Rolling projection DOE
Find the projection setup whose median is closest to actual outcomes.
Replay completed anchors and rank horizon × sample-size settings by projection error.
Best rolling setting
3D · n=80
DOE results snapshot · Window: 2026-01-05 → 2026-01-22 · anchors: 52 · latest snapshot: 2026-01-27 · computed at: 2026年7月9日 下午7:04 HKT
DOE setting comparison
Same idea as score-vs-outcome analysis: compare each parameter setting by observed actual-minus-median error over rolling completed anchors. Lower median abs error and lower bias win.
Best setting 3D · n=80: median projection vs actual return
X = Median projection return. Y = actual return. Solid line = linear trend. Orange line = bucket average. Checked: 40 · Pending: 0 · Avg: -0.57%
Best setting 3D · n=80: median projection vs forecast error
X = Median projection return. Y = actual - median. Solid line = linear trend. Orange line = bucket average. Checked: 40 · Pending: 0 · Avg: -0.96%
DOE ranking
Ranked by a composite DOE score: lower median/mean absolute error, lower bias, enough anchors, and better IQR coverage.
| Rank | Horizon | n | Median abs | Mean abs | Bias | Inside P25/P75 | Anchors | DOE score |
|---|---|---|---|---|---|---|---|---|
| #1 | 3D | 80 | 1.46% | 2.32% | -0.96% | 52.5% | 40 | 1.555 |
| #2 | 3D | 160 | 1.55% | 2.27% | -0.96% | 52.5% | 40 | 1.582 |
| #3 | 3D | 120 | 1.6% | 2.33% | -0.99% | 55% | 40 | 1.627 |
| #4 | 3D | 240 | 1.57% | 2.32% | -1.04% | 50% | 40 | 1.628 |
| #5 | 3D | 40 | 1.66% | 2.3% | -0.86% | 47.5% | 40 | 1.645 |
| #6 | 3D | 320 | 1.6% | 2.32% | -1.06% | 50% | 40 | 1.646 |
| #7 | 3D | 500 | 1.59% | 2.35% | -1.12% | 55% | 40 | 1.648 |