A breakout is price leaving an established range by closing beyond a level that has previously held. The break itself is only the event — what makes it real is the close beyond the level, participation behind the move, and price refusing to fall back inside. Everything else is a wick.
The word does a lot of work in trading, and most of it is done badly. People call it a breakout when price touches a level, when it trades one tick through, when it prints a big green candle in the general vicinity. Those are three different things, and only one of them is a break. Getting the definition right is not pedantry — it is the difference between a rule you can test and a feeling you can rationalise.
What counts as a breakout
Three things have to be true before the word applies.
- There was a level. Something contained price before — a range high, a prior day's low, the top of a consolidation, a trendline touched three times. If the line was drawn after the move, there was no level and there is no break.
- Price closed beyond it. On whatever interval you are trading. A candle that trades through the level and closes back inside did not break it; it tested it and failed. That single distinction removes most of what people call breakouts.
- The level did not immediately reclaim price. A break that is undone on the very next candle is a spike, whatever the first candle looked like.
Notice that none of those conditions is about how the candle looked, how strong it felt, or how much volume printed. Those are secondary. The primary test is positional: where did price finish, relative to a line that existed beforehand?
The four kinds of breakout you will actually meet
The literature names dozens. In practice, intraday and swing traders deal with four, and they behave differently enough to be worth separating.
| Type | The level being broken | What makes it distinctive |
|---|---|---|
| Range break | The high or low of a sideways consolidation | The cleanest to define, because both boundaries are objective and visible to everyone |
| Opening range break | The high or low of the first 15, 30 or 60 minutes of the session | Time-boxed and mechanical, which is why it is the most heavily tested — and the most crowded |
| Trendline break | A sloping line connecting successive highs or lows | The most subjective, because the line moves depending on which touches you chose |
| Pattern break | The boundary of a flag, triangle or wedge | Carries an implied target from the pattern's own height, which is where over-confidence creeps in |
Range and opening range breaks are the two worth starting with, for one reason: their levels are unambiguous. Two traders looking at the same chart will draw the same line. With a trendline, they will not, and a signal that depends on which of five candles you anchored to is not really a signal.
Why the breakout failure statistics you have read are worthless
Search "what percentage of breakouts fail" and you will be told that 50 to 70 percent of intraday breakouts fail, broken down helpfully by time frame — 68 to 72 percent on the one-minute, 40 to 45 percent on the daily, and so on to the nearest point. None of those pages names an instrument, a sample period, a data source, or a definition of failure. They are marketing content, and the numbers were invented to fill a paragraph.
This matters more than it sounds. A "breakout failure rate" is not a property of markets; it is a property of your rulebook. Define a break as one tick through the level and your failure rate is enormous. Define it as a close beyond the level that also holds a retest, and the same market gives you far fewer signals with a far lower failure rate — and you will have missed some real moves getting there. Anyone quoting a single percentage has not told you the only thing that determines it.
What systematic testing has actually found
There is real evidence, and it is sobering rather than damning. A 2026 falsification study of intraday signals in Micro E-mini Nasdaq 100 futures tested fourteen signal families — including opening range breakouts, gap strategies, volume signals and liquidity grabs — across 947 trading days of five-minute data from 2021 to 2025, under out-of-sample walk-forward validation. No signal met all the study's criteria at once. The gross edge available to next-bar-open execution was constrained to roughly 0.07 to 1.50 points per trade, against a fixed two-point round-trip cost (Mesfin, "Structural Limits of OHLCV-Based Intraday Signals in MNQ Futures", arXiv:2605.04004, May 2026).
Read that carefully, because it is easy to over-conclude. It does not say breakouts do not exist or that price never trends after leaving a range. It says that a mechanical breakout rule, fired on bar data alone and executed at the next open, does not clear its own transaction costs on that instrument over that period. The edge was there; it was smaller than the spread and commission.
The practical takeaway is about where the work goes. If the raw signal is worth about a point and costs two, then selection is everything — which levels you take, which sessions you sit out, and how large the move has to be to be worth the ticket. A breakout is a starting condition, not a strategy.
The conditions that separate a break from a spike
You cannot know in advance which break holds. You can stack conditions that historically accompany the ones that do.
- The close, not the touch. Wait for the candle to finish beyond the level. This is the single highest-value filter, and it is the subject of break and hold: why confirmation beats anticipation.
- The level was marked in advance. Pre-session levels are levels. Lines drawn during the move are commentary.
- Time of day. A break in the last twenty minutes of the session, or during a lunchtime lull, has far fewer participants behind it. See the best time of day to trade for when the liquidity actually is.
- Room to the next level. A break with a prior high sitting four ticks above it has nowhere to go. If there is no next level worth reaching, the break is not tradeable even if it is real.
- It survives the return. Broken levels get revisited. Whether the level holds from the other side is the cleanest confirmation available — see what a retest is and why it is the cleanest entry.
Where the fakes come from
Failed breaks are not random noise. They have identifiable causes, and knowing them tells you which ones to avoid rather than which ones to fear.
Resting orders cluster just beyond obvious levels — stops from traders positioned inside the range, and entry orders from traders waiting for the break. Price reaching that cluster triggers both at once, which produces a fast move through the level regardless of whether anyone actually wanted to own the instrument up there. Once those orders are filled, the fuel is gone, and price returns. That is the whole mechanism of the stop hunt, and it explains why the most obvious level on the chart is often the least reliable to trade at the touch.
The second cause is simpler: thin conditions. The same break, in the same place, means something different at 10:00 in the morning and 15:55 in the afternoon, because the number of participants required to move price is different. Holiday sessions, the hour before a scheduled release, and the final minutes before the close all produce moves that look decisive and mean very little.
Frequently Asked Questions
What is a breakout in trading?
A breakout is price leaving an established range by closing beyond a level that has previously contained it. The level can be the high of a range, a prior day's high or low, a trendline, or the boundary of a chart pattern. What makes it a breakout rather than a spike is the close: price has to finish an interval on the far side of the level, not merely touch it.
What percentage of breakouts fail?
Nobody credible knows, and the figures circulating online should be ignored. Search results routinely quote that 50 to 70 percent of intraday breakouts fail, broken down by time frame to the nearest percentage point, without naming an instrument, a sample period, or a definition of failure. Those numbers trace back to marketing blogs rather than any published study. The failure rate depends entirely on how you define a break, and no single number survives that.
How do you tell a real breakout from a false one?
You cannot tell in advance, only after the fact, which is why the question is better framed as what raises the odds. A real break closes beyond the level rather than wicking through it, holds the level as new support or resistance when price comes back to it, and happens in the liquid part of the session rather than in the last twenty minutes or during a holiday. A break that fails all three is a spike.
Should you buy the breakout or wait for the pullback?
They are different trades with different costs, not a right and a wrong answer. Entering on the break gets you the whole move but includes every failed break. Waiting for the pullback gives a tighter stop and a clearer invalidation, but the pullback does not always arrive, so you will miss the fastest moves entirely. Choose one, define it in writing, and stop switching between them mid-trade.
Bottom line
A breakout is a close beyond a pre-marked level that the level then fails to reclaim — and almost everything else that gets the name is a wick through a cluster of resting orders. The published testing suggests the raw signal is worth less than its transaction costs on its own, which means the value is in selection: which levels, which sessions, and how far the next level sits. Build the foundation with finding real support and resistance levels and technical analysis basics, and see how the room's trade callouts work if you want to see levels called before the session rather than after the move.