Scrolling through posts late at night! Polymarket’s big whale made a massive $10.32 million profit by “picking up trash”—did your “anti-hot” strategy work?

In the World Cup arena, teams are fighting for the coveted trophy, while in the prediction market, a whale is quietly cashing in.

On Polymarket, there is an address called swisstony, with total profit of $18.62 million—$10.33 million just in the most recent month. Don’t be fooled by the fact that it was registered only a year ago (July 2025); its homepage views have already surged to 922.2k.

As of July 13, this player has made nearly 140k predictions in total (139,617); it currently holds about $600k. What’s interesting is that its positions are almost entirely placed on the $FIFA World Cup semifinal on July 14, 2026: France vs Spain. It’s betting on France to lose, having already put down $1.6 million.

In addition, it heavily bets “NO” on specific score details, aiming only for stable returns of about 5% to 10%. Its overall win rate is 52.9%, which doesn’t sound especially high. But its total trading exposure is over 245k trades, and its trading volume reaches the multi-hundred-million USD level. In the Polymarket ecosystem, this is absolutely top-tier.

Public research has long said: most retail addresses lose money over the long run, and only a small number of high-frequency, systematized accounts can generate positive returns by executing at scale. swisstony is a typical example—its win rate isn’t extremely high, but it amplifies positive EV little by little through very high trade frequency and position management.

Take a closer look: from creation to now, it executes about 380 operations per day on average—16 per hour, 24/7 without stopping. It can basically be concluded that it’s an API-driven ultra-high-frequency quant trading bot. Its bio says “trash panda” (raccoon). In North American culture, raccoons are masters of surviving by rummaging through trash bins. That signature is spot-on: it’s mining profits from Polymarket’s massive data and tiny price spreads, building an empire worth tens of millions of dollars.

The profit record is even more shocking—there are 17 single-trade profits exceeding $1 million. The biggest one was a bet on Germany win/loss on June 25—betting “NO,” earning $2.22 million, with a return rate of 111.67%. From the screenshots, its entry prices have a big advantage, and single trades often slam in $400,000 to $1 million.

It likes betting heavily “NO” (no win), targeting strong teams that the market has overpriced: Germany, Paraguay (appears repeatedly), England, and Japan. Its entry prices are mostly between 35.8￠ and 53.7￠, corresponding to implied win probabilities of 46% to 64%. And in these cases, the strong teams really lost or failed to win. That’s a typical contrarian-to-the-consensus strategy.

This isn’t everything. It’s also good at going small to win big. Look at another set of data: entry prices are only between 0.2￠ and 1.2￠ (implied probabilities of 0.2% to 1.2%); each bet only puts in a few thousand USD, but once it hits, the profit exceeds $100k. These are things that the market originally deemed almost impossible—yet it uses extremely low costs to achieve high returns.

This “lottery-style” systematized execution—though most of the time it will lose everything (the probability is indeed low)—still, if it hits a few, can massively lift total profit without dragging down principal. This suggests its automated system covers a lot of niche, low-liquidity markets—those tracks are more likely to feature mispricing.

In summary, there are only two lines: large money bets on contrarian strong teams; small money bets on extremely obscure outcomes. With both running in parallel, it not only stabilizes large-scale profits, but also lifts overall return on investment through high-multiple trades. This is the homework you should copy in the prediction market.

Follow me: get more real-time crypto market analysis and insights!

#PreIPOs第二期OpenAI认购 #LAB two-day haircut 53% $BTC $ETH $SOL #World Cup champion prediction