This test uses 67,287 EUR/USD M15 bid candles from 1 January 2024 through 18 September 2026. Every timestamp is converted to Europe/London, so the session follows British daylight-saving changes rather than a fixed UTC offset. One pip is deducted from each completed trade.
Why the London open is testable—and easy to oversimplify
Liquidity and participation often increase as London opens, so the preceding Asian range offers an observable reference. But a line on the chart is not an edge by itself: false breaks, news releases and an unusually wide overnight range can erase the apparent structure.
We deliberately avoid an optimised volatility filter. The benchmark asks a narrower question: what happens if the same first-break rule is applied every eligible day, with a fixed target, a session close and realistic clock handling?
The benchmark, fixed before measurement
Asian range
Record the highest high and lowest low from 00:00 through 06:59 Europe/London. At least 24 M15 candles must be present.
Breakout window
Between 08:00 and 10:59, accept the first candle that closes beyond either side of the range.
Entry
Trade at the next candle open in the breakout direction. Only one trade is allowed per London date.
Stop and target
The opposite side of the Asian range is the stop. The profit target is 1.5 times that initial risk.
Time exit
If neither level is reached, close at the first available 16:00 candle open in London time.
Conservative accounting
If stop and target are both touched inside one M15 candle, the stop is counted first. One pip is deducted per round trip.
One real session plus the complete equity path
The price panel shows 11 September 2026, the latest completed target trade in the sample. The shaded Asian range, confirmed break, next-bar entry, stop and target are derived directly from the rule set. The equity panel includes all 550 trades, not only this example.
EUR/USD · M15 · Europe/London
2026-09-11 · BID · Asian range → Breakout window
EUR/USD · A positive total with almost no margin for error
1% · Rules were fixed before the calculation. Results include the stated cost model and are not a forecast.
A positive total with almost no margin for error
Across 550 trades, 48.2% finished positive. After the one-pip cost assumption, the aggregate result was 4.75R and the profit factor was 1.02.
At a hypothetical fixed risk of 1% of current equity per trade, compounding produced only 0.5%, while maximum drawdown reached 24.5%. The model therefore exposed far more path risk than its final return compensated.
The annual split is more revealing than the final total: 2024 lost -17.02R, while 2025 and the 2026 partial year added 10.81R and 10.96R. A rule that depends heavily on the selected starting year is fragile.
| Year | Trades | Win rate | Net R |
|---|---|---|---|
| 2024 | 225 | 44.4% | -17.02R |
| 2025 | 184 | 48.4% | +10.81R |
| 2026 | 141 | 53.9% | +10.96R |
The unresolved problems
Bid-only candles
Ask prices are not reconstructed. The one-pip deduction is a transparent approximation, not a tick-level execution simulation.
News and slippage
M15 highs and lows cannot reproduce the order queue around major releases or a fast gap through the stop.
Range quality
No width filter is used. Very narrow and very wide Asian sessions are deliberately left in the benchmark.
Calendar effects
Holidays, month-end flows and macro-event days may behave differently; the basic rule does not classify them.
How to improve the research without curve fitting
- Reserve a future year as a true out-of-sample test.
- Compare fixed one-pip costs with broker-specific spread and commission histories.
- Study pre-declared range-width buckets rather than searching for one perfect threshold.
- Report long and short trades separately and inspect news-day sensitivity.
- Forward-test the London clock conversion and next-bar execution on demo.
An instructive benchmark, not a deployable edge
The raw London Breakout idea survived costs by only a narrow margin and endured a deep drawdown relative to its return. Publishing that result is useful: it shows which attractive chart story fails to become a robust trading claim, and it provides a clean baseline for genuinely out-of-sample improvements.
Data and method
Dukascopy Historical Data Export · EUR/USD · M15 BID · UTC normalised to Europe/London
Rules were fixed before the calculation. Results include the stated cost model and are not a forecast.