When this trader buys at $0.70, they imply 70% probability. Perfect calibration = the event happens 70% of the time.
| Bucket | Bets | Expected | Actual | Error |
|---|---|---|---|---|
| 0.20-0.30 | 2 | 27% | 100% | 73.4% |
| 0.30-0.40 | 4 | 39% | 0% | 39.3% |
| 0.40-0.50 | 24 | 44% | 58% | 14.2% |
| 0.50-0.60 | 16 | 55% | 50% | 4.7% |
| 0.60-0.70 | 2 | 64% | 100% | 36.3% |
| 0.70-0.80 | 2 | 70% | 0% | 70.3% |
| 0.90-1.00 | 2 | 93% | 0% | 93.0% |
Skill measures calibration quality (0-100). Variance measures return volatility (0-100, higher = more volatile).
On-chain verification: wallet age 180 days, 2683 txs, provenance grade B. Bot score: 100/100, wash trading score: 80/100.
2000 total trades across 275 markets.
52 bets on resolved markets available for calibration scoring.
Calibration error: 21.5% — needs improvement.
Skill: 44/100 (calibration quality). Variance: 100/100 (higher = more volatile returns).
Brier Skill Score: -15.2% vs naive baseline (>0% = better than always predicting base rate).
Brier decomposition: REL=0.0999 RES=0.0609 UNC=0.2500.
Log loss: 0.8006 (skill: -15.5% vs naive). Lower log loss = better calibration on rare events.
This trader shows some skill signal, but not enough to clearly distinguish from luck. More data needed.
Confidence: C/55 ± 8 (medium confidence, 52 resolved bets). Moderate confidence — score may shift as more markets resolve.
Methodology: Brier Score Decomposition (Murphy 1973), Log Loss, On-Chain USDC Verification. Same approach used by IARPA to identify superforecasters.