch whatever comes next. Lack of an effect on output means that the fee does not have antitrust significance. Once antitrust issues are put aside, how much the NBA charges for national telecasts is for the league to resolve under its internal governance procedures. It is no different in principle from the question how much (if any) of the live gate goes to the visiting team, who profits from the sale of cotton candy at the stadium, and how the clubs divide revenues from merchandise bearing their logos and trademarks. Courts must respect a league’s disposition of these issues, just as they respect contracts and decisions by a corporation’s board of directors. Charles O. Finley & Co. v. Kuhn, 569 F.2d 527 (7th Cir.1978); cf. Baltimore Orioles, Inc. v. Major League Baseball Players Association, 805 F.2d 663 (7th Cir.1986).
According to the league, the analogy to a corporate board is apt in more ways than this. The NBA concedes that it comprises 30 juridical entities — 29 teams plus the national organization, each a separate corporation or partnership. The teams are not the league’s subsidiaries; they have separate ownership. Nonetheless, the NBA submits, it functions as a single entity, creating a single product (“NBA Basketball”) that competes with other basketball leagues (both college and professional), other sports (“Major League Baseball”, “college football”), and other entertainments such as plays, movies, opera, TV shows, Disneyland, and Las Vegas. Separate ownership of the clubs promotes local boosterism, which increases interest; each ownership group also has a powerful *598incentive to field a better team, which makes the contests more exciting and thus more attractive. These functions of independent team ownership do not imply that the league is a cartel, however, any more than separate ownership of hamburger joints (again useful as an incentive device, see Benjamin Klein & Lester F. Saft, The Law and Economics of Franchise Tying Contracts, 28 J.L. & Econ. 345 (1985)) implies that McDonald’s is a cartel. Whether the best analogy is to a system of franchises (no one expects a McDonald’s outlet to compete with other members of the system by offering pizza) or to a corporate holding company structure (on which see Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 104 S.Ct. 2731, 81 L.Ed.2d 628 (1984)) does not matter from this perspective. The point is that antitrust law permits, indeed encourages, cooperation inside a business organization the better to facilitate competition between that organization and other producers. To say that participants in an organization may cooperate is to say that they may control what they make and how they sell it: the producers of Star Trek may decide to release two episodes a week and grant exclusive licenses to show them, even though this reduces the number of times episodes appear on TV in a given market, just as the NBA’s superstation rule doés.
The district court conceded this possibility but concluded that all cooperation among separately incorporated firms is forbidden by § 1 of the Sherman Act, except to the extent Copperweld permits. Copperweld, according to the district court, “is quite narrow, and rests solely upon the fact that a parent corporation and its wholly-owned subsidiary have a ‘complete unity of interest’ ” (quoting from 467 U.S. at 771, 104 S.Ct. at 2741). Although that phrase appears in Copperweld, the Court offered it as a statement of fact about the parent-subsidiary relation, not as a proposition of law about the limits of permissible cooperation. As a proposition of law, it would be silly. Even a single firm contains many competing interests. One division may make inputs for another’s finished goods. The first division might want to sell its products directly to the market, to maximize income (and thus the salary and bonus of the division’s managers); the second division might want to get its inputs' from the first at a low transfer price, which would maximize the second division’s paper profits. Conflicts are endemic in any multi-stage firm, such as General Motors or IBM, see Robert G. Ec-cles, Transfer Pricing as a Problem of Agency, in Principals and Agents: The Structure of Business 151 (Pratt & Zeckhauser eds. 1985), but they do not imply that these large firms must justify all of their acts under the Rule of Reason. Or consider a partnership for the practice of law (or accounting): some lawyers would be better off with a lockstep compensation agreement under which all partners with the same seniority have the same income, but others would prosper under an “eat what you kill” system that rewards bringing new business to the firm. Partnerships have dissolved as a result of these conflicts. Yet these wrangles — every bit as violent as the dispute among the NBA’s teams about how to generate and divide broadcast revenues — do not demonstrate that law firms are cartels, or subject to scrutiny under the Rule of Reason their decisions about where to open offices or which clients to serve.
Copperweld does not hold that only conflict-free enterprises may be treated as single entities. Instead it asks why the antitrust laws distinguish between unilateral and concerted action, and then assigns a parent-subsidiary group to the “unilateral” side in light of those functions. Like a single firm, the parent-subsidiary combination cooperates internally to increase efficiency. Conduct that “deprives the marketplace of the independent centers of decisionmaking that competition assumes”, 467 U.S. at 769, 104 S.Ct. at 2740, without the efficiencies that come with integration inside a firm, go on the “concerted” side of the line. And there are entities in the middle: “mergers, joint ventures, and various vertical agreements” {id. at 768, 104 S.Ct. at 2740) that reduce the number of independent decisionmakers yet may improve efficiency. These are assessed under the Rule of Reason. We see no reason why a sports league cannot be treated as a single firm in this typology. It produces a single product; cooperation is essential (a *599league with one team would be like one hand clapping); and a league need not deprive the market of independent centers of decision-making. The district court’s legal standard was therefore incorrect, and a judgment resting on the application of that standard is flawed.
Whether the NBA itself is more like a single firm, which would be analyzed only under § 2 of the Sherman Act, or like a joint venture, which would be subject to the Rule of Reason under § 1, is a tough question under Copperweld. It has characteristics of both. Unlike the colleges and universities that belong to the National Collegiate Athletic Association, which the Supreme Court treated as a joint venture in NCAA, the NBA has no existence independent of sports. It makes professional basketball; only it can malte “NBA Basketball” games; and unlike the NCAA the NBA also “makes” teams. After this ease was last here the NBA created new teams in Toronto and Vancouver, stocked with players from the 27 existing teams plus an extra helping of draft choices. All of this makes the league look like a single firm. Yet the 29 clubs, unlike GM’s plants, have the right to secede (wouldn’t a plant manager relish that!), and rearrange into two or three leagues. Professional sports leagues have been assembled from clubs that formerly belonged to other leagues; the National Football League and the NBA fit that description, and the teams have not surrendered their power to rearrange things yet again. Moreover, the league looks more or less like a firm depending on which facet of the business one examines. See Phillip E. Areeda, 7 Antitrust Law para. 1478d (1986). From the perspective of fans and advertisers (who use sports telecasts to reach fans), “NBA Basketball” is one product from a single source even though the Chicago Bulls and Seattle Supersonies are highly distinguishable, just as General Motors is a single firm even though a Corvette differs from a Chevrolet. But from the perspective of college basketball players who seek to sell their skills, the teams are distinct, and because the human capital of players is not readily transferable to other sports (as even Michael Jordan learned) the league looks more like a group of firms acting as a monopsony. That is why the Supreme Court found it hard to characterize the National Football League in Brown v. Pro Football, Inc., — U.S. -, -, 116 S.Ct. 2116, 2126, 135 L.Ed.2d 521 (1996): “the clubs that make up a professional sports league are not completely independent economic competitors, as they depend upon a degree of cooperation for economic survival- In the present context, however, that circumstance makes the league more like a single bargaining employer, which analogy seems irrelevant to the legal issue before us.” To say that the league is “more like a single bargaining employer” than a multi-employer unit is not to say that it necessarily is one, for every purpose.
The league wants us to come to a conclusion on this subject (six years of litigation is plenty!) and award it the victory. Yet as we remarked in 1992, “Characterization is a creative rather than exact endeavor.” 961 F.2d at 672. The district court plays the leading role, followed by deferential appellate review. We are not authorized to announce and apply our own favored characterization unless the law admits of only one choice. The Supreme Court’s ambivalence in Brown, like the disagreement among judges on similar issues, implies that more than one characterization is possible, and therefore that the district court must revisit the subject using the correct legal approach.
Most courts that have asked whether professional sports leagues should be treated like single firms or like joint ventures have preferred the joint venture characterization. E.g., Sullivan v. NFL, 34 F.3d 1091 (1st Cir.1994); North American Soccer League v. NFL, 670 F.2d 1249 (2d Cir.1982); Smith v. Pro Football, Inc., 593 F.2d 1173, 1179 (D.C.Cir.1978). But Justice Rehnquist filed a strong dissent from the denial of certiorari in the soccer case, arguing that “the league competes as a unit against other forms of entertainment”, NFL v. North American Soccer League, 459 U.S. 1074, 1077, 103 S.Ct. 499, 500, 74 L.Ed.2d 639 (1982), and the fourth circuit concluded that the Professional Golf Association should be treated as one firm for antitrust purposes, even though that sport is less economically integrated than the NBA. Seabury Management, Inc. v. PGA of *600America, Inc., 878 F.Supp. 771 (D.Md.1994), affirmed in relevant part, 52 F.3d 322 (4th Cir.1995). Another court of appeals has treated an electric cooperative as a single firm, Mt. Pleasant v. Associated Electric Cooperative, 838 F.2d 268 (8th Cir.1988), though the co-op is less integrated than a sports league. These cases do not yield a clear principle about the proper characterization of sports leagues — and we do not think that Copperweld imposes one “right” characterization. Sports are sufficiently diverse that it is essential to investigate their organization and ask Copperweld’s functional question one league at a time — and perhaps one facet of a league at a time, for we do not rule out the possibility that an organization such as the NBA is best understood as one firm when selling broadcast rights to a network in competition with a thousand other producers of entertainment, but is best understood as a joint venture when curtailing competition for players who have few other market opportunities. Just as the ability of McDonald’s franchises to coordinate the release of a new hamburger does not imply their ability to agree on wages for counter workers, so the ability of sports teams to agree on a TV contract need not imply an ability to set wages for players. See Jesse W. Markham & Paul V. Teplitz, Baseball Economics and Public Policy (1981); Arthur A. Fleisher III, Brian L. Goff & Robert D. Tollison, The National Collegiate Athletic Association: A Study in Cartel Behavior (1992).
However this inquiry may come out ■ on remand, we are satisfied that the NBA is sufficiently integrated that its superstation rules may not be condemned without analysis under the full Rule of Reason. We affirmed the district court’s original injunction after applying the “quick look” version because the district court had characterized the NBA as something close to a cartel, and the league had not then made a Copperweld argument. After considering this argument, we conclude that when acting in the broadcast market the NBA is closer to a single firm than to a group of independent firms. This means that plaintiffs cannot prevail without establishing that the NBA possesses power in a relevant market, and that its exercise of this power has injured consumers. Even in the NCAA case, the first to use a bobtailed Rule of Reason, see Diane P. Wood, Antitrust 1984: Five Decisions in Search of a Theory, 1984 Sup.Ct.Rev. 69, 110-12, the Court satisfied itself that the NCAA possesses market power. The district court had held that there is a market in college football telecasts on Saturday afternoon in the fall, a time when other entertainments do not flourish but college football dominates. Only after holding that this was not clearly erroneous did the Court cast any burden of justification on the NCAA. 468 U.S. at 111-13, 104 S.Ct. at 2965-67; see also International Boxing Club v. United States, 358 U.S. 242, 79 S.Ct. 245, 3 L.Ed.2d 270 (1959).
Substantial market power is an indispensable ingredient of every claim under the full Rule of Reason. Digital Equipment Corp. v. Uniq Digital Technologies, Inc., 73 F.3d 756, 761 (7th Cir.1996); Sanjuan v. American Board of Psychiatry & Neurology, Inc., 40 F.3d 247, 251 (7th Cir.1994); Hardy v. City Optical, Inc., 39 F.3d 765, 767 (7th Cir.1994); Chicago Professional Sports Limited Partnership v. National Basketball Association 961 F.2d 667, 673 (7th Cir.1992); Will v. Comprehensive Accounting Corp., 776 F.2d 665, 670-74 (7th Cir.1985); Carl Sandburg Village Condominium Ass’n No. 1 v. First Condominium Development Co., 758 F.2d 203, 210 (7th Cir.1985). During the lengthy trial of this case, the NBA argued that it lacks market power, whether the buyers are understood as the viewers of games (the way the district court characterized things in NCAA) or as advertisers, who use games to attract viewers (the way the Supreme Court characterized a related market in Times-Picayune Publishing Co. v. United States, 345 U.S. 594, 73 S.Ct. 872, 97 L.Ed. 1277 (1953)). College football may predominate on Saturday afternoons in the fall, but there is no time slot when NBA basketball predominates. The NBA’s season lasts from November through June; games are played seven days a week. This season overlaps all of the other professional and college sports, so even sports fanatics have many other options. From advertisers’ perspective — likely the right one, because advertisers are the ones who actually pay for telecasts — the mar*601ket is even more competitive. Advertisers seek viewers of certain demographic characteristics, and homogeneity is highly valued. A homogeneous audience facilitates targeted ads: breakfast cereals and toys for cartoon shows, household appliances and detergents for daytime soap operas, automobiles and beer for sports. If the NBA assembled for advertisers an audience that was uniquely homogeneous, or had especially high willingness-to-buy, then it might have market power even if it represented a small portion of air-time. The parties directed considerable attention to this question at trial, but the district judge declined to make any findings of fact on the subject, deeming market power irrelevant. As we see things, market power is irrelevant only if the NBA is treated as a single firm under Copperweld; and given the difficulty of that issue, it may be superior to approach this as a straight Rule of Reason ease, which means starting with an inquiry into market power and, if there is power, proceeding to an evaluation of competitive effects.
Perhaps this can be accomplished using the materials in the current record. Although the judge who presided at the trial died earlier this year, the parties may be willing to agree that an assessment of credibility is unnecessary, so that a new judge could resolve the dispute after reviewing the transcript, exhibits, and stipulations, and entertaining argument. See Fed.R.Civ.P. 63. At all events, the judgment of the district court is vacated, and the ease is remanded for proceedings consistent with this opinion. Pending further proceedings in the district court or agreement among the parties, the Bulls and WGN must respect the league’s (and the NBC contract’s) limitations on the maximum number of superstation telecasts.