Not in our testing. Across 1005 trades on 2.54 years of XAUUSD M15 data, the opening range breakout produced a profit factor of 0.85 and lost $424.07 net of costs. We swept reward-to-risk from 1.8:1 to 4:1 and it never exceeded 0.92. We retired it.
Almost nobody publishes the strategies that did not work, which is exactly why beginners think everything works. We ran three strategies over the same data; two lost money. This is one of them, reported with the same detail we gave the one that survived.
The opening range breakout is a genuinely respected concept with real academic and practitioner heritage — Toby Crabel's work on opening ranges is a serious body of research. That is precisely why the result is worth publishing: a well-regarded idea can still fail on a specific instrument and timeframe.
Define the range formed during the London opening window, then trade a decisive close beyond it in the direction of the break, with the stop at the opposite side of the range. Entries required agreement with the prevailing trend and were restricted to the high-volume session.
Same dataset as everything else on this site: 60 000 M15 bars of XAUUSD, 2024-01-05 to 2026-07-21, with 30 points round-trip spread subtracted from every trade.
Over 1005 trades — a large sample, not a small one:
| Metric | Opening range breakout | For comparison: sweep fade |
|---|---|---|
| Trades | 1005 | 173 |
| Win rate | 32.5% | 25.4% |
| Profit factor | 0.85 | 1.12 |
| Expectancy/trade | $-0.42 | $0.37 |
| Total P&L | $-424.07 | $63.53 |
| Worst losing streak | 18 | 11 |
Before retiring a strategy it is worth checking whether the entry is sound and only the exit is wrong. So we swept reward-to-risk across 1.8:1, 2:1, 2.5:1, 3:1 and 4:1 and re-ran the full dataset each time.
Profit factor never exceeded 0.92. Unlike the sweep fade — which climbed from 0.87 to 1.12 as targets widened, documented in what reward-to-risk works best — the ORB stayed under water at every setting. That pattern is diagnostic: when no exit rescues a strategy, the problem is the entry.
The most likely explanation is that gold's opening range break is heavily contaminated by liquidity behaviour. The obvious range high and low are exactly where stop orders cluster, so price frequently pushes through them to collect that liquidity and then reverses — which is a sweep, not a breakout.
In other words, the same market behaviour that made our sweep-fade profitable is what made the breakout unprofitable. They are opposite sides of one phenomenon, and on this instrument the fade side was the one that paid. The mechanics are in liquidity and sweeps.
Its 18-trade worst losing streak also tells a story: long strings of losses are what you get when you repeatedly buy the top of a raid.
It does not prove the opening range breakout is a bad idea everywhere. It was tested on one instrument, one timeframe, one 2.5-year window, with one specific set of rules. Other markets, sessions or filters may behave differently.
What it does show is that a respected concept applied mechanically to gold M15 lost money across a thousand trades — and that no amount of target adjustment fixed it. That is enough for us to stop allocating risk to it.
A high win rate meant nothing. The ORB won 32.5% of the time — more often than the strategy we kept — and still lost. This is the clearest possible demonstration of why we judge strategies on profit factor rather than accuracy.
The other lesson is procedural: we only know this because we tested and modelled costs. Traded live on faith, this strategy would have bled slowly enough to be blamed on discipline rather than on the strategy itself.
Not in our test. Profit factor 0.85 over 1005 trades and 2.54 years, losing $424.07 net of costs.
No. We tested reward-to-risk from 1.8:1 to 4:1 and profit factor never exceeded 0.92.
Most likely because gold's obvious range extremes are liquidity targets — price sweeps through them and reverses, which punishes breakout entries and rewards the fade.
No. We retired it and reallocated the risk budget to the one strategy that survived costs.