What is a breaker block?
A breaker block is an order block that price has closed through, after which the same zone is treated as support or resistance in the opposite direction from its original role.
The sequence is specific. A bullish order block forms, price returns to it as expected, and instead of holding it is traded through to the downside. The zone that was meant to supply buying is now the place where those buyers are offside, and the breaker idea holds that a later return to the zone from below will meet their exit orders and reverse. A bearish order block that fails becomes a bullish breaker by the mirror image.
The concept is therefore a composite. It needs an order block, which needs its own definition. It needs a failure of that block, which needs a threshold. And it needs a return, which needs a definition of what counts as touching the zone and within what time. Each of those is a parameter, and the result depends on all three.
The relationship to the plain flipped level is close. Support that becomes resistance is the same logic without the order-block machinery, and much of what is known about testing flips carries over directly.
How is a breaker different from a flip zone or a mitigation block?
A breaker is a flip zone whose original level was an order block rather than a swing point, and it differs from a mitigation block in that the breaker requires the original block to have been closed through rather than merely reached.
The distinction from a flip zone is one of construction rather than mechanism. A flip zone is drawn from a swing high or low that was broken; a breaker is drawn from the candle body of a block that was broken. The trading logic, that participants trapped on the wrong side supply the reaction, is the same, and a test that measured flips and found them conditional on their parameters should be expected to find the same for breakers.
The distinction from a mitigation block is about the depth of the failure. Vocabularies differ, but the usual line is that a mitigation block is an order block price reached and reacted from without a full break, while a breaker is one that was traded through and then revisited. If that line is not drawn with a number, the two labels get applied after the fact to whichever outcome occurred, and neither can be graded.
For testing purposes the labels matter less than the thresholds. A zone, a break condition, a return condition and a horizon define the event fully. Whatever it is called, the same four numbers decide the result.
Which parameters decide the result?
Four parameters decide the result: how the original order block is defined, how far through it price must close to count as a failure, what counts as a return into the zone, and how long the breaker remains valid before it expires.
The block definition is inherited, and it carries the same choices as any order-block test: whether the zone is the candle body or the full range, whether a displacement or a gap is required after it, and how far back the block may sit. Every choice changes which zones exist to be broken.
The failure threshold is the parameter unique to breakers. Requiring a close beyond the far edge of the block is the cleanest; requiring only a wick through it admits many events that are ordinary order-block reactions with an overshoot. The two definitions produce different sets of breakers and typically different results, and the choice has to be made before the results are seen.
The return definition and the expiry decide the denominator. A return that counts any touch, at any later time, produces a set dominated by stale zones and makes the failure case rare by construction. A return that requires a touch within a stated number of bars produces a smaller set with an honest failure rate, and that is the number worth knowing.
What makes breakers difficult to test?
Breakers are difficult to test because the stop and the entry sit on opposite edges of a narrow zone, the same zone can be labelled a block, a breaker and a failed breaker in sequence, and the events are rarer than plain order blocks because they require a failure first.
The stop problem is the practical one. The entry is at the near edge of the zone and the stop beyond the far edge, and on many timeframes both fit inside a single candle. Whichever way a tester resolves that candle decides the trade, and an optimistic resolution makes a losing rule look profitable. Finer data or a pessimistic rule is required.
The labelling problem is the analytical one. A block that fails becomes a breaker; a breaker that fails becomes something else, or is quietly forgotten. Unless the test enumerates every block, records which ones failed, and grades every return to a failed one, the chain of relabelling selects the successful cases into the sample.
The rarity problem affects the sample size. Breakers are a subset of order blocks, and order blocks under a strict definition are not frequent. A decade of hourly bars on one pair may yield only a few hundred qualifying events, which is enough for a grade but not for a fine breakdown by parameter.
How does QuantParadox define a breaker?
QuantParadox defines a breaker as an order block whose zone price closed through, and the breaker primitive fires when price later returns into that zone from the far side, with the block itself carrying the same fair-value-gap requirement as every order block the engine finds.
That inheritance is worth noting, because it fixes two of the four parameters at the platform's definitions rather than the trader's. A block on the platform is the last opposing candle before a displacement that left a gap, and a breaker is that block after a close through it. A trader whose own definition of a block differs will find the platform's set of breakers differs too.
The entry-and-stop candle is resolved with finer data where the archive has it and booked as a loss where it does not, which for a rule with a stop this close to the entry is the difference between a real result and a flattering one.
The remaining parameters, the return window and the expiry, are set in the rule rather than by the platform, and the out-of-sample split is applied to whatever is chosen. The per-instrument breakdown is the place to check whether the rule's result comes from many markets or from one.