Author: Frank, PANews
On February 25, 2026, the prediction market platform Kalshi issued a fine of $20,397.58 to a YouTube editor. This exact amount marked the first publicly disclosed insider trading fine in the prediction market industry.
The person fined was Artem Kaptur, the visual effects editor for MrBeast, the world’s largest influencer. He invested approximately $4,000 on Kalshi to trade contracts related to MrBeast’s YouTube events, earning $5,397.58. However, this modest profit triggered a major industry-level regulatory signal, prompting the CFTC (U.S. Commodity Futures Trading Commission) to issue a formal enforcement advisory on prediction markets.
However, Kalshi is a KYC platform that requires real-name verification, making it relatively easy to identify a video editor trading under their real identity. What truly warrants investigation is what happens when someone with the same information moves to Polymarket, which does not require identity verification. PANews analysis found that on Polymarket contracts for the second season of MrBeast’s reality show “Beast Games,” the winning probability for the eventual champion was pushed to 94% three weeks before the season ended—exhibiting textbook characteristics of insider trading.
PANews will use Kalshi’s penalty as a starting point, combined with on-chain data anomalies from Polymarket, to deeply analyze how insider trading—once an exclusive term of Wall Street—has become a gray game accessible even to editing assistants in the era where everything can be wagered on.
The first insider trading fine in a prediction market
According to Kalshi’s disciplinary notice, Kaptur traded event contracts related to the MrBeast channel during August to September 2025, using his position at Beast Industries.
Kalshi’s monitoring system detected extremely anomalous statistical patterns: Kaptur achieved an “almost perfect win rate” in low-odds markets. Simultaneously, because Kalshi’s trading data is fully public, multiple platform users identified this anomaly and voluntarily reported it. Triggered by both findings, Kalshi froze Kaptur’s account and launched an investigation. The final penalty: forfeiture of all illicit gains amounting to $5,397.58, plus a $15,000 punitive fine, totaling $20,397.58, and a two-year platform ban.
On the same day, another even more absurd case was disclosed: California Republican gubernatorial candidate Kyle Langford bet approximately $200 on Kalshi that he would win the election, then immediately posted a screenshot of the trade on X to boast about it. Kalshi froze his account that same day and ultimately imposed a five-year ban and a fine of $2,246.36.
The fines on these two notices are not substantial, but their signaling value far exceeds the monetary amounts. On the same day, the CFTC issued an official enforcement advisory, explicitly citing Section 6(c)(1) of the Commodity Exchange Act, stating that both cases may constitute federal violations. CFTC Chair Mike Selig stated on X: "Our exchanges are the CFTC’s first line of defense against insider trading in prediction markets. If you attempt to manipulate, defraud, or engage in insider trading, we will find you and take action." This marks the first time a U.S. federal regulatory agency has issued such a direct warning against insider trading in prediction markets.
Beast Industries issued a statement saying it takes a "zero-tolerance" approach to insider trading by employees and has initiated an independent internal investigation. However, the company also recommended that Kalshi should communicate its findings more openly in the future.
However, all these efforts are predicated on one assumption: that Kalshi is a centralized KYC platform where users’ identities, bank statements, and IP addresses are fully known. Catching an editor who trades under their real identity reveals little. The real question is: what if someone with the same information chose a platform that requires no identity verification and settles transactions using anonymous wallets and USDC?
94% on Polymarket: Beast Games champion leaked on-chain
At the same time window that Kaptur was penalized for earning over $5,000 on Kalshi, MrBeast was advancing a much larger-scale project. The second season of the reality show "Beast Games," in collaboration with Amazon Prime Video, premiered on January 7, 2026, with 200 contestants competing for a record-breaking $5.1 million prize. The season finale, aired on February 25, revealed the ultimate champion: Player 167, former U.S. Air Force pilot and former University of Pennsylvania wide receiver Tyler Lucas.
But on Polymarket, this outcome seemed to have been "priced in" three weeks ago.
PANews analyzed the odds movement on Polymarket’s contract “Who will win Beast Games Season 2?” and discovered an extremely anomalous capital flow. Even though the show still had many contestants remaining and the finals were far off, the Yes shares representing Player 167 experienced sustained buying pressure that cannot be explained by normal market logic.
From the timeline, the anomaly is clear and significant. From late January to early February 2026, before the show had even entered the latter half of the elimination phase, Player 167’s win probability began to sharply diverge from fundamentals. By February 4, with three weeks remaining until the final finale, Tyler Lucas’s odds of winning had surged to 84%. By February 18, just one week before the season finale was revealed, the contract’s implied probability was pinned above 94%.
In stark contrast, other top performers who stood out on the show were almost entirely "priced at zero." In a reality show with 200 participants relying on physical and intellectual challenges, without definitive insider information, no rational capital would assign a single contestant a win probability above 90% at the midpoint of the elimination rounds.
The Reddit community and Polymarket comment sections have already exploded. Multiple posts directly state in their titles that “the winner has basically been spoiled by Polymarket.” Community members have drawn comparisons to the first season’s winner, Jeff Allen, whose victory was also leaked in advance—but this time, the data patterns are even more blatant.
The odds themselves are merely superficial. PANews fully scraped and analyzed the on-chain transaction data for this market, uncovering evidence more direct than odd fluctuations.
The entire Season 2 Beast Games betting market recorded 111,000 trades involving 2,640 unique addresses. Among these 2,640 addresses, one data point stands out sharply: 795 addresses traded exclusively on the contract of Player 167 throughout the entire market lifecycle. Among 25 players, "coincidentally" selecting only the eventual champion represents a concentration far beyond what normal betting logic could explain.
PANews further retrieved and cross-referenced the full transaction histories of all suspicious addresses, analyzing dimensions including: percentage of Beast Games trading, overall platform win rate, and cross-address connectivity. Ultimately, 147 highly suspicious addresses were identified. Among them, 16 addresses exhibited textbook insider trading characteristics—they participated exclusively in Beast Games markets on Polymarket and had no trading records in any other markets.
Among these 16 addresses, the most suspicious one, named "0xA1F3Cf8Ba7410956a2955D5300A9be7Ff1dBc07E-1767992471439," participated in only three Beast Games sub-markets, all of which were profitable, resulting in a 100% win rate and a total profit of $3,237. Several similar addresses exhibit comparable behavior—while individual profits are modest, their operational patterns are highly consistent. This suggests that insider traders may be deliberately dispersing their bets to reduce scrutiny.
From a volume perspective, more attention should be paid to mixed traders who made substantial profits on Beast Games but did not trade exclusively in MrBeast. Trading behavior patterns further heighten suspicion. PANews identified multiple highly synchronized "address clusters" in terms of timing and behavior. On January 27, the day with the highest overall trading volume (reaching $44,547 in a single day), the top suspect address completed all 12 trades within 17 minutes, earning $11,830. Two anonymous addresses executed sell trades at exactly 09:41 on January 30, each earning $3,542—with identical amounts, timing, and mirrored behavior.
So, who has the ability to place bets with such certainty mid-season? The trail points to a very limited number of groups: Beast Industries’ large post-production team, the 200 contestants themselves and their close social circles, and staff involved in scheduling and promotion (investigations revealed multiple addresses focused on film-related contracts with unusually high win rates). Kaptur’s trades on Kalshi amounted to just $4,000, yielding over $5,000 in profit. But on Polymarket, the cumulative profits from just the top suspected insider addresses tracked by PANews have already exceeded $100,000—and this is likely only the tip of the iceberg.
It's a feature and an unfair game.
Kalshi can precisely seize $5,397.58 in illicit proceeds because, as a regulated centralized exchange, all traders' identity information, bank statements, and IP trajectories are fully visible to the audit team. On Polymarket, however, users can trade simply by connecting a decentralized wallet like MetaMask; while on-chain transactions are public and transparent, the real identities behind the addresses remain anonymous.
A deeper divergence lies in philosophy. Kalshi’s head of enforcement, Robert DeNault, explicitly defines information asymmetry as a violation that must be aggressively punished. In contrast, Polymarket CEO Shayne Coplan has publicly expressed a fundamentally different view: insider trading is a “feature, not a flaw,” of prediction markets.
The Beast Games champion was "spoiled" at 94% on Polymarket three weeks in advance—perhaps a direct result of this arbitrage opportunity. For ordinary players without any insider advantage, participating in such event predictions essentially means serving as fodder.
From a broader perspective, the original intent of prediction markets was to transform collective wisdom into price signals. But when they become large-scale insider operations, this unfair game may reveal not collective wisdom, but the shadow of information privilege.