Retest in Trading: What 1,446 Breakouts Say About Waiting

Every breakout trader gets torn between two fears: the fear of chasing a candle that's about to reverse, and the fear of watching a real move leave without them. The retest is where those two fears negotiate.

We can add something to that negotiation that most explainers can't: a clock. Every signal our desk publishes is timestamped at entry and exit, winners and losers. Pull every completed breakout signal in the book — 1,446 of them across forex majors, metals and indices, February 24 to August 31, 2026 — and a lopsided pattern falls out:

Failed breakouts died in a median of 2 hours. Winners took 11 to pay.

Hold that asymmetry in your head. It's the entire case for the retest, and we'll come back to it with the full table.

What a retest actually is

Price pushes through a level everyone was watching — yesterday's high, a weekly pivot, the edge of a range. That's the break. Then, instead of running, it turns around and comes back to touch the level it just broke. That's the retest.

The idea traders care about: a level that held as resistance is supposed to act as support once it's broken (and the mirror for shorts). So the retest is the market's second offer. You skipped the breakout candle; here's the same trade, at the original level, with proof the level flipped.

"Break and retest" is simply the strategy of only taking the second offer.

Why price comes back at all

No mystique required — a few mechanical reasons cover most of it:

• The chasers get shaken. Breakout buyers with stops just under the level are an easy pool of liquidity. A dip back to the level fills those stops, then the move can carry on without them.
• Early buyers take profit. The traders who bought below the level sell into the breakout pop. That selling drags price back toward the level.
• Orders live at obvious prices. Limit buyers who missed the move park bids at the broken level, because it's the reference everyone shares. Their bids are what "old resistance becomes support" looks like in the order book.

And sometimes price comes back because the break was simply false. Which is most of the time: in our 1,611-signal failure study, 75.5% of tracked breakouts failed to reach their target. The retest question is really a sorting question — how do you avoid paying full price for the 75% while keeping enough of the 25%?

Retest timing, measured: failures are sprints, winners are marathons

Here's what the clock says, by timeframe. "Died" means the signal hit its stop; "paid" means it closed in profit.

Two details worth staring at:

• 41% of all failed breakouts were dead within one hour. Nearly two-thirds were dead within four.
• 67% of winners were still working after four hours. Real moves take time to travel; failures reverse almost immediately.

That's why waiting is not the coward's option — it's the statistically cheap one. A failed break usually collapses back through the level in the first few candles, which means the retest entry never triggers: the level doesn't hold, you never click, the loss belongs to someone faster than you. A real break, meanwhile, usually gives you hours. The wait filters the sprinting losers and keeps most of the marathon winners.

The honest cost: some real moves never look back. You will watch a few trains leave. The table says that trade-off is priced in your favor — losers resolve three to six times faster than winners on every timeframe we track.

How our desk trades a break and retest

The mechanics, in the order they happen:

1. Mark the level before the break. A level you drew after the move is a level you're negotiating with. Ours come from the pivot ladders and session ranges our detectors watch.
2. Let the break go without you. The breakout candle is information, not an invitation. Note where it closed — a break that closes beyond the level means more than a wick through it.
3. Wait for price to return to the level. Give it a few candles of its timeframe. On M15 that's an hour or so of patience; on H1, an afternoon.
4. Demand proof it holds. You want rejection at the level — a candle that touches it and closes back in the breakout's direction. No hold, no trade. This is the step that quietly declines most false breaks for you.
5. Stop goes just beyond the level. This is the retest's quiet gift: your stop sits where the idea is provably wrong, usually much closer than the chaser's stop. Same risk budget, more size — or same size, less risk. Our position size calculator does the arithmetic.
6. Target the next level, not a feeling. Measured objectives; ours are the next rung on the ladder.

When we don't wait: news-driven moves and strong-session momentum breaks (think London open expansion) retest least often — that's exactly when the desk's continuation signals fire on the break itself rather than the pullback. If the market is sprinting, demanding a retest is how you politely miss everything.

Retest, throwback, pullback — same animal, different collars

You'll meet three words for this move, and they're worth thirty seconds so no chart book confuses you:

• Retest — the general term, and the one traders actually say: price returns to any broken level and tests it from the other side.
• Throwback — the classical-charting word for a retest upward through resistance: price breaks out, then gets "thrown back" to the old ceiling.
• Pullback — in the strict old-book sense, the downward twin: price breaks support and pulls back up to it. (In everyday trading talk, "pullback" also just means any dip inside a trend — context tells you which.)

Different collars, one animal: the market offering the level a second time. Everything in the table above applies to all three.

When the retest never comes

Accept it cheerfully. The move you missed cost you nothing — no risk spent, no stop hit. Trading pays on expectancy, not attendance: the same book that fails three breaks in four still nets positive, because the winners pay several times what losers cost (the full ledger is public, sample sizes and all).

The retest is one more filter in the same spirit as trading the right sessions and the right timeframes: it doesn't make any single trade certain. It shifts the population of trades you take away from the fast deaths and toward the slow winners.

Methodology: 1,446 completed breakout-family signals, Feb 24 – Aug 31, 2026, across the 15 instruments our desk tracks; entries and exits timestamped at fire time. "Died" = stopped out (−1R); "paid" = closed in profit (realized R > 0). Retest frequency itself isn't measured in our book yet — when it is, it'll be published the same way everything else is.

Want the numbers as they update, every week, losers included? The Dossier is free.