A head and shoulders pattern is three touches instead of two. That’s the whole difference between this and a double top. Price makes a high, pulls back. Makes a higher high, pulls back again. Makes a third high that fails to reach the second one, pulls back a third time. Three peaks, the middle one taller than the other two — a shoulder, a head, a shoulder. Connect the two pullback lows and you have a neckline. That’s the entire pattern before anyone attaches a story to it.
I don’t treat it as more special than a double top just because it has a third touch and a nicer name. What the third touch actually buys you is more information. A double top tells you price failed twice at a price. A head and shoulders tells you price failed at a price, failed again higher, and then failed a third time lower than the second attempt. That’s a shrinking effort curve — buyers pushed hard, pushed harder, then couldn’t even match their second try. The shape is describing weakening demand, not predicting the future.
Why the middle touch matters more than the outline
Most explanations of this pattern spend their words on the outline — head taller than shoulders, shoulders roughly even, neckline connecting the lows. That’s shape-matching, and shape-matching is where the pattern gets oversold as something mechanical. What I actually watch is the character of the head itself, the tallest peak in the middle.
A head that forms on strong volume-adjacent price action — wide-range candles pushing decisively into new highs — followed by a fast, sharp rejection tells a different story than a head that grinds slowly to a marginal new high and drifts back down over a week. The first version is a failed breakout attempt: buyers tried to run it and got overwhelmed. The second version is closer to exhaustion — there was barely any push left to begin with. Both can end up looking identical on a weekly chart zoomed out. They are not the same setup, and I size them differently.
The right shoulder is the second thing I check before the outline. If the right shoulder forms on noticeably lower momentum than the head — smaller candles, a slower approach, an earlier rollover before even reaching the head’s price — that’s confirmation the second failed push had less behind it than the first. If the right shoulder comes in just as strong as the head and only misses it by a hair, I’m less confident the pattern is really telling me demand is fading. It might just be a level that got tested three times and is still undecided.
The neckline break is the trigger, not the pattern itself
None of the above is tradeable on its own. A head and shoulders pattern isn’t complete until the neckline breaks, and until that happens all you’ve got is a shape that might resolve either way. I’ve watched plenty of textbook-looking head and shoulders setups get bought right back up through the right shoulder and never come near the neckline. The pattern existing doesn’t create the trade. The break does.
What I want on the neckline break is a close through it, not a wick through it. A candle that pokes below the neckline intraday and closes back above it hasn’t broken anything — that’s often a stop run, not a breakdown. A candle that closes clearly below the neckline, ideally with some room rather than sitting a penny under it, is the version I act on. I’d also rather see the break happen with some pace to it than a slow drift that takes a week to finally close under the line. A slow drift gives the level too much time to attract buyers defending it.
The retest matters too, and it’s the step most people skip. After the neckline breaks, price frequently comes back up to test the underside of that same line before continuing down. That retest, if it happens, is often the better entry than chasing the initial break — you get to see whether the old support is now acting as resistance, which is the confirmation that the level actually flipped.
The measured target, and why I don’t treat it as a promise
The standard measured target for a head and shoulders is the distance from the head’s peak down to the neckline, projected downward from the point where the neckline breaks. If the head is $12 above the neckline, the target sits roughly $12 below the break point. It’s a useful reference for where a reasonable move could reach, not a number the market owes you.
I use it to frame expectations and to decide where partial profit-taking makes sense, not as a certainty I’m holding for no matter what. Plenty of head and shoulders breakdowns travel a fraction of the measured move and then base out. Plenty travel well past it. The target is a ruler, not a floor plan.
A head and shoulders that worked: ROKU on the daily chart
I’ll walk through one so this isn’t abstract. ROKU had been climbing through the spring and put in a left shoulder at $94.20 before pulling back to $86.50. It then pushed to a head at $101.80 on a wide green candle with strong follow-through, before rolling over hard and dropping back to $85.90 — a lower pullback low than the first one, which is normal and not a red flag on its own.
The right shoulder formed next, and this is where the setup earned my attention. Price rallied back toward the head but stalled at $96.40, well short of $101.80, on noticeably smaller candles than the run into the head had shown. That’s a weaker second attempt at new highs — the same tell I look for on any reversal setup, just showing up as the third touch instead of the second. I drew the neckline connecting the two pullback lows, which sloped slightly, running from $86.50 down to roughly $85.10 under the right shoulder.
Eleven sessions later, price closed at $83.70, a clean close below the neckline with room to spare, on a wide red candle. I didn’t chase that candle. Three sessions after, price retested the neckline area at $85.60, got rejected, and closed back down at $84.20. That retest and rejection is what I acted on. I shorted at $84.00 with a stop at $87.10, just above the retest high. The measured target from a $15.60 head-to-neckline distance put a reasonable objective near $69.50. Price reached $71.40 over the following three weeks before I covered the bulk of the position, short of the exact target but well past what I needed for the trade to have worked.
If the retest hadn’t rejected — if price had closed back above $85.60 and held there — I’d have treated the neckline break as a failed one and stood aside, measured target or not.
How this is different from a double top
A double top is two touches at roughly the same price with a weaker second push. A head and shoulders is three touches where the middle one is the extreme, and the third one fails to reach the second. The information content is similar — both are asking whether the second (or third) attempt at a price came in weaker than the one before it — but the head and shoulders gives you an extra data point and, because of the extra swing, a built-in way to measure the target once the neckline breaks. A double top can be measured too, off the height of the level, but the head and shoulders’ shape makes that measurement feel more natural because the head is right there marking the top of the range.
I don’t rank one pattern above the other. Both are level-and-reaction analysis wearing different names. What matters in either case is the same short list: is the level real, is the latest push into it weaker than the one before, and did price actually close through the line that’s supposed to confirm the move, rather than just wick at it.
How I’d use this on your own chart
Find the three touches before you look for the name. A left shoulder, a head that’s the tallest of the three, a right shoulder that fails to reach the head. Draw the neckline off the two pullback lows between them. Then wait — don’t act on the outline alone. Watch how the right shoulder forms relative to the head, watch for a close through the neckline rather than a wick, and if you get a retest, watch whether the old neckline holds as new resistance before you commit size.
The measured target gives you a number to plan around, not a guarantee to hold for. I’ve had these fail to reach half the projected distance and I’ve had them run well past it. The pattern tells you demand weakened across three attempts and a level gave way. It doesn’t tell you exactly how far the reaction travels once it starts.
I mark levels like these — including the ones that happen to form a head and shoulders — on a blank chart in Static, the free daily chart room run by Draw Lines Make Money. If you want to watch how a neckline break and retest get read in real time, you can sit in:
Join the free Static chart room →
Disclosure: that’s an affiliate link — I may earn a commission if you join a paid tier later, at no extra cost to you. The free room is free.