Methodology
Every other prediction-market leaderboard sorts by total profit. That rewards one thing: having made money. It doesn't tell you whether that money came from real skill or one enormous lucky bet. Sirtio Score is our attempt to tell the difference.
Trader A
+$500,000
One long-shot bet on a single election market. Got it right. Total career trades: 3.
Trader B
+$500,000
A steady 8% edge across 1,400 trades over a year, spanning a dozen categories.
Same profit. A raw-PnL leaderboard ranks them the same. Trader B is the one worth paying attention to. Sirtio Score exists to tell them apart.
Sirtio Score uses a statistical technique for estimating true skill from a limited, noisy sample. Each trader's raw average return gets blended toward the tracked pool's average, weighted by how much real history exists on that wallet -- a few lucky trades get pulled back toward average, while a long, consistent record gets judged mostly on its own numbers. The final score is how far that blended estimate sits from break-even, relative to how confident the model actually is in it. We keep the exact formula, weights, and constants private; what's above is the real shape of it, not a simplified stand-in.
Every trader with a computed Sirtio Score lands in one of six tiers, fixed at points along the 0-100 scale -- not a forced top-5%-gets-a-trophy ranking. A tier can sit empty if nobody in the current pool has actually earned it, or hold most of the pool if that's genuinely where performance clusters.
Aug 26, 2026: fixed a bug where a losing position that got redeemed on Polymarket (settled for $0) was being counted as a win. Some traders' realized PnL dropped as a result -- correctly. Aug 27, 2026: that fix surfaced a separate issue in the scoring math that was compressing every trader's score, so the top-ranked wallet on the site was reading well below what the data actually supported. Both are fixed; scores reflect real performance again, and tiers moved from a forced top-5% ranking to the fixed cutoffs described above.