An opening-range breakout tells you almost nothing about direction

The opening range — the high and low of the first half hour — is one of the most watched lines on an intraday chart, and breaking it is one of the most traded events on it. So we tested whether the opening range breakout (ORB) predicts price direction, produces the false breakouts it is famous for, or at least signals a bigger move.

We took every session of 22 large-cap US equities and ETFs over four years, found the first bar that closed outside the opening range, and measured the following 15 minutes. There are 20,650 of them.

By itself, and under this definition, the breakout carries almost no directional information.

What the tape says about direction#

Fifteen minutes after the breakout, price was still moving the same way 51.3% of the time. Over 30 minutes, 52.7%.

Those numbers need something to be compared against, and this is where most versions of this analysis stop. The base rate for any moment of the session in the same data is 51.4% up — US equities drift upwards, and that drift is there whether or not anything happened.

Continuation is not above the base rate. Resampling trading days, the difference is −0.03 percentage points, 95% CI −1.2 to +1.1. Whatever the breakout is doing, its directional effect is bounded inside about one point of nothing.

Splitting by side shows where the small asymmetry lives: upside breakouts continue 52.5% of the time (CI 51.2–53.9), downside breakouts 50.1% (CI 48.5–51.5). Read against a 50% coin the upside looks like something; read against the 51.4% base rate it is the drift again, showing up where drift always shows up.

The same holds if you ignore the event entirely and just ask where price is sitting. Across 235,004 arbitrary decision points, the next 15 minutes were up 51.6% of the time when price was above the opening range, 51.6% when it was inside, and 50.8% when it was below. Three states, one answer.

The false-breakout number is real, and misleading#

Here is the number that gets quoted: 46.5% of breakouts are back inside the range within 15 minutes, and 57.6% within 30. Half of them fail, and most fail eventually.

That looks damning until you ask how far outside the range the breakout actually closed.

Opening range breakout false-breakout rate by ATR distance#

Nearly half of all "breakouts" clear the line by less than a quarter of the symbol's own average range. Price closed a hair outside a level and is now sitting on it. That it wanders back is not the market rejecting a breakout; it is a rounding error with a name.

The deeper problem is definitional. "Breakout" is a binary word — inside or outside — laid over a quantity that is continuous. A close 0.01× ATR past the line and a close 1.5× ATR past it are both called the same thing, and then the failure rate of the category is reported as though the category described one event. It does not.

So the honest version of the statistic is not "breakouts fail half the time". It is "most things people call breakouts are barely outside the line, and those are the ones that come back".

Distance predicts the box, not the direction#

The obvious next thought is that the far breakouts are the real ones. The first half of that is true: past one ATR beyond the range, price is back inside within 15 minutes only 12% of the time. If your question is will it stay outside the box, distance answers it.

It does not answer the other question.

Continuation sits between 50.3% and 53.2% across every distance bucket. The strongest breakouts — the 5% that clear a full ATR and rarely come back within the quarter hour — continue 50.9% of the time, and there are only 1,113 of them, so that figure carries a confidence interval running from 47.3% to 54.5%.

That interval is the finding, not a caveat on it. The most decisive-looking breakouts on the chart are the ones where we can say least about what happens next.

Staying outside the range and going up are different questions, and only the first one has an answer here.

It does not even predict a bigger move#

The fallback position is that a breakout at least tells you something is happening — that volatility is about to expand, even if direction is unknowable.

That can be checked, but it has to be done carefully. The median breakout happens at 10:15 ET, which is one of the most volatile stretches of the day, so comparing a breakout against a session-wide average would credit the setup for the time of day it tends to occur at.

Comparing each breakout against the median move at the same bar index of the session instead — where "move" is the absolute close-to-close return over the next 15 minutes, divided by ATR — the move after a breakout is 0.575× ATR against a same-time baseline of 0.562×, and a breakout exceeds its own time-matched baseline 49.5% of the time.

Not larger. Not smaller. The same.

What we measured, and what we did not#

Four limits, because they decide how far this travels.

The universe. SPY, QQQ, DIA, IWM, SMH, XLE, XLF, XLU, XLV, AAPL, MSFT, NVDA, AMD, GOOGL, AMZN, TSLA, COIN, PLTR, WMT, PG, JNJ, KO. That list was fixed in July 2026 when the question bank was built — chosen for liquidity and for spread across sectors and volatility regimes, then tested backwards to 2022. It contains no delisted names, so it carries the usual survivorship caveat of any universe picked at the end of its own sample.

The market. These are large caps and ETFs. Opening-range setups are frequently traded on small caps and gap-ups, where float, liquidity and a catalyst behave differently. We have not measured those, and this piece does not speak for them.

The event. The first close outside the range, on 5-minute bars. Traders use many variants: a 5- or 15-minute opening range, an intrabar break rather than a close, a retest entry, a relative volume condition, a gap filter, a regime filter. Each is a different event, and one of them may well behave differently from this one. Nothing here rules that out — it rules out the bare version.

The horizon. Direction only, over a fixed 15 minutes, chosen because it is the horizon the whole question bank is built on and measured for its own reasons. Nothing here is about where a stop or a target would go, which is where a real trading decision mostly lives.

What this changes in the question bank#

Every question we generate is tagged with where price sits against the opening range — above, below or inside — and that tag has never been used to choose an answer or to weight the bank. This is why. It describes a setup for analysis and variety; it is not evidence about the outcome, and a bank built to over-represent breakouts would be built around a shape that carries no signal.

It is also part of why the filter deciding which setups become questions is about magnitude rather than about setups. A 1–3× ATR move over the horizon is a statement about whether there is anything there to read. "Price broke the opening range" is a statement about the chart's furniture.

What is left#

None of this says the opening range is uninteresting. It is one of the day's real reference points, plenty of chart readers use it for context, and where price sits relative to it changes what the rest of the chart means.

What the measurement says is narrower, and it is about a rule rather than about the level: on this data, in this market, under this definition, the breakout event by itself does not carry direction, does not carry size, and its most quoted failure statistic is mostly measuring how close to the line the close was. The parameters behind our own numbers are on the methodology page.