The Head and Shoulders Pattern

A head and shoulders pattern is three peaks in a row where the middle one is highest, followed by a break below the lows between them. It's the most famous chart pattern there is, and most of the people who think they see one are looking at something else.

This article covers what the pattern actually is, how to draw it so you're not fooling yourself, what to do when a stock you own forms one, where it fails, and why I'd rather you use it as a seatbelt than as a trade.

What the pattern is

Three pieces, in order.

The left shoulder. The stock rallies to a peak and pulls back. Nothing unusual yet. Every uptrend has these.

The head. It rallies again, to a higher peak, and pulls back again, roughly to the same level as the first pullback. Still nothing unusual. This is what a healthy uptrend looks like.

The right shoulder. It rallies a third time and fails to get anywhere near the head. The peak lands about where the left shoulder did. Then it rolls over. That third attempt is the tell. Buyers had a chance to make a new high and couldn't.

Draw a line across the two pullback lows. That's the neckline. The pattern isn't complete until price closes below it. Until then you have three bumps on a chart and an opinion.

Why it works, when it works

Chart patterns are pictures of crowd behavior. A head and shoulders in 1929 looks like one in 2008 because the psychology underneath is the same. Buyers push to a peak and exhaust themselves. Sellers test lower. Buyers try once more and can't make a higher high. Sellers take control.

The right shoulder is where that story becomes visible. Everyone who bought near the head is now underwater and would love to get out even. Everyone who bought the first pullback is watching their gain shrink. The third rally runs into all of them selling, which is why it stalls below the head. When the neckline goes, the last group holding on, the ones who bought the dips and were told it always comes back, finally sells too.

The pattern is complete when price breaks below the neckline on volume. Not before. Three peaks without the break is a stock that hasn't decided yet.

How to draw it so you're not fooling yourself

Four rules I hold to. They disqualify most of the patterns people point at.

Use closes to validate the pattern, wicks to place the stop. A daily bar that spikes below the neckline and closes back above it is not a breakdown. It's a liquidity grab, and often it's the opposite of a breakdown. I validate patterns on closing prices and place stops using the highs and lows. Pick a convention and keep it for the whole trade; switching between them mid-analysis is how you talk yourself into whatever answer you already wanted.

The shoulders should be roughly level. Not identical. Roughly. If the right shoulder is a lot higher than the left, the sellers aren't as tired as the pattern needs them to be. If it's a lot lower, the stock may already be in a downtrend and you're late.

The neckline can slope, but not much. A gently rising or falling neckline is fine. A steep one means the two pullbacks landed in different places, which means the support level the pattern depends on isn't really a level.

Weeks and months, not hours. A head and shoulders that forms over three months on a daily chart represents three months of buyers and sellers fighting over a price. One that forms over three hours on a five-minute chart represents noise. The longer the pattern took to build, the more it means when it breaks.

The volume rule

Volume is what separates a real breakdown from a false one, and it's easy to ignore because it's a second chart underneath the one you're looking at.

The textbook version has volume heaviest on the left shoulder and the head, lighter on the right shoulder, then expanding on the break. The part I treat as non-negotiable is the last one. A close below the neckline on ordinary volume is a pattern that hasn't confirmed. A close below it on volume well above the stock's recent average is sellers showing up in size, and that's the signal.

Where the stop goes, and what a failure looks like

Patterns are useful mainly because they tell you where you're wrong. For a head and shoulders, that's simple: price recovering above the neckline after the breakdown means the pattern failed. That's your stop. Specific, defensible, and based on the structure of the pattern rather than on how much pain you're willing to tolerate.

A failed breakdown is information too. It means sellers had their chance and buyers absorbed it, which is sometimes the start of a strong move up. But it doesn't hand you a new bullish trade automatically. It hands you a reason to look again.

On the ABC Company chart above, the stock closed below $48 on heavy volume. If it then closed back above $48, the setup is finished. You don't argue with it, you don't give it another day, you don't decide the neckline was really $47. You're out.

The measured move: how far it's supposed to fall

Measure the height of the pattern from the top of the head to the neckline. Subtract that from the neckline. That's the target.

Two things to say about that number. It's a projection, not a promise, and plenty of breakdowns run out of sellers before they get there. It's also better than guessing, and it gives you a reward-to-risk figure you can check before you act: $10 of expected move against however far above the neckline your stop sits. Book Four's rule is that a setup needs to offer at least 3:1 to be worth taking. If your stop has to be $4 above the neckline because the stock is volatile, a $10 target doesn't clear the bar and the pattern, however pretty, isn't a trade.

What to do when a stock you own forms one

Here is where I'll part company with most of what's written about this pattern. Nearly all of it is written for short sellers. Most readers of this series don't sell stocks short, and I don't think they should start. For you, the head and shoulders is a defensive tool, and it's a good one.

If you own ABC and it has built a left shoulder, a head, and a right shoulder, you don't need to do anything yet. The pattern isn't complete. You should, however, know exactly where the neckline is and have decided in advance what you'll do if it breaks on volume.

When it breaks, the honest answer is that the market has just told you the uptrend you were riding is over, at least for now. Whether you sell depends on why you own it. If you bought it as a trade off the chart, you sell, because the thing you bought it for is gone. If you bought it because the business is growing and the price was reasonable, go back to the fundamentals and ask whether anything changed. Sometimes a head and shoulders forms because earnings estimates are quietly being cut and the chart knew before you did. Sometimes it forms because a large holder was selling for reasons that have nothing to do with the company. The chart can't tell you which. The 10-K can.

What you should not do is decide the pattern doesn't apply to this stock because you like the company. The pattern isn't a judgment on the company. It's a count of buyers and sellers, and the sellers just won.

The inverse head and shoulders

Flip everything over. Three troughs, the middle one lowest, with a neckline drawn across the two rally highs between them. A close above the neckline on volume completes it and signals a bottom. The measured move works the same way: head to neckline, added to the neckline.

For a long-only investor this is the more useful version, because it's an entry rather than an exit. It's also the one to be more skeptical of, since bottoms take longer to build than tops and the third trough often isn't the last. Wait for the break. The stock that's "about to break out" has cost more people more money than the stock that already did.

How reliable is it, really

Among the more reliable patterns when it's properly formed, which is the caveat doing all the work in that sentence. Tom Bulkowski has done the most thorough statistical work on chart patterns that I know of, and his results for head and shoulders tops are favorable. What's worth noticing is that he doesn't report one success number. He separates the failure rate, the average decline, and how often price reaches the measured target, because those are three different questions and the answers aren't the same. If you want the figures, read them in the original rather than in a summary, mine included.

Every pattern fails some percentage of the time. The discipline isn't finding one that doesn't. It's sizing the position so that losing on the failures still leaves you ahead when the wins outpay them, and the only way to know whether that's true for you is a journal with enough entries in it.

The three mistakes I see most

Seeing it everywhere. Once you learn the shape you'll find it on every chart, because any three bumps can be made to fit if you squint. Hold to the drawing rules above and most of them disappear.

Acting before the break. The right shoulder looks obvious in hindsight. In real time it's a pullback in an uptrend until proven otherwise, and plenty of them go on to make new highs. The pattern pays you for patience and charges you for anticipation.

Ignoring volume. A neckline break on thin volume is a coin flip dressed up as a signal. If you only take one rule from this page, take that one.

What to do with this

Pull up a daily chart of any stock you own over the past year. Look for three peaks with the middle one highest. If you find one, draw the neckline across the two lows and note whether price ever closed below it, and on what volume. Then check what happened in the following month.

Do that for ten stocks and you'll have learned more about this pattern than any article can teach you, including how often the thing that looked like a head and shoulders was just a stock going sideways. That's the point of the exercise. The pattern is real. Your eye isn't trained yet, and the only way to train it is on charts where you already know the ending.