Another API (c) case with false advertising and contract claims too

Trackman, Inc. v. GSP Golf AB, 2024 WL 4276497, No. 23 Civ. 598 (NRB) (S.D.N.Y. Sept. 24, 2024)

Trackman makes the golf simulator game Perfect Golf, which offers users the ability to virtually play some of the most famous golf courses in the world. Defendants allegedly copied key components of Trackman’s copyrighted software and falsely suggested, in promotions and advertisements, that defendants were authorized to use the well-known courses in their game.

Although the court dismissed a contract claim, copyright and false advertising claims survived.

Plaintiff’s Perfect Golf simulator allows users to design golf courses; has “an API4 for external tournament sites to be able to fully integrate into Perfect Golf for online real-time scoring and tracking”; and allows users to play each other on courses designed in the simulator. Using a combination of radars and cameras, plaintiff’s launch monitors track the full trajectory of a golf shot. Launch monitors incorporated into plaintiff’s simulator technology, which allows users to play golf indoors using real clubs and balls in front of an “impact screen” that displays the simulation and keeps golf balls from ricocheting back at the player after they are hit.

Perfect Golf has a EULA that bans reverse engineering.

Although Perfect Golf used to be compatible with third-party launch monitors, as of August 2020, Perfect Golf users had to buy plaintiff’s launch monitors to play the game.

Defendant saw the compatibility-breaking as an opportunity to replace Perfect Golf and be compatible across a number of launch monitors.

This first required developing golf simulator software, eventually called GSPro. GSPro allegedly copied Perfect Golf’s course-creating code as well as copied Perfect Golf’s ‘combine’ feature,” which “enables golfers to identify strengths and weaknesses in their game.” Defendants also allegedly developed an online platform to host tournament play by GSPro users that copied Perfect Golf’s API data structures for simulating golf competitions. And they allegedly copied golf courses created on Perfect Golf’s course design platform.

In addition, defendants allegedly claimed that course selection included “iconic, branded courses like St. Andrews in Scotland and various PGA Tour Tournament Players Club courses throughout the United States” without having the licenses “required” to offer those courses, while plaintiff had “diligently sought and obtained permission[ ], including trademark licenses, from the owners of branded golf courses,” including St. Andrews and various PGA Tour courses. “Eventually, in 2023, the trademark owners of the St. Andrews and PGA Tour courses sent cease-and-desist letters to defendants, after which defendants ‘removed, disabled access to, or renamed the St. Andrews and PGA Tour courses,’” but plaintiffs argued that the damage had been done.

The court refused to hold on a motion to dismiss that the API data structures at the center of the dispute (which sound like they’re needed for interoperability) were not copyrightable, relying on the Federal Circuit’s decision in Google v. Oracle. The structures at issue include “shared naming conventions that allows a simulated golf tournament site … both to communicate with [client] software … and to process data, like how many shots it takes for a player to complete a hole in the golf simulation.”

The court found that plaintiff sufficiently pled the “modest” requirements of originality by alleging that it spent “years” developing the program, which provides, among other things, “an immersive experience centered on high-resolution visuals” and “hyper-realistic gameplay.” (Why does this make the API protectable?) It also alleged that it “built” an API. (That very verb signals the issue.) But: “Such allegations, at this stage, are more than sufficient to demonstrate that Perfect Parallel both independently developed the subject API structures and made numerous creative decisions in doing so.” Anyway, questions of originality are generally inappropriate for determination on a motion to dismiss. Likewise, whether plaintiff’s API structures were a protectable process or method of operation couldn’t be determined on a motion to dismiss.

Estoppel/license defenses were also premature, and the complaint satisfied the discovery rule on its face for statute of limitations purposes.

However, the court dismissed the breach of contract claim, finding the EULA’s anti-reverse engineering provisions preempted by copyright law. “Put simply, plaintiff claims that defendants breached the [reverse engineering] Provision by ‘studying and analyzing’ plaintiff’s software as part of its efforts to develop its own competing golf simulator software that would be compatible with Course Forge courses and third-party hardware.”

The disputed work was clearly within the scope of copyright—software/literary work. For express preemption to apply, “the state law claim must involve acts of reproduction, adaptation, performance, distribution, or display.” That was true here. But a claim isn’t preempted if it has an extra element that makes it qualitatively different. Unlike other circuits, the Second Circuit has instructed that the “extra element” inquiry is not “mechanical” but instead “requires a holistic evaluation of the nature of the rights sought to be enforced, and a determination whether the state law action is qualitatively different from a copyright infringement claim.”

While some courts have held that the promise element of a contract claim suffices, categorically exempting contract claims from preemption, the Second Circuit hasn’t. (And in an age of unavoidable contracts of adhesion, saying that as a matter of law there’s an actual “promise” and then that the promise avoids preemption seems wrong.) In the Second Circuit, “a breach of contract claim is preempted if it is merely based on allegations that the defendant did something that the copyright laws reserve exclusively to the plaintiff (such as unauthorized reproduction, performance, distribution, or display).”

Plaintiff argued that its contract claim was distinguishable because it is specifically (and carefully) “directed to the non-copying acts of studying and analyzing copyrighted works.” But, evaluating the nature of the rights sought to be enforced “holistically” showed that the contract claim was centrally about copying (or studying) in order to create competing works. Defendants’ “studying and analyzing” were “part and parcel of their broader infringing conduct that is at the heart of plaintiff’s copyright claims (i.e., unlawfully developing, producing, and distributing plaintiff’s software).”

Then, in a footnote revealing a fundamental misunderstanding, the court noted the strategic reasons for a breach of a contract claim, if the API structures turn out not to be copyrightable—in that case plaintiff would be “wholly or partly without a remedy.” Thus, the court dismissed the contract claim without prejudice if there are “substantial changes in the law.” This is a flat-out mistake about 301’s scope, which doesn’t just apply when there are valid © claims. It applies when the subject matter is the same as ©’s subject matter, whether or not the material at issue is protectable. That’s why you can’t use state law to protect works whose copyright has expired, or the facts in a work. If it’s a claim whose gravamen is copying a fixed work, then the unprotectability of the copied material doesn’t matter.

And then the court upheld a false advertising claim that seems quite problematic to me. Plaintiff alleged that defendants “sought to commercially advertise and promote the availability of iconic branded golf courses for simulator play on the SGT platform,” which misleadingly suggested that they were “authorized to offer genuine, trademarked courses,” when, in reality, defendants “lacked the rights to offer these branded courses.” The court agreed that plaintiff didn’t allege literal falsity, but found that implied falsity was plausible.

The claims weren’t literally false because “iconic, branded golf courses” were available for play. But they might have falsely implied licensing/endorsement, and plaintiff didn’t need to provide extrinsic evidence of deception at this stage. Also (ugh), intentional deception might obviate the need for evidence of confusion, and the facts here might allow that theory: Defendants allegedly

(1)  knew that they did not have the requisite authorization to make available the trademarked courses; (2) made a litany of statements on social media and elsewhere suggesting that they had such authorization; and (3) made these statements with the intention of influencing “a significant number of users” to purchase SGT subscriptions and GSPro downloads on the basis that they could play “at some of the most coveted courses around the world.”

Even though (3) was true, that was enough for the court.

What about materiality? It was plausible that consumers would care about whether the courses endorsed defendants, because the courses are official partners of plaintiff, which was plausibly related to plaintiff’s success.

What about Dastar? Some courts have rejected Lanham Act claims premised upon false representations of licensing status.” E.g., Sybersound Records, Inc. v. UAV Corp., 517 F.3d 1137 (9th Cir. 2008), rejected the plaintiff’s argument that “the licensing status of each work is part of the nature, characteristics, or qualities of the karaoke products” because they weren’t characteristics of the goods themselves. But the court found this line of cases to be irrelevant to the false advertising claim here. “Dastar and the like are concerned about impermissibly blurring the lines between trademark and copyright law.” The claims here were “based solely on the SGT Defendants’ misleading statements regarding the licensing of trademarked courses, not the licensing of expressive copyrighted (or copyrightable) material…. [T]he misrepresentations at issue have nothing to do with claims of authorship of an expressive work or creation of an invention.” Plaintiff wasn’t suing over copying the simulated version of the course.

This is a little weird given that the representations of the courses were copyrightable representations—the license was a license to represent the courses, not to play on them or replicate them in the physical world. More to the point, though, the analysis does not fit well with today’s textualism. Dastar says that is about the meaning of “origin” in the Lanham Act, even if that interpretation was motivated in significant part by avoiding a TM/© conflict. Dastar says that “origin” does not include the origin of intangible content—including who “stands behind” that intangible content. It specifically addresses the argument that intangible origin could well matter to consumers for “communicative” goods. You can get there textually by saying that “nature, characteristics or qualities” is broader than “origin,” for sure. But I think these days you have to take that step.

It’s also worth noting that the TM claim here is based on the representation of golf courses in the game, which shouldn’t require permission any more than a book about the golf courses would—there are pretty significant Rogers issues as well, and the trademark claimants aren’t even actually here to make their claims.