Not who bets loudest. Who keeps being right.
Polymarket fills carry the taker's wallet, so accuracy per account is a real measurement rather than an inference. Once a market settles, every fill on it is scored, and the accounts behind the winning side accumulate a record.
| Account | Score | Win rate | Record | Volume | Avg size |
|---|---|---|---|---|---|
| shugafa | 100 | 100% | 2430/2432 | $36.0k | $15 |
| FountainCleaner | 100 | 100% | 1677/1677 | $8.7k | $5 |
| 0x50f7 size | 99 | 100% | 1694/1700 | $456.5k | $268 |
| lalala67 | 99 | 100% | 532/532 | $5.4k | $10 |
| ppyjanato | 99 | 100% | 636/637 | $81.1k | $127 |
| V1nch0u size | 99 | 100% | 447/447 | $109.1k | $230 |
| shota23 | 99 | 100% | 462/462 | $8.8k | $18 |
| thatwrong | 99 | 100% | 433/433 | $8.8k | $20 |
Why the score is not the win rate
A raw win rate rewards small samples: two correct trades reads as 100 percent and would outrank forty correct out of fifty. The score shrinks each record toward a coin flip with six pseudo-observations before rescaling, so sustained accuracy earns a high number and a lucky pair does not. The raw record is shown next to it so you can see what the shrinkage did.
Why Kalshi has no leaderboard
Kalshi's public tape is anonymous. There is no account to attribute a fill to, so there is no honest way to build a track record from it. Rather than invent one, Kalshi contributes flow rather than identity, and the smart-money signal is skipped there instead of being reported as a real zero.
See the money move
The full leaderboard, the whale feed, and the flow graph between accounts and markets.