Original | Odaily Planet Daily (@OdailyChina)
Author | Azuma (@azuma_eth)
Barron’s reporter Nick Devor posted a very “strange” phenomenon on X this morning — on Polymarket’s “2028 Presidential Election” prediction market, approximately 70% of trading volume is concentrated on candidates with near-zero likelihood (real-time probability less than 1%).
For example, the highest trading volume was for NBA star LeBron James ($48.41 million), followed by celebrity socialite Kim Kardashian ($33.84 million), and even further down were candidates who don’t even hold U.S. citizenship (and thus don’t meet eligibility requirements), such as the world’s richest person, Elon Musk ($23.14 million), and New York City Councilmember Zohran Mamdani ($18.39 million)...
As for the true high-probability candidates, such as Vice Presidential frontrunner JD Vance (trading volume: $10.58 million), California Governor Gavin Newsom (trading volume: $15.71 million), and Secretary Marco Rubio (trading volume: $9.32 million), their trading volumes are all significantly lower than those of the aforementioned "influencer candidates."
According to official Polymarket data, the total trading volume for the "2028 Presidential Election" has reached $549 million, making it the platform's most popular betting event. However, examining the trading volumes of the 36 candidates reveals the counterintuitive situation described above. Why is this? Are people out of their minds, betting on candidates who are nearly or completely unlikely to win?
The answer is naturally not so. Earlier this year, Odaily published an article titled “Who Is Betting Against the Market’s Common Sense?”, which used examples such as “the Second Coming of Jesus” and “the Flat Earth theory” to explain that participants trading or providing liquidity on these absurd event markets can be categorized into three groups: “lottery players, bots, and reward hunters.”
Nick Devor's own explanation aligns with our third point: he found that multiple large addresses held equal amounts of YES and NO shares on the same candidate, essentially capturing risk-free profits from Polymarket's subsidies—Polymarket provides an annualized 4% position reward on certain market positions based on total position value to maintain long-term pricing accuracy, and the "2018 Presidential Election" was one such subsidy event.
Nick Devor noted that a 4% annualized yield exceeds the current U.S. Treasury yield (3.98% for 5-year Treasuries), making such low-risk holdings—like buying NO shares of James or Kardashian (it doesn’t matter who’s trending, buy whichever is popular)—more attractive to whales for capturing this yield; holding both YES and NO shares simultaneously enables risk-free profit.
As for why some users hold small amounts of YES on such low-probability candidates, another X user, A5 (@probablythenuts), explained that in this multi-option market, Polymarket offers a feature that allows users to convert a set of NO shares into the corresponding set of YES shares.
Many users utilize this feature due to considerations of liquidity depth or quote quality—instead of directly buying YES shares for the candidates they believe will win, they first buy NO shares for those they believe will not win, then convert those NO shares into a corresponding set of YES shares. Additionally, they can simultaneously buy NO shares for multiple candidates and, after conversion, hold a corresponding set of YES shares for other candidates, including any new candidates added to the event in the future.
So, users trading James and Kardashian on Polymarket’s “2028 Presidential Election” event are neither crazy nor foolish—they are either seeking stable annualized returns or better execution pathways; the actions may seem absurd, but they are still driven by rationality.