Accuracy — lstm_s5
validation only · 2.3M
Summary
| horizon | days | token-days | rank IC | t | LS decile | hit (alpha) | hit (price) | MAE (alpha) | MAE (price) |
|---|---|---|---|---|---|---|---|---|---|
| +1d | 265 | 179k | +0.0510 | 10.6 | +0.64% | 55.8% | 54.4% | 0.0356 | 5.801255 |
| +7d | 265 | 178k | +0.0639 | 14.3 | +1.53% | 62.1% | 61.9% | 0.0933 | 16.339215 |
| +30d | 265 | 177k | +0.0891 | 15.7 | +5.99% | 62.3% | 70.7% | 0.2078 | 33.019014 |
By month
| month | IC +1d | hit +1d | LS +1d | IC +7d | hit +7d | LS +7d | IC +30d | hit +30d | LS +30d | days |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-01 | +0.059 | 56.0% | +0.40% | +0.079 | 63.0% | -0.47% | -0.029 | 47.5% | -4.37% | 28 |
| 2025-12 | +0.028 | 56.0% | +0.79% | +0.042 | 60.3% | +2.16% | +0.034 | 64.5% | +1.94% | 31 |
| 2025-11 | +0.047 | 51.4% | +1.43% | +0.034 | 46.3% | +1.28% | +0.064 | 63.5% | +5.03% | 30 |
| 2025-10 | +0.046 | 58.5% | +0.71% | +0.057 | 66.2% | +2.36% | +0.139 | 45.4% | +16.23% | 31 |
| 2025-09 | +0.051 | 58.8% | +0.47% | +0.056 | 71.2% | +1.78% | +0.070 | 85.1% | +7.41% | 30 |
| 2025-08 | +0.046 | 50.0% | +0.16% | +0.027 | 49.7% | +0.55% | +0.074 | 58.0% | +4.96% | 31 |
| 2025-07 | +0.069 | 53.7% | +0.76% | +0.089 | 53.8% | +1.83% | +0.108 | 51.0% | +2.22% | 31 |
| 2025-06 | +0.037 | 59.1% | +0.16% | +0.046 | 74.1% | +0.06% | +0.082 | 64.2% | +1.38% | 30 |
| 2025-05 | +0.085 | 59.6% | +0.93% | +0.174 | 78.5% | +4.77% | +0.302 | 87.6% | +22.10% | 23 |
t-stats at +7d and +30d are optimistic: consecutive windows overlap, so the effective sample size is closer to the number of distinct days than the number of rows. +1d is the clean read. Long-short spreads are before trading costs and slippage, and the tail of the top-1000 by volume is thin.