Why do so many breakouts fail?

A price leaves a range decisively, everything looks clean — and days later it's back inside, as if nothing happened. This is one of the most common experiences in chart-reading, and unlike most pattern folklore, the mechanism behind it is well understood.

The self-defeating part of a visible level

A level everybody can see is a level where stop orders cluster. Participants who own the stock place protective stops just below an obvious floor; participants positioned the other way place them just above an obvious ceiling. Those orders are, by design, triggered automatically once the level is touched — a stop-market order becomes a market order at that instant, and a stop-limit becomes a live limit order.

Now watch what that produces. The price reaches the level. Stops trigger. A burst of automatic buying (or selling) hits the market, volume spikes, and the price moves fast through the level. That looks exactly like a genuine breakout — the volume expansion, the speed, the clean break — but the buying came from orders that were already there, not from anyone forming a new opinion.

Once those stops are exhausted, that source of demand is simply gone. If no genuine new interest follows, the price drifts back inside the range. And now the participants who joined on the break are offside, so their own exits add supply on the way back — which is why failed breakouts frequently reverse sharply rather than gently.

A worked example

Rounded and illustrative. A stock ranges between $60 and $64 for six weeks. On Tuesday it prints $64.90 intraday on triple average volume — then closes at $63.60, back inside the range. By Thursday it's at $61.20.

Read what happened: the level was touched, a burst of activity carried the price through it briefly, and by the close the range's sellers had reabsorbed it. Two details did the work: the price closed back inside, and the volume spike didn't persist beyond that single burst. Tuesday's $64.90 high is now itself a level — one that traders will watch on the next approach.

The same mechanism runs at a much smaller scale, and looks like this — thirty cents through an obvious low, and a close back above it in the same session:

What traders watch (and what they can't know)

The commonly watched distinctions, all of them descriptive:

• Close versus poke. An intraday extreme can be one impatient order; a close is where the whole session settled.
• Volume persistence. Did participation stay elevated for several sessions, or spike once and vanish?
• Behaviour on the return. Does the old ceiling now absorb selling, or does the price cut straight back through?

None of these are signals to act on, and none of them tell you in advance which case you're in. They're the questions that separate "the supply at this level was genuinely consumed" from "some stops got triggered."

The number nobody honestly has

How often do breakouts fail? Nobody knows, and that's not evasion — it's the actual state of things. The answer depends entirely on how you define a level, what counts as a break, and what counts as a failure. Tighten the definitions and the failure rate falls; loosen them and it rises. Across reasonable definitions on the same data, you can produce failure rates anywhere from modest to overwhelming.

That range is the finding. A breakout is not an object that exists in the market and can be counted. It's a classification you applied, and the statistic follows from your choices. Anyone quoting a precise failure rate has hidden those choices from you — possibly from themselves.

In the data

Whether a break held is a question about resolution. A daily bar is one row per session, so a level crossed at 10:05 and lost by 15:00 leaves a high above the level, a close below it, and nothing in between. Put a Level through a recent range edge on the month of the small-company fund below and look for exactly that shape: a wick through, a body back inside.

Only finer bars show the sequence, and the finest bars are kept for the shortest history, so the old cases worth checking are often the ones you can no longer check.

Try it now

1. On the two schematics above, write down the intraday extreme and the close of each excursion. Both went through a visible level; both were back on the wrong side of it by the bell. That pair of numbers is the entire observable difference between a probe and a break on the day it happens.
2. Now find both cases in real prices. In the year of the same fund below, find one stretch where it left a multi-week range and was back inside within three sessions, and one where it left a range and stayed out for a month. Compare the extreme, the close, and the session's volume against the 50-session average drawn over the volume bars. Are they clearly different, or uncomfortably similar?

3. Now the twist: write down a precise rule that would have classified the first as a failure and the second as a success — using only information available on the break day. If you can't, you've just learned the most important thing in this lesson.