Inverse head and shoulder pattern (inverse head and shoulders): how to spot and trade it

The inverse head and shoulders pattern is a bullish reversal formation: three lows, the middle one (the head) being the deepest, connected by a neckline that price has to break to the upside. Every trader thinks they understand it, right up until they jump in too early on a fake breakout and watch their stop get tagged ten minutes later.

• Three-low structure: left shoulder, head (the deepest low), right shoulder, all connected by a neckline.
• Volume is a reading, not a guarantee: StockCharts requires a volume expansion on the breakout, and Bulkowski finds it highest on the left shoulder or head, lighter on the right shoulder, trending down across the pattern 65% of the time.
• The target is measurable: vertical distance between head and neckline, projected upward from the breakout point.
• The stop sits below the right shoulder, where the pattern is invalidated: below the head, it costs almost twice the distance for the same target.
• The numbers exist: an 11% failure rate and a 45% average rise on 3,197 bull-market cases per Bulkowski; the academic studies, on US stocks and currencies with daily data, say nothing about intraday or crypto.

Where this pattern comes from and why it never went away

Robert Edwards and John Magee described this pattern in Technical Analysis of Stock Trends, a book first written in 1948 according to the Routledge page for its eleventh edition, and of which the Internet Archive holds a 1958 edition. No computers, no candlestick software back then. Just traders plotting prices by hand on graph paper, hunting for recurring visual regularities in market behavior. The head and shoulders formation, in its bearish version and its inverse cousin, is still in today's manuals, and it is one of the chart patterns academic research has actually tested, as we'll see below.

If it's still taught, it's because it tells a coherent market story, not just an arbitrary shape. A downtrend exhausts its sellers in waves. A first selling wave creates a low, the left shoulder. A partial bounce follows. Then a second, more violent wave pushes price even lower, forming the head. Finally a third selling wave, weaker than the previous two, fails to push price as low again, the right shoulder. That last part matters most: the failure of the third wave to make a new low is the tell that selling pressure is running out of gas.

There's nothing mystical about it, it's crowd psychology you can literally read off a chart. Each low is a skirmish between buyers and sellers, and the fact that the right shoulder doesn't reach the depth of the head shows buyers are slowly taking back control. That underlying logic is exactly why this pattern is still taught more than seventy-five years after it was first described.

Anatomy of the pattern: left shoulder, head, right shoulder, neckline

Let's break down the structure, because this is where reading errors start, with traders spotting inverse head and shoulders patterns everywhere on their charts when half of them don't qualify. StockCharts' ChartSchool, on its Head and Shoulders Bottom page, describes the sequence this way. The left shoulder forms at the tail end of a downtrend: price falls, marks a new reaction low, then bounces. That bounce creates an intermediate peak. Price then falls again and exceeds the previous low, that's the head, the lowest point of the whole structure. A new bounce follows, creating a second peak.

Then comes the third leg down, the one that shapes the right shoulder. And here's the detail that makes or breaks the whole pattern: this low is always higher than the head, and usually in line with the low of the left shoulder, as StockCharts puts it. Bulkowski asks for the two shoulders to bottom near the same price, at a similar distance from the head, and finds that a left shoulder higher than the right one leads to worse performance after the breakout. Moderate asymmetry won't invalidate the pattern. A right shoulder that drops below the head, on the other hand, kills it completely. At that point you don't have an inverse head and shoulders, you have something else, maybe a range that's still trending down.

The neckline connects the two intermediate peaks (the one between left shoulder and head, and the one between head and right shoulder). It can be horizontal, upward-sloping or downward-sloping. On that slope, the two sources disagree: StockCharts reads an upward slope as more bullish, while Bulkowski's page on head-and-shoulders bottoms finds that patterns with down-sloping necklines perform marginally better. Don't build an entry on it. What matters is confirmation: for Bulkowski, the pattern is valid when price closes above a down-sloping neckline, or above the right armpit when the neckline slopes upward.

The role of volume: what separates a real signal from a mirage

Looking only at the shape of price and ignoring volume is a mistake on this particular pattern, because both sources make it part of the definition. StockCharts writes that volume plays a larger role in bottom formations than in top ones, that it should be light on the decline of the right shoulder, and that a valid breakout needs a volume expansion on the advance through the neckline. Bulkowski finds it highest on the left shoulder or head, diminished on the right shoulder, and trending downward across the pattern 65% of the time.

Picture two scenarios. In the first, price breaks the neckline on volume roughly in line with previous sessions, nothing special. In the second, the break comes with a volume spike clearly above the recent average. The manuals pick the second, and that's the rule to follow. But none of the sources read here quantify the gap in success rate between the two for this pattern, and across all patterns Bulkowski even finds that an upward breakout on volume above its 30-day average is followed more often by a return to the level, as explained in the guide on break and retest. Volume is a reading condition, not a measured guarantee.

Volume on the right shoulder deserves your attention for that reason. If it's clearly lighter than the head's volume, that's the profile both sources expect. If it's heavy on the right shoulder, close to what you saw on the head, it departs from the described profile and should put you on alert. You have to read it yourself off the volume histogram under your chart, and log it in your journal to check later what it changes on your own trades.

Spotting the pattern on a real chart: a step-by-step approach

In practice, here's the sequence I use to validate the pattern before I even think about an entry. First, confirm you're actually at the end of an established downtrend, not just inside a pullback within a bigger uptrend. Prior trend context isn't a minor detail, it's the very definition of the pattern: without a prior downtrend to reverse, StockCharts writes, there cannot be a head and shoulders bottom. And Bulkowski adds that price must have something to reverse: if the decline leading to the pattern is small, expect a small rise.

1. Spot a clear downtrend that's been running for a while, not just a handful of candles.
2. Identify the first low followed by a bounce: that's your candidate left shoulder.
3. Wait for a second, deeper low followed by a stronger bounce: that's your candidate head.
4. Draw a provisional neckline between the two intermediate peaks.
5. Watch for a third low forming that stays above the head's level: your candidate right shoulder.
6. Track volume behavior across all three phases, especially any gradual weakening.
7. Wait for the actual neckline break before acting, never before.

That last point is worth repeating because it's mistake number one for beginners: jumping the gun on the breakout. You see the right shoulder forming, you're convinced the pattern is about to complete, so you enter before price even touches the neckline. Except the market never signed a contract with you. The right shoulder can easily keep falling and invalidate the whole pattern right in front of you, with your capital already on the line.

Calculating the price target: the classic measured move

The theoretical target for an inverse head and shoulders pattern is easy to remember, and both sources give the same rule. You measure the vertical distance between the lowest point of the head and the neckline directly above it. That distance is then added to the price where the neckline breaks. Bulkowski adds a useful nuance: he first multiplies that height by the percentage of patterns that meet their target, 71% in his records, to get a target that is hit more often.

Take an illustrative example to fix the idea. Say the head bottoms out at 100, and the neckline, measured at that same point, sits at 120. The distance is 20 points. If the neckline break happens at 122, the theoretical target is 142 (122 plus 20), and the Bulkowski-style adjusted target is 136 (122 plus 20 times 0.71, rounded). That number isn't a guarantee, it's an estimate built from the pattern's geometry, more of a compass than a precise GPS.

Scaling out of the trade in stages rather than aiming for the full target in one shot is an option: a partial exit halfway there, at 132 in our example, a second chunk aimed at the full projection, and sometimes a third piece left running if the uptrend clearly extends beyond that. This approach is reasoned with the risk/reward ratio: you lock in part of the gain early while still leaving room for the best-case scenario to play out, and you then measure what it yields in R on your closed trades.

Where to place the stop-loss: the question that decides whether you survive this setup

Stop placement isn't a cosmetic detail, it's what determines whether the trade is even worth taking from a risk/reward standpoint. The simplest, and most logical, rule is to place the stop just below the low of the right shoulder. The reasoning is simple: if price falls back under that level after breaking the neckline, it means the pattern has failed, sellers are back in control, and your original thesis is dead. That is exactly the principle of the guide on where to place a stop loss: where the thesis is invalidated, not at a distance chosen for comfort. Back to the example: entry at 122, right shoulder at 110, stop at 109. The risk is 13 points for a 20-point target, a ratio of 1.5. A stop below the head, at 99, would push the risk to 23 points for the same target, and the ratio would fall to 0.9.

Some traders prefer placing their stop just below the neckline itself, closer to the entry price: at 119 in our example, the risk drops to 3 points and the ratio climbs to 6.7 on paper. The price you pay is the risk of getting stopped out by a simple retest of the neckline, and that return is not rare: Bulkowski records it after 65% of head-and-shoulders bottom breakouts. There's no universal answer here, it depends on your tolerance for market noise and the volatility of whatever you're trading. On a jumpy instrument, a tight stop under the neckline will shake you out before the move a good share of the time.

What matters, fundamentally, is that this stop gets defined before the entry, not adjusted afterward based on your mood. Position size should be derived from the distance between your entry price and that stop, never the other way around: with a $10,000 account and 1% risk per trade, so $100, a 13-point stop and a point worth $1, the size is 100 divided by 13, so 7 units rounding down, exactly as in the guide on how to calculate position size. That's a basic principle covered in our guide on fixed versus dynamic risk management, and it applies just as much here as on any other setup.

Inverse head and shoulders versus the classic bearish version

The regular head and shoulders pattern is the mirror image: it forms at the top of an uptrend and signals a bearish reversal, with the same three-part structure flipped upside down. A left shoulder, a head as the highest point, a right shoulder, and a neckline that gets broken to the downside. Conceptually they're twins, flipped upside down, and the rules above apply to both in mirror image.

What the two versions share: the same measure rule, a confirmation on a close beyond the neckline, and a stop beyond the shoulder that would invalidate the pattern. What separates them in the sources read here: StockCharts writes that volume plays a larger role in bottom formations than in top ones, and the Bulkowski figures quoted in this article are for the bottom only, so a number read for one version should never be quoted for the other. If you trade both, log them under two different tags.

The practical rule is simple: inverse means lows, not highs, and a bullish reversal, not a bearish one. If you see three highs above a neckline, you're looking at the classic pattern, whatever the neckline looks like, and the breakout to watch is to the downside.

Common mistakes and false signals (the fakeout problem)

The single biggest mistake, already mentioned, is entering before the neckline break confirms. But there are others just as damaging. One classic error: forcing the pattern onto a chart where it simply isn't there. That isn't a personal flaw: the study by Lo, Mamaysky and Wang published in 2000 in the Journal of Finance opens on that very point, the presence of geometric shapes in price charts is often in the eyes of the beholder, which is why its authors handed pattern recognition to an algorithm. You want to see a bottom so badly that you connect three random lows and call it an inverse head and shoulders, even though the right shoulder is way lower than it should be, or the whole thing formed inside a sideways range with no real preceding downtrend.

Fakeouts are the other recurring trap. Price breaks the neckline, you enter, and within a few candles it reverses back below the line, stopping you out. Keep it apart from the simple throwback to the neckline, which Bulkowski counts after 65% of breakouts of this pattern and which, he notes, hurts the performance that follows: price can come back to touch the line and leave again, or come back and never leave. The difference is judged on a close below the line, not on a wick. One way to reduce exposure is waiting for a confirmed close above the neckline rather than reacting to an intraday wick, and checking that volume actually backs up the move instead of assuming it does.

There's also a subtler mistake: treating the price target as a guarantee rather than an estimate. Traders get emotionally attached to that number, refuse to take partial profits earlier, and end up giving back gains when price stalls well short of the projection. If you've ever felt that specific frustration of watching an open profit evaporate because you were fixated on a textbook target, you already know how this ends. Keeping a written record of these situations in a trading journal is the fastest way to notice you're repeating the same error.

The market context that makes this pattern reliable

Context is not optional here, but it doesn't work the way intuition says. Intuition says a long, extended downtrend gives the pattern more weight; Bulkowski's records say otherwise: a short-term drop of zero to three months leading to the pattern gives the best post-breakout performance, and a pattern that follows a small decline can only reverse a small distance. The signs of seller exhaustion, volume highest on the left shoulder or head then diminished on the right shoulder, are part of the described profile. What isn't: a pattern forming right under nearby overhead resistance, one of the failure cases Bulkowski shows on his page.

Broader market conditions matter too, and here the honest thing to say is that the numbers quoted in this article are bull-market numbers: Bulkowski's table is explicitly labeled that way. Nothing read here tells you how the pattern performs on an individual stock during a broad market downturn. This is one of the reasons experienced traders check the bigger picture, the index, the sector, sometimes the correlated asset, before trusting a single-instrument reversal signal in isolation, and then measure the result in their own log.

Time frame is the other honest limit: the academic studies quoted here all work on daily data, and Bulkowski's page doesn't state the timeframe of his records. None of the sources read measures the pattern on a five-minute chart. Saying it's 'more reliable' on a daily chart than intraday is a reasonable intuition, not a result: nobody has published the intraday number, and your journal is the only place it will ever exist.

Backtesting the pattern: what the data actually suggests

Chart patterns have a reputation problem: everyone quotes success rates with total confidence, and almost nobody shows their methodology. Here is what the sources actually say. Bulkowski, on his page devoted to head-and-shoulders bottoms, gives for bull markets an 11% break-even failure rate, a 45% average rise, a throwback 65% of the time and a price target met 71% of the time, on 3,197 perfect trades, ranking the pattern 13th out of 39. His page states neither the market, nor the period, nor the timeframe of those records, and a 'perfect trade' measures the pattern, not a trader's result with fees and stops. Read those numbers as an order of magnitude, not a promise.

Academic research has tested the pattern on two markets. On currencies, Staff Report No. 4 of the Federal Reserve Bank of New York, by Osler and Chang in 1995, applies a head-and-shoulders trading rule to daily exchange rates of major currencies against the dollar from March 1973 to June 1994, with a recognition algorithm and a comparison against 10,000 simulated series; the version published in 1999 in the Economic Journal concludes that the rule is profitable but not efficient, since it is dominated by simpler trading rules. On US stocks, Savin, Weller and Zvingelis, in the Journal of Financial Econometrics in 2007, reuse the Lo, Mamaysky and Wang algorithm on S&P 500 and Russell 2000 stocks from 1990 to 1999: little or no support for a stand-alone trading strategy, but strong evidence that the pattern had power to predict excess returns, 5 to 7% a year risk-adjusted. And Lo, Mamaysky and Wang themselves, on US stocks from 1962 to 1996, conclude that several patterns, head-and-shoulders included, provide incremental information, without measuring the profitability of a trading rule. The number that matters still comes from your own log, the same way it does for the trading patterns in your own behavior.

What these studies don't say matters as much. None covers intraday: all work on daily data. None covers crypto assets, which didn't exist over their periods. The academic papers speak of 'the head-and-shoulders' without publishing, in their abstracts, a success rate specific to the inverse version; only Bulkowski quantifies the bottom on its own. And Bulkowski's figures are bull-market figures. If you want a real answer for your own trading, the only reliable path is running your own review: pull up historical charts, mark every instance where you'd have honestly called the pattern in real time (not with hindsight bias), and track what happened after the neckline break.

How Tradoshi helps you

Trading a pattern like the inverse head and shoulders isn't about spotting the shape once and getting lucky, it's about doing it consistently enough, with consistent risk, that the edge actually shows up in your results over time. Tradoshi's trading journal lets you import trades automatically from your broker (MT5, MT4, cTrader) or crypto exchange, or log them manually, so every neckline breakout you take gets recorded with the context you noted at the time rather than reconstructed from memory later.

Once you've got enough of these setups logged, the statistics module gives you win rate, profit factor, expectancy, drawdown, R-multiple and average win/loss ratio, plus an overall Oshi Score, so you can see whether your pattern-based entries are actually adding value or just feeling good in the moment. The position size calculator and customizable risk rules help make sure your stop placement below the right shoulder translates into a consistent, pre-defined risk per trade rather than a number you improvise.

Trade Review lets you replay a specific inverse head and shoulders trade, attach your own free labels (say, 'clean volume confirmation' or 'weak right shoulder'), note your emotional state through the check-in feature, and check plan adherence afterward. Over dozens of these reviews, patterns in your own decision-making start to surface, the kind you'd never catch from memory alone, and that's exactly where the trading playbook approach becomes genuinely useful.

Frequently asked questions