Does the head and shoulders pattern always work?

No. Across 44 real patterns detected on 1-hour candles of BTC, ETH and SOL since 2021, the price moved in the pattern's direction about half the time, very close to what chance would give.

Head and shoulders pattern: what it is, how to spot it and how often it fails

Q: Is the head and shoulders pattern bullish or bearish?

The classic version appears after a rise and the theory reads it as bearish; the inverse version appears after a fall and is read as bullish. Reading it that way does not guarantee what the price will do.

Q: Which timeframe should I look at?

It appears on any timeframe. The shorter the timeframe, the more patterns form and the more noise there is; the figures in this article come from 1-hour candles.

What it is

The head and shoulders is one of the best-known chart patterns. It is three rallies in a row: the middle one, the head, goes higher than the other two, the shoulders, which end up at a similar height. Underneath, the two pullbacks separating the shoulders from the head mark a line: the neckline.

Classic theory reads it as the end of a rise: the price tried three times, the third attempt fell short and, once it closes below the neckline, the move should turn down. This article explains that reading and then puts it to the test with real data.

How to spot it

By eye, almost any chart seems to have heads and shoulders. To count them without cheating, Kronochart uses fixed rules, the same in the app and in these figures:

Five alternating points: high (left shoulder), low, high (head), low and high (right shoulder). Each high or low has to be the extreme of the 5 candles before it and the 5 after it.
A clear head: at least 1.5% above the average of the two shoulders, with several candles above them, not a single wick.
Similar shoulders: the difference between them is less than 40% of the head's height above the neckline, and neither shoulder is more than twice as far from the head as the other.
An almost flat neckline: the two lows within 1.2% of each other, and the shoulders well clear of the neckline.
A sensible size: between 12 and 80 candles from shoulder to shoulder.
Breakout: the pattern only exists once a candle closes below the neckline. Before that it is a possibility, not a pattern.

One detail changes the figures a lot: the detection never looks ahead. Each point only counts once the candles needed to confirm it have passed, just as it would for anyone looking at the chart at that moment. Patterns drawn with hindsight look neater and seem to work more often.

The inverse head and shoulders

It is the same pattern turned upside down, after a fall: three lows, the middle one deepest, and a neckline above joining the two highs in between. Theory reads it as the end of the fall, and the breakout is a close above the neckline. It is spotted with the same rules, mirrored.

What classic theory says

Entry: after the close below the neckline (above it, in the inverse pattern), or on a pullback, when the price comes back to touch the neckline from the other side.
Textbook target: the height of the head above the neckline, projected from the breakout point.
Invalidation: if the price climbs back above the right shoulder, the pattern no longer makes sense. Many traders place their stop-loss there.
Volume: theory adds that volume tends to fall on the right shoulder and rise on the breakout. Our rules leave it out, so as not to mix two ideas in a single figure.

What happened in real data

We searched for every pattern that meets the rules on 1-hour candles of BTC, ETH and SOL from January 1, 2021 to October 17, 2026: about 152,000 hours of market. 44 turned up: 21 classic and 23 inverse. That strict rules find so few already says something: the perfect pattern is rare.

To see what happened next, we use the same measure as everywhere in Kronochart. From the close of the breakout candle, two bands are drawn, above and below, at 1.5 times the average hourly volatility (14-candle ATR), and we check which one the price touches first over the next 48 hours. The expected way means touching first the band the theory points to: the lower one for the classic pattern and the upper one for the inverse.

Binance 1-hour candles. Figures calculated on October 18, 2026.
Pattern	Cases	Expected way	Chance
Head and shoulders	21	42.9%	50.4%
Inverse	23	60.9%	49.6%
Both	44	52.3%	50.0%

The chance column is not an assumed 50%: it is what happened at more than 30,000 ordinary moments in the same data, measured the same way. Someone who bought or sold without looking at the chart would have been right that often.

How to read it: in this data, the pattern went the expected way 52.3% of the time, against 50.0% for chance. With only 44 cases, the margin of error is around ±15 points: the gap with chance is far too small to say the pattern, on its own, helps anticipate the price. The differences between classic and inverse, or between coins (BTC 8 cases, 62.5%; ETH 16, 43.8%; SOL 20, 55.0%), sit within that same noise.

The textbook target, the height of the head, was reached within 48 hours in 52.3% of cases.

This does not mean the pattern is useless: it helps structure your reading, decide where an idea is invalidated and place the stop. What the data says is that, on its own, it does not tell you where the price will go. The figures are recalculated whenever the data is extended.

One case that worked and one that failed

BTC/USDT, March 3, 2022. A clean pattern, a close below the neckline and a drop to the lower band in 9 hours; it also reached the textbook target.

ETH/USDT, June 26, 2025. Same rules, an equally clear breakout, and the price rose to the upper band in 7 hours. Knowing this happens almost as often as the other outcome is part of becoming a seasoned trader.

Test yourself with 3 real cases

Three patterns picked at random from all that turned up, in their real proportion of hits and misses. You will see the chart up to the breakout candle, with no drawings and no date, as in the app. Decide, then see what happened: repeating this is how you build fluency.

Preparing the cases…

Historical data. Education, not financial advice. Being right here guarantees nothing in the real market.

Common mistakes

Seeing it before it exists. Entering on the right shoulder, without a close below the neckline, is jumping ahead of a pattern that has not formed yet. Many never break.
Forcing the shoulders. If you have to hunt for it to fit, it probably is not there.
Ignoring the timeframe. A 1-hour pattern inside a strong daily trend carries little weight.
Overconfidence. Being famous does not make it reliable: in the data it goes the expected way roughly half the time.
Moving the stop when the price climbs back above the neckline, hoping the pattern will “be right”.

Managing risk with this pattern

Since the pattern alone does not decide the direction, what matters most is how much you risk on each attempt:

Decide where it is invalidated before entering: above the right shoulder (below it, in the inverse pattern).
Size the position from that distance, so a miss costs a small, fixed share of the account.
With leverage, check that the liquidation price sits well beyond your stop; otherwise liquidation comes first.
Compare the distance to the stop with the distance to the target. If you risk more than you expect to make, you would need to be right far more than half the time.

Check what you have learned

Five questions about this article. Each answer is marked straight away, with the reason why.

1. When does a head and shoulders really exist? When the right shoulder forms When a candle closes on the other side of the neckline When the head rises above both shoulders

2. In the inverse pattern, what does theory expect after the breakout? A rise A fall A sideways range

3. In the classic pattern, where is the idea invalidated? If the price climbs back above the right shoulder If the price touches the neckline If the price drops 1%

4. In BTC, ETH and SOL since 2021, how often did the pattern go the expected way? Almost always, around 80% About half the time, very close to chance Hardly ever, around 20%

5. If on each attempt you risk twice what you expect to make, what do you need to break even? To be right more than 2 times out of 3 To be right half the time It does not matter if the pattern is clear

Frequently asked questions

Is the head and shoulders bullish or bearish?

The classic version appears after a rise and theory reads it as bearish; the inverse version appears after a fall and is read as bullish. Reading it that way does not guarantee what the price will do.

Does it always work?

No. Across 44 real patterns in BTC, ETH and SOL since 2021, it went the expected way 52.3% of the time, against 50.0% for chance measured the same way.

What is the neckline?

The line joining the two lows between the shoulders and the head (the two highs in the inverse pattern). The pattern is complete when a candle closes on the other side.

Which timeframe should I look at?

Any. The shorter it is, the more patterns form and the more noise there is. The figures in this article come from 1-hour candles.

What is a pullback to the neckline?

The price returning to the neckline after breaking it. Sometimes it touches it and carries on; other times it crosses back and the pattern is cancelled.

Sharpen your reading of this pattern

On Kronochart you can revisit it in sessions of 10 situations of the same type, on your own chart setup, polish what goes wrong, reinforce what goes right and see over the months whether your reading beats chance.

Train with real market moments