VIGIL Trust Score for the AI Agent Economy
andikundimangaandi
0x758d6ff3fe2dd3b5e7ff22bb2b83b3680fe43ebb
Polymarket Trader
DANGER
High probability of further losses.
Calibration Analysis
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 | 26 | 16% | 8% | 8.8% |
| 0.20-0.30 | 2 | 28% | 0% | 28.0% |
| 0.70-0.80 | 48 | 77% | 21% | 56.2% |
| 0.80-0.90 | 24 | 82% | 8% | 74.0% |
Calibration Error: 47.6%
Reliability (CAL)
0.2867
Lower = better calibrated
Resolution (RES)
0.0044
Higher = stronger opinions
Brier Skill Score
-233.4%
vs naive baseline
Log Loss
1.0541
Skill: -160.3% vs naive — sensitive to rare events
Skill & Variance Analysis
Skill measures calibration quality (0-100). Variance measures return volatility (0-100, higher = more volatile).
Signals
✓ Well-diversified: 1921 unique markets
✓ On-chain: human-like trading patterns
✓ On-chain: diverse counterparty network
⚠ Weak Brier score: 0.4014
⚠ Overconfidence bias: 0.476
⚠ High luck component: 97/100 — returns may not persist
⚠ Brier Skill Score: -233.4% — performing worse than naive baseline
⚠ Low resolution: forecasts cluster near base rate — no genuine opinions
⚠ Log loss penalty: severe overconfidence on wrong bets detected
⚠ Receive-only wallet: zero outbound transactions — possible proxy/settlement address
⚠ Low win rate: 14%
⚠ Concentrated portfolio: 100% in single market
On-Chain Verification (Polygon)
Protocols: USDC.e (Bridged)
PnL verified: $512.38 gap between API and on-chain USDC
✓ Heavy on-chain activity: 2839 txs
✓ Significant USDC flow: $11,506
✓ Human-like transaction patterns
✓ Diverse trading network: 33 counterparties
⚠ Receive-only wallet: never sent a transaction
Reasoning
On-chain verification: wallet age 91 days, 2839 txs, provenance grade B. Bot score: 10/100, wash trading score: 0/100.
2000 total trades across 1921 markets.
100 bets on resolved markets available for calibration scoring.
Calibration error: 47.6% — needs improvement.
Skill: 4/100 (calibration quality). Variance: 97/100 (higher = more volatile returns).
Brier Skill Score: -233.4% vs naive baseline (>0% = better than always predicting base rate).
Brier decomposition: REL=0.2867 RES=0.0044 UNC=0.1204.
Log loss: 1.0541 (skill: -160.3% vs naive). Lower log loss = better calibration on rare events.
What Does F/34 Mean?
No demonstrated forecasting skill. This trader performs at or below random chance based on available data.
Confidence: F/34 ± 8 (medium confidence, 100 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.