Accuracy — lstm_s2
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 | +0.0713 | 12.5 | +0.84% | 54.1% | 53.3% | 0.0357 | 5.691646 |
| +7d | 265 | 178k | +0.0828 | 17.0 | +2.19% | 57.3% | 58.0% | 0.0947 | 16.098387 |
| +30d | 265 | 177k | +0.1100 | 24.7 | +7.89% | 57.4% | 65.6% | 0.2129 | 33.215407 |
By month
| month | IC +1d | hit +1d | LS +1d | IC +7d | hit +7d | LS +7d | IC +30d | hit +30d | LS +30d | days |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-01 | +0.082 | 56.4% | +0.91% | +0.092 | 62.8% | +1.04% | +0.088 | 47.6% | +8.17% | 28 |
| 2025-12 | +0.048 | 53.2% | +0.80% | +0.069 | 58.3% | +4.06% | +0.076 | 63.4% | +4.85% | 31 |
| 2025-11 | +0.053 | 49.6% | +1.39% | +0.033 | 44.8% | +2.64% | +0.065 | 61.4% | +6.77% | 30 |
| 2025-10 | +0.062 | 54.4% | +0.45% | +0.056 | 63.7% | +0.33% | +0.088 | 45.9% | +10.41% | 31 |
| 2025-09 | +0.067 | 57.6% | +0.55% | +0.053 | 70.6% | +0.43% | +0.105 | 84.2% | +3.24% | 30 |
| 2025-08 | +0.073 | 50.0% | +0.46% | +0.091 | 49.5% | +1.74% | +0.094 | 57.8% | +6.76% | 31 |
| 2025-07 | +0.093 | 53.7% | +0.94% | +0.118 | 54.0% | +3.06% | +0.110 | 51.4% | +8.60% | 31 |
| 2025-06 | +0.070 | 59.1% | +0.79% | +0.091 | 69.7% | +2.13% | +0.125 | 62.2% | +8.04% | 30 |
| 2025-05 | +0.102 | 53.0% | +1.38% | +0.161 | 38.0% | +4.75% | +0.281 | 37.8% | +16.12% | 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.