Accuracy — lstm_s1
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 | +0.0684 | 13.7 | +0.81% | 56.0% | 54.4% | 0.0356 | 5.716118 |
| +7d | 265 | 178k | +0.0894 | 18.6 | +2.97% | 61.8% | 61.9% | 0.0933 | 16.186023 |
| +30d | 265 | 177k | +0.1155 | 24.1 | +9.49% | 60.3% | 67.3% | 0.2092 | 34.240679 |
By month
| month | IC +1d | hit +1d | LS +1d | IC +7d | hit +7d | LS +7d | IC +30d | hit +30d | LS +30d | days |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-01 | +0.072 | 56.3% | +0.73% | +0.098 | 63.3% | +2.32% | +0.049 | 48.6% | +6.97% | 28 |
| 2025-12 | +0.041 | 56.4% | +0.88% | +0.060 | 60.1% | +3.89% | +0.056 | 60.5% | +6.19% | 31 |
| 2025-11 | +0.068 | 51.3% | +1.78% | +0.074 | 45.2% | +4.76% | +0.104 | 60.7% | +8.71% | 30 |
| 2025-10 | +0.058 | 59.6% | +0.59% | +0.061 | 66.0% | +0.52% | +0.079 | 43.6% | +11.40% | 31 |
| 2025-09 | +0.050 | 58.8% | +0.00% | +0.042 | 71.2% | +0.88% | +0.073 | 81.4% | +5.78% | 30 |
| 2025-08 | +0.075 | 49.8% | +0.51% | +0.106 | 49.3% | +3.90% | +0.131 | 58.0% | +7.01% | 31 |
| 2025-07 | +0.088 | 53.5% | +0.86% | +0.112 | 54.0% | +3.58% | +0.127 | 52.4% | +7.01% | 31 |
| 2025-06 | +0.072 | 59.9% | +0.71% | +0.094 | 74.6% | +2.20% | +0.168 | 64.4% | +13.21% | 30 |
| 2025-05 | +0.099 | 59.7% | +1.39% | +0.181 | 76.4% | +5.19% | +0.291 | 77.5% | +22.06% | 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.