Accuracy — lstm_s7
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 | +0.0701 | 13.1 | +0.71% | 56.0% | 54.4% | 0.0356 | 5.778596 |
| +7d | 265 | 178k | +0.0876 | 17.3 | +2.58% | 62.1% | 62.0% | 0.0930 | 16.477448 |
| +30d | 265 | 177k | +0.1071 | 20.4 | +8.75% | 62.2% | 70.6% | 0.2075 | 32.750338 |
By month
| month | IC +1d | hit +1d | LS +1d | IC +7d | hit +7d | LS +7d | IC +30d | hit +30d | LS +30d | days |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-01 | +0.079 | 56.2% | +0.97% | +0.102 | 63.4% | +3.01% | +0.007 | 47.5% | +4.49% | 28 |
| 2025-12 | +0.050 | 56.6% | +0.97% | +0.085 | 60.6% | +4.49% | +0.095 | 64.5% | +5.75% | 31 |
| 2025-11 | +0.073 | 51.5% | +1.60% | +0.094 | 46.3% | +5.91% | +0.119 | 63.6% | +12.80% | 30 |
| 2025-10 | +0.050 | 58.9% | +0.36% | +0.049 | 66.1% | +0.41% | +0.089 | 45.1% | +15.93% | 31 |
| 2025-09 | +0.062 | 58.8% | +0.14% | +0.047 | 71.3% | +0.27% | +0.055 | 85.1% | +2.18% | 30 |
| 2025-08 | +0.074 | 49.9% | +0.33% | +0.057 | 49.3% | +0.82% | +0.101 | 58.0% | +6.69% | 31 |
| 2025-07 | +0.083 | 53.5% | +1.01% | +0.099 | 53.9% | +2.93% | +0.090 | 51.2% | +3.96% | 31 |
| 2025-06 | +0.070 | 59.9% | +0.47% | +0.093 | 74.6% | +0.70% | +0.137 | 64.2% | +9.81% | 30 |
| 2025-05 | +0.097 | 59.9% | +0.54% | +0.188 | 77.5% | +5.39% | +0.314 | 85.7% | +19.47% | 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.