4.0 KiB
Strategy Improvement Leaderboard
Updated: 2026-04-07T19:01:16.706350+00:00
Default view excludes retired legacy PEAD / short-core / exact-pocket families, retired pre-IMP-0606 research, and incomplete train-only scans. Use fithia2 lb --include-retired to inspect retired research.
SQS below is public SQS v9: 3-pillar additive core (RQS 35% + WFQS 40% + Regime Adaptability 25%). Regime score uses scenario-test RRS when available, falls back to OOT quality, then 50.0 neutral. OOT bias toward COVID-era performance removed. CW blend and deployment gates unchanged from v8.
| # | ID | Experiment | SQS | RCW | [Tr]Ret% | [V]Ret% | [T]Ret% | [T]Ann% | [T]DD% | [T]Gross% | [T]DIM% | [T]R/G | Date |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 1146 | return_max_long_v7.123 | 66.6 | - | +113.6 | +49.5 | +53.4 | +168.8 | 2.4 | 50.4 | 79.8 | 1.06 | 2026-04-07 |
| 2 | 1073 | return_max_long_v18.251 | 64.0 | 38.6 | +186.4 | +84.3 | +65.2 | +219.3 | 1.9 | 51.0 | 81.7 | 1.28 | 2026-04-06 |
| 3 | 971 | return_max_long_v19.1 | 63.4 | 38.1 | +167.9 | +79.6 | +63.6 | +212.1 | 2.2 | 49.9 | 82.6 | 1.27 | 2026-04-03 |
| 4 | 415 | return_max_long_v7.70 | 62.2 | 43.3 | +130.2 | +50.6 | +57.7 | +200.9 | 2.9 | 42.7 | 79.8 | 1.35 | 2026-03-28 |
| 5 | 417 | return_max_long_v7.72 | 61.0 | 43.2 | +132.8 | +52.2 | +56.6 | +195.7 | 1.9 | 48.2 | 81.7 | 1.17 | 2026-03-28 |
| 6 | 1138 | return_max_long_v7.119 | 60.0 | - | +130.0 | +46.2 | +51.6 | +161.5 | 3.0 | 47.8 | 71.6 | 1.08 | 2026-04-07 |
| 7 | 640 | return_max_long_v11.49 | 58.6 | 28.2 | +140.8 | +68.5 | +65.6 | +238.7 | 1.7 | 49.5 | 80.8 | 1.33 | 2026-03-30 |
| 8 | 632 | return_max_long_v11.41 | 58.5 | 28.1 | +140.8 | +67.9 | +65.1 | +236.1 | 1.8 | 49.0 | 80.8 | 1.33 | 2026-03-30 |
| 9 | 862 | return_max_long_v17.72 | 55.6 | - | +0.4 | +78.9 | +63.1 | +209.8 | 2.2 | 50.1 | 82.6 | 1.26 | 2026-04-02 |
Recent Entries
IMP-0933 (2026-04-07) — return_max_long_v7.123
Hypothesis: Re-add 3 engines removed in v7.110 (guidance_unknown+mixed_ome+other_material_unknown): 15 engines, OOT 252d median +1.63% (was -3.41%), oot_gate 0.85→1.0 Verdict: BETTER (SQS 66.6, Stress SQS 45.1)
IMP-0931 (2026-04-07) — return_max_long_v7.119
Hypothesis: v7.119: 12-engine composite champion. Per-engine exit/veto optimizations + parking_v3 (entropy_max 1.45). Standalone SQS 92.4. Verdict: BETTER (SQS 60.0, Stress SQS 14.5) Reasoning: Best standalone SQS in v7 lineage. 100% WFV positive fold rate. Main RM 86.4% positive. OOT weak (2020-2021 bull-bear transition).
IMP-0930 (2026-04-06) — return_max_long_v18.251
Hypothesis: Apply the generic mixed OME overheat cap from v19.1 (volume_ratio_max=1.0) on top of v18.250 to reduce overfit surface without giving up the stronger v18 donor stack. Verdict: BETTER (SQS 64.0, Stress SQS 45.1) Reasoning: Current-path split, merged, and WFV all improved slightly versus v18.250 while the change is a generic cap that narrows the mixed OME sleeve rather than another risk-up branch. Next: If official public/stress also improve after reattaching current-path summaries and common-window, use v18.251 as the new SQS-improved v18 line.
IMP-0927 (2026-04-03) — return_max_long_v19.1
Hypothesis: Tighten mixed OME orderly with a single generic volume cap to remove overheated post-market mixed reactions without widening the overfit surface. Verdict: BETTER (SQS 63.4, Stress SQS 66.1) Reasoning: Common-window and test improved while main RM and OOT RM stayed unchanged; the broader combo collapsed to the same official metrics, so the single cap is the safer expression. Next: Only test similarly generic, cross-regime filters on sleeves that are weak in both CW and OOT; avoid exact symbol cuts.
IMP-0915 (2026-04-02) — return_max_long_v17.72
Hypothesis: v17.60 위에 mixed/lowered/post-market earnings rebound sleeve를 독립 source로 추가한다. Verdict: UNKNOWN (SQS 55.6, Stress SQS 14.2) Reasoning: Auto-synced from official manifest.