Asymmetric MOO -> 11:00 Strategy Research

Loading data...
Asymmetric payoff thesis: WR in 20-50% (low), but mean_win/|mean_loss| >= 3:1, EV>0, profit_factor >= 1.8. Stops inside window 09:30-11:00. Hedge stack reduces variance. Few trades (1-3 per week per edge) - need to combine all edges into a basket.

Best Hedge Stack (LONG x SHORT)

LONG only cum SHORT only cum Combined cum

All 33 Edges (sorted by Profit Factor on test)

Edge Equity Curves (top 8 by PF)

Cumulative return % on test set (chronological).

Hedge Stack Combinations (top 20)

LONG edgeL stopL tgtL N SHORT edgeS stopS tgtS N DaysDaily meanSharpe annCum %Max DD %

Ticker Frequency (top 40)

Trade List (test set)

DateTickerSideEdgeStopTgt Gap %Return %Result

Methodology

Universe: ADV 2M+ (1808 stocks). Window: Cache 2024-04-26 to 2026-04-24 (600 days). Train/Test split: before 2025-09-01 / after.

Honest features only (known by 09:30:00 ET): gap, pm_drift, pm_dvol, prev_day_ret, sector, mcap, earnings BMO d0 / AMC d-1, beat/miss (where coverage exists).

Stops: {none, +/-0.5%, +/-1%, +/-1.5%, +/-2%, +/-3%, +/-5%}. Targets: {none, +5%, +7%, +10%}. Conservative: if stop+target hit same bar - assume stop first.

Quality gate: WR_test in [20%,50%] AND EV_test > 0 AND sign agreement train-test AND win/loss >= 3.0 AND profit_factor >= 1.8 AND N_test >= 30.

Caveats: Leveraged ETFs (TQQQ/SOXL/UVIX etc) may give fake gaps from split-adjustment - flagged per-edge. Concentration risk: small-cap pumps (CRML/UAMY/SGML/NUAI/APLD) - in live cap 1 entry/ticker/week.