Accuracy — lstm_s8
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 | +0.0653 | 11.9 | +0.78% | 56.0% | 54.4% | 0.0356 | 5.780942 |
| +7d | 265 | 178k | +0.0845 | 15.6 | +2.38% | 61.6% | 61.5% | 0.0932 | 16.285246 |
| +30d | 265 | 177k | +0.1151 | 23.6 | +8.10% | 61.9% | 70.2% | 0.2079 | 32.430457 |
By month
| month | IC +1d | hit +1d | LS +1d | IC +7d | hit +7d | LS +7d | IC +30d | hit +30d | LS +30d | days |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-01 | +0.074 | 56.2% | +0.88% | +0.087 | 63.2% | +2.19% | -0.005 | 47.6% | +2.69% | 28 |
| 2025-12 | +0.044 | 56.3% | +0.87% | +0.081 | 60.0% | +4.71% | +0.077 | 64.2% | +6.01% | 31 |
| 2025-11 | +0.062 | 51.5% | +1.68% | +0.079 | 46.3% | +5.63% | +0.122 | 63.3% | +15.12% | 30 |
| 2025-10 | +0.055 | 58.9% | +0.42% | +0.059 | 66.1% | -0.30% | +0.122 | 45.5% | +12.37% | 31 |
| 2025-09 | +0.060 | 58.9% | +0.26% | +0.055 | 71.1% | -0.26% | +0.092 | 84.8% | +1.24% | 30 |
| 2025-08 | +0.070 | 49.9% | +0.42% | +0.073 | 49.2% | +1.24% | +0.114 | 57.9% | +7.47% | 31 |
| 2025-07 | +0.082 | 53.5% | +1.19% | +0.110 | 53.7% | +4.54% | +0.139 | 51.1% | +9.73% | 31 |
| 2025-06 | +0.061 | 60.0% | +0.58% | +0.075 | 74.2% | +0.07% | +0.131 | 64.1% | +5.13% | 30 |
| 2025-05 | +0.085 | 59.7% | +0.78% | +0.159 | 74.5% | +3.93% | +0.275 | 83.9% | +14.09% | 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.