TNT Amusements, Inc. v. Torch Electronics, LLC

• Court: District Court, E.D. Missouri
• Decided: February 20, 2025
• Docket: 4:23-cv-00330
• Status: Unknown

Opinion

UNITED STATES DISTRICT COURT EASTERN DISTRICT OF MISSOURI EASTERN DIVISION

TNT AMUSEMENTS, INC., ) d/b/a PLAY-MOR COIN-OP, ) ) Plaintiff, ) ) v. ) Case No. 4:23CV330 JAR ) TORCH ELECTRONICS, LLC, et al., ) ) Defendants. )

MEMORANDUM AND ORDER This matter is before the Court on the parties’ respective motions to exclude expert testimony. Both motions will be denied. Background Plaintiff TNT Amusements leases traditional arcade games in retail locations throughout Missouri. Defendant Torch Electronics leases “no-chance” gaming devices in the same market. Plaintiff alleges that the parties are competitors in the amusement device industry. A central issue in this case is whether Torch’s devices are illegal slot machines being operated in violation of Missouri gambling laws. That question turns on whether the devices involve an element of chance, as contemplated under Mo. Rev. Stat. § 572.010. To explain the functionality of the devices in question, TNT retained expert witness Stacy Friedman and Torch retained expert witness Nick Farley. Friedman is a professional game designer and casino gaming mathematician with a bachelor’s degree in computer science from Harvard University. He is President of Olympian Gaming and has 25 years of experience in software design and slot machine mathematics. He has served as a subject matter expert in over 40 cases involving gaming-related intellectual President of Nick Farley & Associates, d/b/a Eclipse Compliance Testing, a regulatory compliance test laboratory and electronic gaming device consulting firm serving the regulated gaming, amusement, and vending industries. Farley has been testing electronic gaming equipment for regulated clients for nearly 40 years and has frequently testified as a computer gaming expert in state and federal courts. Both experts submitted their initial reports on November 6, 2023. As relevant to the instant motions, Farley’s report was largely informed by his staff’s review of five Torch devices between May 2018 and January 2021. He opined that Torch devices are designed to eliminate randomized outcomes so that chance does not play a role. The devices do not use random

number generators to determine the outcome of a turn but instead use predetermined finite pools that deliver outcomes in sequential order. A useful analogy would be a Rolodex of outcomes from 1 to 1000, eventually cycling back to 1. Additionally, each Torch device has an optional prize viewer feature that allows a player to preview the outcome of the next turn before paying to play it. (Doc. 241-1). Upon review of Farley’s report, Friedman issued several supplemental reports based on his examination of additional program drives. After reviewing a drive provided by TNT, Friedman issued a supplemental report dated December 15, 2023, in which Friedman noted that the starting point for each pool of outcomes and the combination of symbols displayed on each turn are random. (Doc. 87-7). After analyzing four more drives provided by Torch purporting to

be identical to those reviewed by Farley, Friedman issued a second supplement dated February 26, 2024, in which Friedman again identified random selection of starting points in each pool and random number generation determining display symbols as well as a shuffle function randomizing payout amounts. (Doc. 161-1). Upon review of Friedman’s reports, Farley then provided four newly cloned drives represented to be exact duplicates of those reviewed by Farley for his initial report. In his third and final report dated March 18, 2024, Friedman noted that the then-correct drives still implemented random initial selection into the pools. For example, “in the listing of database tables there are 12 pools related to Lucky Striker, three sets of four each for four different payback settings for three different game configurations.” (Doc. 241-2 at 9, 28-29). Each game listed had 16 possible “start index” options, some used and others not. (Id. at 10). Friedman concluded that the processing flow in all four drives “performs equivalent functionality to randomly select the initial index into the game pool.” (Id. at 19). In addition, Friedman again

observed that the software used random “reverse mapping” to arrive at a display of visual symbols based on the prize amount called from the database. (Id. at 19-26). In short, Friedman opined that, though the random shuffling noted in his second supplement was no longer present, there were still random elements to the devices, namely the starting points into the outcome pools as well as the displays of symbols generated for each outcome. (Id. at 29). After review of Friedman’s third supplement, Farley issued a declaration dated April 26, 2025, acknowledging the discrepancies in the drives provided to Friedman and attributing the error to his employee and to Banilla. (Doc. 188-5). Farley also issued a rebuttal report stating: …the two instances of alleged randomness Mr. Friedman noted in his Third Supplemental report are immaterial to game classification and do not alter my prior opinions. The process used to select the “current_index” start point position after device activation or startup can happen any number of ways before the game is actually played by a customer in the normal stream of commerce. This is a very limited system setup issue that should only happen one time after device activation. Once “current_index” is set, the starting point for the given pool is set and the outcome of each turn is determined based on the fixed and sequential list of finite and predetermined outcomes that continually repeats once the pool of outcomes is depleted. Further, whatever randomness Mr. Friedman desires to highlight specific to how the entertaining symbols are displayed to the player is irrelevant. As Mr. Friedman admits, the symbols are “reverse-mapped” based on the predetermined outcome the player knows or can know by using the preview feature; the entertaining symbols do not dictate or influence the outcome. In other words, the actual game outcome is derived from the sequential delivery of game outcomes from a finite pool where no randomness is used to determine the game outcome.

(Doc. 194-1 at 7-8).

The Court held a hearing on October 1, 2024, and heard testimony from both Friedman and Farley. (Doc. 300). Both experts now agree on the technical functionality of Torch devices. Specifically, they agree that there is a random starting point into each finite pool of outcomes, in each game and on each play level within each game. From the random starting point in a given game and play level, outcomes are produced in a predetermined sequential order, not shuffled. They also agree that the symbols displayed for each turn (solely for the player’s entertainment) are randomly “reverse-mapped” to correspond to the outcome dictated by the pool. The experts’ only disagreement now is whether these characteristics are relevant to the Court’s analysis of “chance” under Missouri law. Torch’s motion to exclude Friedman’s opinions centrally attacks Friedman’s qualifications to opine on game classification under Missouri law and also his initial findings of November 6, 2023, which relied on Farley’s factual description of the device software and not Friedman’s own examination. (Doc. 228). TNT’s motion to exclude Farley’s opinions likewise challenges Farley’s qualifications to opine on the legality of Torch devices under Missouri law, particularly regarding whether specific random functions are relevant to the analysis of “chance.” (Doc. 241). Additionally, TNT argues that Farley’s methodology was unreliable insofar as he depended on subordinate staff and Banilla representatives to supply the information for his reports rather than examining the software himself until confronted with Friedman’s findings. Torch responds that these circumstances bear only on Farley’s credibility and the weight to be given to his opinions, not their admissibility. Federal Rule of Evidence 702 governs the admission of expert testimony. It states: A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or otherwise if the proponent demonstrates to the court that it is more likely than not that: (a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert’s opinion reflects a reliable application of the principles and methods to the facts of the case. Fed. R. Evid. 702. The rule was amended in 2000 in response to Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579 (1993), which charges trial judges with a “gatekeeping” role to screen expert testimony for relevance and reliability. Id. at 590–93; see also Russell v. Whirlpool Corp., 702 F.3d 450, 456 (8th Cir. 2012) (“The main purpose of Daubert exclusion is to prevent juries from being swayed by dubious scientific testimony.”) “To satisfy the relevance requirement, the proponent must show that the expert’s reasoning or methodology was applied properly to the facts at issue.” Barrett v. Rhodia, Inc., 606 F.3d 975, 980 (8th Cir. 2010). “To satisfy the reliability requirement, the party offering the expert testimony must show by a preponderance of the evidence both that the expert is qualified to render the opinion and that the methodology underlying his conclusions is scientifically valid.” Id. (citation omitted). The inquiry envisioned by Rule 702 is a flexible one, designed to exclude “vague theorizing based on general principles” or “unsupported speculation,” but not requiring an opinion to be “a scientific absolute” in order to be admissible. Adams v. Toyota Motor Corp., 867 F.3d 903, 914–16 (8th Cir. 2017); see also Kumho Tire Co. v. Carmichael, 526 U.S. 137, broad latitude when it decides how to determine reliability as it enjoys in respect to its ultimate reliability opinion.”). In the most recent amendment to Rule 702, effective December 1, 2023, the Advisory Committee noted that “many courts have held that the critical questions of the sufficiency of an expert’s basis, and the application of the expert’s methodology, are questions of weight and not admissibility. These rulings are an incorrect application of Rules 702 and 104(a).” Fed. R. Evid. 702 advisory committee’s note to 2023 amendment. Rather, proponents of expert testimony must establish admissibility of the proffered evidence by a preponderance of the evidence. The amendment does not impose any specific new procedures but instead “is simply intended to

clarify that Rule 104(a)’s requirement applies to expert opinions under Rule 702.” Id. Discussion As discussed during the Daubert hearing on October 1, 2024, and further confirmed in the parties’ post-hearing briefs, the parties now agree that neither of their experts may opine on the ultimate question of whether Torch devices are illegal under Missouri law. And both experts now agree on the technical functionality of Torch devices. Thus, there is little left for the Court to resolve. Friedman Torch first asserts that Friedman is unqualified to opine on game classification under any state’s laws, and particularly Missouri’s. This argument is now moot given that neither expert

will be permitted to opine on that ultimate issue. That matter aside, the Court readily finds Friedman to be qualified as a technical expert in this case. As previously noted, Friedman is a Harvard-educated computer scientist with decades of experience in game design and casino gaming mathematics. Having heard Friedman’s credible testimony during the Daubert hearing, issues. The remainder of Torch’s motion attacks Friedman’s initial report of November 6, 2023, in which, at that early stage, Friedman relied on information provided by Farley’s team and not his own examination. This argument is also moot in light of Friedman’s subsequent examination of multiple drives and resulting supplemental reports, which Farley has since accepted as accurate. Torch advances no other viable arguments for exclusion of Friedman’s final technical opinions. Torch’s motion will therefore be denied Farley Similarly, TNT does not challenge Farley’s expert qualifications except to opine on legal

issues, and the Court finds him otherwise qualified based on his education and experience in the field of game design and testing. However, TNT raises a colorable concern with respect to the reliability of Farley’s opinions given his failure to verify the accuracy of information provided by subordinates and Banilla representatives to generate his reports. TNT characterizes this issue in several ways applying Rule 702 standards. To summarize: (1) Farley lacks personal knowledge of critical aspects of the devices; (2) Farley’s “methodology” (i.e., practice) of trusting staff and Banilla proved to be defective and unreliable; (3) Farley’s initial “methodology” to determine what game functions are relevant was to simply believe Banilla, without independent investigation; and (4) once confronted with evidence of random functions, Farley’s “methodology” to conclude that those functions were irrelevant is unscientific and subjective.

While TNT’s critiques are valid given the evolution of events in this case, the Court is not persuaded that Farley’s operational practices vis-à-vis staff and clients – though clearly fallible – are fairly characterized as his scientific “methodology.” In the Court’s basic understanding of both experts’ reports, the standard accepted methodology for game design once the discrepancies in Farley’s initial report came to light, he personally conducted an independent review of the database and stored procedures and ultimately arrived at the same technical conclusions as Friedman made. Any previous oversight resulting from Farley’s delegation of responsibilities goes to his credibility. Thus, despite early errors, the Court finds that the preponderance of more recent evidence supports the admissibility of Farley’s final technical findings with respect to device functionality. TNT’s motion will be denied in this respect, but TNT may test Farley’s personal knowledge on cross-examination. Questions of reliability and objectivity go to weight rather than admissibility. DiCarlo v. Keller Ladders, Inc., 211 F.3d 465, 468 (8th Cir. 2000) (“An expert witness’s bias goes to the weight, not the

admissibility of the testimony, and should be brought out on cross-examination.”). “Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible evidence.” Daubert, 509 U.S. at 596. Finally, to the extent TNT challenges the scientific basis for Farley’s professional opinions as to whether the random functions identified by both experts are relevant to the ultimate question of chance, as previously discussed, such opinions are improper and will not be permitted by either expert. Given the parties’ agreement on this point, TNT’s motion is moot as to these arguments and will be denied on this basis as well. Conclusion The Court orders that the parties’ respective experts shall not offer legal conclusions or opinions as to the legality of Torch devices under Missouri law. However, these witnesses may offer expert testimony explaining the technical mechanics of the devices, including any functions involving random selection. Accordingly, IT IS HEREBY ORDERED that Plaintiff's motion to exclude the testimony of Nick Farley is DENIED. (Doc. 240). IT IS FURTHER ORDERED that Defendants’ motion to exclude the testimony of Stacy Friedman is DENIED. (Doc. 227). Dated this 20th day of February, 2025. A few

UNITED STATES DISTRICT JUDGE