Signal Performance
Transparent, honest track record from Sentinel's prediction ledger. All numbers are computed from pre-registered signal predictions and realised forward returns.
Hit Rate by Signal
Minimum 10 resolved predictions per signal. A "hit" means the signal's direction matched the sign of the 30-day forward return.
| Signal | Tier | Hit Rate | Correct / Total | Avg 30d Return |
|---|---|---|---|---|
| Conflict Intensity | 2 | 46.5% | 369 / 794 | +5.47% |
| Revenue Yoy | 1 | 50.5% | 369 / 731 | +5.24% |
| Institutional Flow | 1 | 63.5% | 380 / 598 | +5.38% |
| Options Put Call Skew | 2 | 52.7% | 313 / 594 | +4.94% |
| Social Options Interaction | 2 | 51.6% | 281 / 545 | +4.21% |
| Sentiment Combined | 2 | 50.9% | 244 / 479 | +4.20% |
| Stocktwits Net Sentiment | 3 | 55.7% | 230 / 413 | +3.22% |
| Factor Grade Quality | 2 | 50.1% | 194 / 387 | +1.51% |
| Polymarket Divergence | 2 | 55.6% | 197 / 354 | +1.68% |
| Borrow Cost Acceleration | 1 | 48.8% | 166 / 340 | +5.36% |
| Factor Grade Growth | 2 | 44.9% | 141 / 314 | +1.09% |
| Earnings Confidence | 2 | 58.1% | 169 / 291 | +10.89% |
| Price Momentum 5d | 1 | 55.4% | 150 / 271 | +7.16% |
| Hiring Momentum | 1 | 49.6% | 121 / 244 | +2.83% |
| Factor Grade Momentum | 2 | 57.8% | 85 / 147 | +2.25% |
| Factor Grade Value | 2 | 42.1% | 51 / 121 | +10.28% |
| Insider Cluster Sell | 1 | 30.2% | 35 / 116 | +15.80% |
| Earnings Tone Divergence | 2 | 50.0% | 48 / 96 | +15.87% |
| Si Daily Momentum 5d | 2 | 68.0% | 51 / 75 | +2.08% |
| Contract Momentum | 2 | 36.2% | 17 / 47 | +1.76% |
| Insider Cluster Buy | 1 | 54.2% | 13 / 24 | +1.45% |
| Options Sentiment Pcr | 2 | 66.7% | 16 / 24 | -2.25% |
Hit Rate by Market Regime
Monthly Hit Rate
Monthly accuracy of directional predictions. Green bars beat 55%, red below 45%, blue is the inconclusive zone in between. The dashed line marks 50% (coin-flip baseline).
| Month | Signals | Directional | Correct | Hit Rate |
|---|---|---|---|---|
| Jun 2026 | 45 | 34 | 18 | 52.9% |
| May 2026 | 4530 | 3301 | 1708 | 51.7% |
| Apr 2026 | 5013 | 3685 | 1919 | 52.1% |
Calibration Curve
Signal magnitude (|value|) as a proxy for confidence, bucketed against observed hit rate. A well-calibrated model follows the diagonal — points above the line mean the model is under-confident, below means over-confident.
| Confidence Bucket | Predicted Prob | Observed Hit Rate |
|---|---|---|
| 5% | 5% | 54.9% |
| 15% | 15% | 49.5% |
| 25% | 25% | 51.5% |
| 35% | 35% | 52.0% |
| 45% | 45% | 60.4% |
| 55% | 55% | 47.4% |
| 65% | 65% | 45.6% |
| 75% | 75% | 49.4% |
| 85% | 85% | 54.4% |
| 105% | 105% | 51.7% |
Signal Information Coefficients
30-day Information Coefficient (Pearson correlation between signal value and forward return). Higher |IC| = more predictive. Green = positive correlation with returns, red = inverse.
| Signal | IC (30d) | |IC| | Sample Size | Strength |
|---|
Important Caveats
- Not investment advice. Sentinel is a research tool, not a recommendation engine. Past signal performance does not predict future results.
- Look-ahead bias risk. Although signals are recorded at extraction time (pre-registered), the composite weighting is updated monthly from the same ledger. This creates a mild feedback loop that inflates apparent IC vs true out-of-sample.
- Small sample sizes. Many signals have fewer than 100 resolved predictions. Hit rates with n < 50 have wide confidence intervals and should not be treated as reliable.
- Survivorship bias. The company universe has changed over time. Companies removed from coverage are excluded from recent calculations but included in historical periods.
- Forward-only sources. Signals derived from Polymarket, Bluesky, Reddit, and StockTwits only accumulate from their deployment date. Historical depth is thin.
- Unaudited. These statistics are computed internally and have not been independently verified. We publish them in the spirit of transparency, not as performance guarantees.