Scoreboard

Every strategy we’ve tested, in one table — the winners, the maybes, and the dead. Profit factors are net of slippage on the stated window. ✅ Survived means it held an edge through a full-year test. 🔬 In testing means the short-window result is promising but the year-long out-of-sample gauntlet isn’t done. ❌ Failed means it died under honest testing — those stay on the board, because the failure rate is the base rate.

The board

StrategyMarket / TFResultWindowVerdict
Opening Range Breakout (IB structure)NQ / 1mPF ~1.7365d + slippage✅ Survived
IB bottom-quartile close, longNQ / 5mPF ~1.5365d✅ Survived
Linear-regression stretch fade, short-onlyNQ / 5mPF ~1.3365d✅ Survived (thin)
Gap-down half-fill longYM, ES / 5mPF ~2.6–3.090d🔬 In testing
Session VWAP reclaim, long-onlyNQ / 5mPF ~2.0~100d🔬 In testing
3pm power-hour continuationNQ / 5mPF ~3.090d🔬 In testing
First-FVG-of-the-open continuationNQ / 5mPF ~1.3~100d🔬 In testing
15m fair-value-gap first-touch bounceNQ / 5mPF ~2.1~100d🔬 In testing
“Magic hour” overnight sweep-reclaim fadeNQ / 5mPF 0.86 (claimed 2.17)faithful reproduction❌ Failed
RSI mean-reversion, tight target / wide stopNQ / 5m1.40 → 0.91collapsed at 365d❌ Failed
1-minute VWAP-cross trend flipQQQ / 1mPF 0.97~1yr❌ Failed
NQ shorts (breakout variants, as a class)NQ / allnegative edge2025–26 uptrend❌ Failed

The graveyard, briefly

  • Magic hour sweep — a shared dashboard with a suspiciously smooth equity curve. Faithful reproduction with costs: losing system. Cause of death: no transaction costs + cherry-picked hours. Full autopsy here.
  • RSI mean-reversion — great on 100 days, dead on 365. Cause of death: tight-target/wide-stop geometry hiding a fat left tail the short window never sampled.
  • VWAP trend flip — legit structure, but the win/loss math only worked in one regime. Cause of death: regime dependence.
  • NQ shorts — nearly every short-side variant we test bleeds in a market that goes up. Cause of death: fighting the tape.

Rows move down this page over time — that’s the point. When an “in testing” strategy clears (or fails) its year-long retest, it gets a verdict post and the board updates. Numbers are rounded; exact parameter sets aren’t published for the live survivors. Everything here is subject to the risk disclaimer.