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 |
|---|---|---|---|---|
| 0.10-0.20 | 2 | 18% | 0% | 17.7% |
| 0.20-0.30 | 3 | 27% | 0% | 26.9% |
| 0.30-0.40 | 19 | 36% | 0% | 36.4% |
| 0.40-0.50 | 28 | 45% | 0% | 45.1% |
| 0.50-0.60 | 38 | 56% | 0% | 55.8% |
| 0.60-0.70 | 29 | 63% | 0% | 62.7% |
| 0.70-0.80 | 3 | 75% | 0% | 75.1% |
| 0.80-0.90 | 1 | 87% | 0% | 86.6% |
| 0.90-1.00 | 1 | 96% | 0% | 96.0% |
On-chain verification: wallet age 16 days, 2089 txs, provenance grade B. Bot score: 0/100, wash trading score: 0/100.
Polymarket on-chain coverage: $6,173 in / $0 out across 37 withdrawal tx since 2026-04-18.
3100 total trades across 114 markets.
124 bets on resolved markets available for calibration scoring.
Calibration error: 51.7% — needs improvement.
Skill: 0/100 (calibration quality). Variance: 21/100 (higher = more volatile returns).
Brier Skill Score: -13.5% vs naive baseline (>0% = better than always predicting base rate).
Brier decomposition: REL=0.2829 RES=0.0000 UNC=0.0000.
Log loss: 0.7739 (skill: -9687.1% vs naive). Lower log loss = better calibration on rare events.
Timeliness: avg entry 2.6 days before resolution, 0% early mover.
Below average. The data shows poor calibration, thin evidence, or both. When this trader expresses high confidence, events don't happen at the rate they imply.
Confidence: D/40 [CI95: D→D, 39-41] (124 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.