Accuracy — lstm_s4
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 | +0.0662 | 13.6 | +0.80% | 53.0% | 52.7% | 0.0357 | 5.694870 |
| +7d | 265 | 178k | +0.0875 | 15.5 | +2.34% | 56.5% | 56.1% | 0.0949 | 16.233114 |
| +30d | 265 | 177k | +0.1043 | 19.6 | +7.79% | 59.2% | 66.8% | 0.2099 | 33.151904 |
By month
| month | IC +1d | hit +1d | LS +1d | IC +7d | hit +7d | LS +7d | IC +30d | hit +30d | LS +30d | days |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-01 | +0.075 | 54.7% | +0.83% | +0.099 | 60.1% | +1.30% | -0.050 | 47.3% | -4.01% | 28 |
| 2025-12 | +0.040 | 51.0% | +1.03% | +0.079 | 55.4% | +4.95% | +0.107 | 63.5% | +10.03% | 31 |
| 2025-11 | +0.068 | 53.6% | +1.54% | +0.090 | 46.9% | +4.32% | +0.124 | 62.7% | +14.05% | 30 |
| 2025-10 | +0.055 | 53.2% | +0.35% | +0.040 | 55.1% | -0.08% | +0.115 | 43.5% | +15.46% | 31 |
| 2025-09 | +0.063 | 55.4% | +0.45% | +0.041 | 66.9% | -0.12% | +0.021 | 82.9% | -1.74% | 30 |
| 2025-08 | +0.077 | 51.3% | +0.48% | +0.104 | 49.7% | +1.99% | +0.163 | 58.0% | +7.76% | 31 |
| 2025-07 | +0.093 | 54.0% | +1.15% | +0.129 | 57.9% | +4.42% | +0.112 | 51.6% | +5.01% | 31 |
| 2025-06 | +0.059 | 55.0% | +0.50% | +0.080 | 69.1% | -0.08% | +0.133 | 63.6% | +5.93% | 30 |
| 2025-05 | +0.066 | 48.2% | +0.87% | +0.140 | 45.0% | +4.82% | +0.231 | 60.1% | +19.23% | 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.