Break and retest is the patient trader’s breakout: instead of buying the candle that clears the level, you let it go, then enter when price returns to the broken level and holds it — old resistance acting as new support, or the reverse. The appeal is structural. You enter at the level rather than a full candle beyond it, so the stop is tighter; the return-and-hold doubles as confirmation that the breakout was real; and it spares you the worst feeling in trading, chasing a green candle. It may be the most-taught entry pattern on the internet.
It is also quietly two strategies in one — a breakout definition and a retest definition — and the retest half is the part nobody pins down: how close must price come back, within how many bars, and what invalidates the trade? Our engine ships an N-bar breakout primitive today, but not a return-to-broken-level retest primitive, and we publish only numbers our own deterministic engine produced on real data. So the rules below pre-register the exact test: the plain-breakout baseline is already runnable on Secuora, and the retest results appear here the moment the primitive ships.
The exact rules we’ll test
- Level: the high or low of the last 20 15-minute candles — an obvious, recent extreme, defined by arithmetic rather than eyeballing.
- Breakout: a 15-minute CLOSE beyond the level (close, not wick; first cross only — staying beyond does not re-signal).
- Retest: within the next 20 bars, price trades back to within 0.1×ATR(14) of the broken level and prints a candle that touches it and closes back in the breakout direction.
- Entry at that retest candle’s close; no qualifying retest within 20 bars means no trade — missed runners are part of the strategy’s honest stats.
- Invalidation and exit: a close back through the level before entry cancels the setup; after entry, stop beyond the retest extreme, target 2R, risk 1% per trade, 0.05% commission per side.
- Missing primitive: the engine ships the N-bar breakout detector today, but not the return-to-level retest detector — results publish when it lands.
Backtest data: in the queue
We publish only numbers our own deterministic engine produced on real data — no borrowed or estimated stats. This setup needs a detection primitive the engine doesn’t ship yet, so its 12-month results will appear here the moment that lands. The strategies with completed research are on the strategies hub, and you can already practice this setup bar-by-bar in the free replay demo.
What counts as a level — and what counts as a retest?
Every part of this pattern sounds objective until you write it down. The level might be a confirmed swing, a range boundary, an N-bar extreme or a round number; the breakout might be a wick beyond it or a close beyond it; and whether a second break of the same level counts again splits textbooks down the middle. Our pinned reading — 20-bar extreme, close-through, first cross only — is one defensible choice, stated so the eventual numbers can’t be quietly re-interpreted.
The retest is looser still. An exact touch almost never prints, so you need a tolerance band (we pin 0.1×ATR) and a time window (we pin 20 bars) — and the window creates the pattern’s built-in survivorship trap: the strongest breakouts often never come back, so a retest-only entry silently skips some of the best moves while dodging some of the worst. Whether that trade-off nets out positive after costs is an empirical question, not a philosophical one — which is exactly why the retest primitive is worth shipping and testing.
How to backtest break and retest on Secuora
You can test the breakout half mechanically today; the retest half is a replay exercise until that primitive ships.
- Run the plain-breakout baseline first at /backtest/ai — "long on the first 15-minute close above the 20-bar high, structure stop, 2R" compiles today through the N-bar breakout primitive.
- Then open the free replay demo at /backtest/demo and mark an obvious level — a recent extreme or range boundary — with the drawing tools.
- Step forward bar by bar until a candle CLOSES beyond the level, then keep replaying and wait for price to come back to it within your window.
- Sign up free and enter on the retest trigger in the replay terminal with a simulated order — a touch of the level plus a close back in the breakout direction — stop beyond the retest extreme, take-profit at 2R.
- Journal both populations — retest entries taken, and breakouts that never returned — because the no-retest cases are the half of the data most traders never see.
Frequently asked questions
What is the break and retest strategy?
Instead of buying a breakout as it happens, you wait: let price close beyond a meaningful level, then enter when it returns to that level and holds it — the old ceiling acting as the new floor (or the reverse for shorts). The appeal is a tighter stop, built-in confirmation, and never chasing an extended candle.
Do all breakouts get retested?
No — and that is the strategy’s central trade-off. Strong breakouts often run without looking back, so a retest requirement filters out some of the best moves while protecting you from some of the worst fakeouts. How often each case occurs is an empirical question, which is why our pre-registered rules fix a retest window (20 bars) instead of "wait and see".
Is break and retest better than entering on the breakout itself?
Unknown until both run on the same data with the same costs — that comparison is the entire point of this page. The plain breakout half already compiles in our engine, and our published Opening Range Breakout research shows how demanding raw breakout entries are once fees are on; the retest variant gets its own numbers when the return-to-level primitive ships.
Can I backtest break and retest on Secuora today?
The breakout half, yes — the AI backtester at /backtest/ai ships an N-bar breakout primitive, so the no-retest baseline runs today. The retest leg is manual for now: mark the level in the free replay demo, then sign up free and step forward bar by bar in the replay terminal with simulated orders, pushing the results to your journal until the primitive lands.