What the engine actually trades, why it holds up, and how to read the rest of the site.
On the New York opening range, enter in the direction of the first 5-minute 9/20 EMA alignment after the range closes — but only when it agrees with the daily SMA20 trend. Target and stop are each one range-width (1R:1R). We call this signal Bt.
Three things had to be separated before the edge was clear. The raw opening-range breakout alone does not survive realistic costs. The 5-minute 9/20 crossover alone has no edge at any hour of the day — tested across ~22,000 all-day trades it is a coin flip. The edge lives in the combination: a directional break of the opening range, entered on momentum confirmation, filtered by the daily trend. The whole is far more than the parts.
The trend filter does two good things at once: it raises the per-trade edge (the counter-trend entries it removes were the weakest) and it roughly halves the trade count, which cuts the cost drag. Out-of-sample (2024 onward) the filtered edge held or strengthened — the sign that it is structural, not curve-fit.
Same signal, every session and window. The edge appears everywhere; whether it is fundable depends on the opening range being wide enough to outrun commission.
| Session / window | Gross | Net | Tier | Fundable |
|---|---|---|---|---|
| NY 60m | +0.286R | +0.216R | Primary | Yes — ~0% bust |
| NY 30m | +0.282R | +0.189R | Complement | Yes — ≤1% bust |
| NY 15m | +0.205R | +0.082R | Edge present | No — cost-bound |
| London 30m | +0.214R | +0.124R* | Add-on | Only at low cost |
| Asia 60m | +0.184R | −0.083R | Research only | No — cost-killed |
Bt gross/net per trade, ~$4/contract cost. *London shown at $2.50 (it is not fundable at $4). The lesson: wider ranges dilute cost, so the 60-minute New York window is the sweet spot.
Edge is measured in R (range-agnostic); fundability is decided in dollars against a prop-firm trailing drawdown. On a 50K account the sweet spot is 0.25–0.30% risk per trade (~$125–150). At that size the NY 60-minute signal passes a typical evaluation in the high-90s percent of simulations with a near-zero bust rate, even under pessimistic cost. Sizing is in micro Gold (MGC) so risk is granular.
Paper record — the forward, prospective track record of Bt in dollars, with backfilled history and genuinely-live trades kept separate.
Daily recaps — what each session's opening range did (Asia, London, New York), with key levels known before the day.
Research log — every idea the AI layer proposes and how it tested out. Proposals never change the live strategy until they pass validation and are manually promoted.
The Breakout · GC opening-range research. Historical and simulated results for educational purposes only — not financial advice.