A false break sounds attractive: wait for EUR/USD to close outside the 00:00–07:00 London range, then trade against the move if price closes back inside within four 15-minute candles. Does this survive a realistic cost assumption?
The question
The idea in plain words
Here we test whether a move that breaks the quiet overnight range but quickly returns inside it tends to travel back towards the range’s middle. This is a hypothesis about price behaviour, not an automatic buy signal.
Walk through the rule
- Draw the rangeFind the highest and lowest prices of 28 complete 15-minute candles from 00:00 to 07:00 London time. These become the upper and lower borders.
- Wait for a failed breakBetween 07:00 and 10:00, require a candle to close beyond a border, then close back inside within four candles. A wick that merely touches or crosses a border is insufficient.
- Trade the returnIf price first broke above, sell short at the open of the candle after the return inside; if it first broke below, buy. The stop is one pip beyond the furthest point of the break, and the target is the range midpoint.
- Limit the risk and exitConsider only stop distances of 2–40 pips and one trade per day. Close at the stop, the target or at 15:45 London time. The model risks 1% of its current account value at the stop and subtracts another one pip for trading costs.
A real observation from the test
On May 7, 2025, the actual upper border was 1.13669, the lower 1.13248. Price broke above and returned; the model sold at 1.13659, aimed for 1.13458 and placed its stop at 1.13741 (8.2 pips away). The stop was hit in the entry candle. With the one-pip cost, this trade lost -1.12R. The chart shows this very trade.
What the numbers mean: One R is the distance from entry to stop before trading costs. A gain of +1R equals that planned price risk; after an extra pip of costs, a stopped trade can lose more than −1R. The win rate counts profitable trades, but does not show whether winning trades outweigh losses and costs.
Exact trading rules
Build the range from 28 complete M15 BID candles. Look for the first close outside it between 07:00 and 10:00 London time. After a close back inside within four bars, enter at the next open against the breakout; set the stop one pip beyond the break’s extreme, target the range midpoint and close any open position by 15:45. Accept stop distances of 2–40 pips and at most one trade per day. If stop and target touch in one candle, count the stop first.
M15 · 00:00–07:00 → 07:00–10:00
1 pip · 1% · Europe/London
Real market charts
EUR/USD · Historical data · M15
2025-05-07 · BID · Europe/London
Net result · USD
2024-01-01 — 2026-09-18 · 1% / Trades
Historical data: Dukascopy · EUR/USD M15 · BID · Europe/London. 2024-01-01 — 2026-09-18.
What the test found
Across 336 trades, +56.50% finished profitable after the stated cost. The sum was -20.22R and the model account returned -19.68% with -26.73% maximum drawdown. Yearly results differ: the sample does not support a stable positive edge.
Trades: 336 · Win rate: +56.50% · Profit factor: 0.87 · Maximum drawdown: -26.73%
Profit factor compares total winning amounts with total losses. Below 1 means the losses were larger.
| Year | Trades | Net result | Win rate |
|---|---|---|---|
| 2024 | 133 | -11.34R | 57.1% |
| 2025 | 120 | +1.48R | 59.2% |
| 2026 | 83 | -10.36R | 51.8% |
Costs and robustness
Before the assumed one-pip round-trip cost, the same trades produced +21.93R; after that cost they produced -20.22R. At two pips the total falls to -62.39R. The calculation uses BID-only OHLC plus a fixed cost allowance; it cannot reconstruct a broker’s actual ask quotes, spread widening or slippage.
Where the model can fail
The test omits incomplete Asian sessions and does not claim a fill inside a 15-minute candle. A stop and target hit in one bar is resolved against the model. Holiday liquidity, execution outages, news events and the chosen 2–40-pip risk filter could change real outcomes. The rules and cost assumption must be fixed before forward testing.
Editorial verdict
The cost-adjusted result is negative, despite a win rate above 50%. On these data this is a useful example of why a plausible chart pattern is not automatically a tradable strategy. A new untouched sample and actual bid–ask execution records would be needed to revisit the hypothesis.
Data and method
Dukascopy · EUR/USD M15 · BID · Europe/London · 2024-01-01 — 2026-09-18.