What Are the Rules of Failed Breakout in Forex Price Action?

Direct answer

Failed breakout is a price-action situation where price moves beyond a previously defined boundary (often a support or resistance level, or another chart-defined range edge) but then fails to maintain that move. The “rules” of failed breakout are not about predicting outcomes; they are about using consistent, observable conditions to (1) identify the attempted breakout and (2) decide, by predefined criteria, that the move did not hold.

Because the term is used in different ways, the most important rule is to specify your own definitions before you measure anything: which level counts as “the boundary,” how you confirm the “break,” what time window you use, and what “failure to hold” means in measurable terms.

Mechanism or definition

A rule-based definition typically has four parts: boundary, breach, hold, and outcome.

Boundary (the level you are watching)

Choose a specific price level that is fixed for the test, such as a recent swing high/low or the top/bottom of a well-defined range.
The rule should state how that boundary is chosen (for example, “the most recent confirmed swing high before the event”) and whether you use closing prices, intrabar extremes, or both.

Breach (what counts as breaking the boundary) Common measurable breach rules include:

Close-based breach: the market closes beyond the boundary.
Extreme-based breach: price trades beyond the boundary at any time (e.g., intrabar high/low) even if it closes back inside.

You must pick one, because these lead to different counts. If you allow both close-based and extreme-based breaches without separating them, your results become hard to verify.

Hold (what it means for the breakout to “succeed”) A failed breakout implies that the breakout does not hold. So you need a hold criterion, such as:

Time window hold: after the breach, price must remain beyond the boundary for at least N candles (or a set number of minutes/hours).
Retest-based hold: after breaching, price revisits the boundary and does not reclaim the inside/failed direction.

Again, pick one measurable hold rule. The phrase “does not hold” is ambiguous unless you translate it into a rule that can be checked on a chart.

Failure to hold (the failed breakout trigger) A typical failure criterion is one or more of the following, depending on your chosen framework:

Return inside: within the window after the breach, the market returns to the inside of the boundary.
Sustained rejection: the market repeatedly closes back on the inside side for several candles.

A key rule here is to separate “first return inside” from “confirmed failure.” If you define failure as the first touch back inside, you will label many marginal cases as failures.

Evidence or example (rule-based, testable)

No real-time market data is assumed here, so the example is a measurement template you can apply to historical charts.

Example measurement template (assumptions stated):

Boundary: Resistance at 1.1000, defined as a prior confirmed swing high.
Breach rule: A breach occurs when a candle closes at or above 1.1000.
Hold rule: The breakout “holds” only if price remains at or above 1.1000 for at least 2 consecutive candles after the breach candle.
Failure rule: A failed breakout is recorded if, within the next 2 candles, the market closes below 1.1000.

How you would apply it:

Scan the chart for a candle that closes at/above 1.1000. Mark that candle as the breach candle.
Look at the next two candles only (your time window).
If both next candles close at/above 1.1000, you label it a “held breakout” (not a failed breakout).
If any of the next two candles closes below 1.1000, you label it a failed breakout.

This produces a rule set that you can independently verify by re-checking the same boundary and candle sequence. If you change the hold window (e.g., 1 candle instead of 2), you change the classification. That difference is part of the “rules.”

Compare two common definitions (to avoid mixing meanings)

Two people can look at the same chart and disagree on “failed breakout” because they used different breach or hold rules.

Definition A (close-based): counts only closes beyond the level.
Definition B (extreme-based): counts intrabar touches beyond the level.

If you use extreme-based breaches, many cases will look like breakouts occurred briefly but never held. If your failure rule also uses closes back inside, then those cases will heavily depend on your timeframe and candle construction.

Limitations and risks

Failed breakout is a concept about labeling and measuring price behavior, not a certainty about future movement. Several material limitations affect how reliably you can apply any rule set.

Market microstructure effects Even with fixed chart rules, trading conditions can make what “happens” differ from what the chart shows. Costs such as spreads and execution quality can affect whether a boundary breach that looks minor on the chart corresponds to real entry/exit experiences.
Timeframe sensitivity The breach and hold rules depend on your timeframe. A “2-candle hold” on one timeframe may correspond to a completely different duration than on another timeframe. A failed breakout on a short timeframe can be a normal fluctuation within a longer timeframe structure.
Boundary selection bias If the boundary is chosen differently (swing-based vs. range-based, or slightly shifted levels), the set of breaches changes. A rule set is only self-consistent if the boundary-selection method is stated and repeated.
False failure vs. genuine failure A single close back inside can be noise. If your failure criterion is too sensitive, you will label many borderline situations as failures. If it is too strict, you may miss legitimate failures.
Regime changes Historical patterns do not establish future results. Market volatility, liquidity, and behavior can change, so a rule set that labels failures in one period may label different behavior in another.

Verification or next question

To independently verify the rules of failed breakout, do three checks using the same historical dataset:

Reproducibility: Can another person apply your boundary, breach, hold, and failure rules and get the same labels? - Sensitivity: If you adjust one rule parameter slightly (e. g. , hold window from 2 candles to 3), how many labels flip?