Do EURUSD Breakouts Actually Follow Through? 344 Events, 16 Years
Success rate (%)
"Trade the breakout" is one of the oldest instructions in the book, and one of the least examined. We took a strict, mechanical definition of a EURUSD breakout — price closing decisively beyond a tight consolidation — and measured what happened next across every occurrence in 16 years of hourly data: 344 events.
The headline: better than a coin flip, worse than the reputation
| Outcome | Count | Share |
|---|---|---|
| Success (extends cleanly) | 217 | 63.1% |
| Fakeout (closes back inside) | 127 | 36.9% |
Sixty-three percent success sounds tradeable — and it is — but a third of breakouts snapping straight back is exactly the experience that makes people distrust the pattern. The average is not the story. The story is which breakouts.
Displacement is the whole game
The single most powerful filter is how hard the breakout candle pushed past the level. The chart above is nearly monotonic:
A limp break that barely clears the level fails more often than it works. A break that displaces decisively past the level succeeds nine or ten times out of ten. This is the "conviction" principle made quantitative — the market either means it or it doesn't, and the size of the move tells you which.
The retest is not the confirmation people think it is
A popular rule says: wait for the breakout to retest the broken level, then enter. The data is blunt about this:
- 82% of all breakouts retested the level within a few hours.
- Successful breakouts retested 71% of the time.
- Fakeouts retested 100% of the time.
Because winners and losers both retest, the retest itself carries almost no information. What matters is what happens at the retest — does the level hold, or does price close back through it? The retest is a decision point, not a green light.
Timing and compression tilt the odds
Two context filters moved the numbers meaningfully:
- Session timing. Breakouts during the London-open window (roughly 07:00–08:00 UTC) succeeded ~70–72% of the time; breakouts in the quiet Asian session succeeded only ~52%. Liquidity has to be present for a break to run.
- Compression tightness. The tighter the consolidation before the break, the higher the success and the larger the follow-through — the "coiled spring" thesis holds up in the data.
What this means
The exact entry, stop, and target parameters we run stay in the engine — a published edge is a dead edge. But the durable lessons transfer to any instrument:
- A breakout is a claim; displacement is the evidence. Size the candle before you trust the level.
- The retest tells you almost nothing on its own — only its outcome does.
- Breakouts need liquidity to run. The same pattern is worth far less in a dead session than at the London open.
Sixty-three percent becomes something much better once you refuse to take the weak ones. That refusal — not the entry trigger — is the edge.
Studies like this become the filters inside a documented playbook — the research → playbook → backtest → live loop, locked to one instrument at a time. That playbook runs live on Cortiq.