Most guides to price action patterns are a glossary. They line up a dozen shapes, label each one bullish or bearish, and leave the reader believing every formation is worth the same attention. The market does not work that way. Some patterns reach their target more than 80% of the time once they complete, others barely clear a coin flip, and the gap between them decides whether a setup belongs in a trading plan or in the bin.
A price action pattern is a repeatable formation read directly off price, with no derived indicators sitting on top of the chart. The premise behind trading it is simple: price reflects everything known about a stock, formations tend to repeat, and what happened after a pattern in the past hints at what might happen the next time it appears. That premise is sound enough to build on. It is also where most traders stop thinking, which is exactly why they lose.
Continuation versus reversal
Every pattern below falls into one of two families. Continuation patterns mark a pause inside an existing trend, a stock catching its breath before the move resumes. Flags, pennants, triangles, channels, and rectangles all belong here. Reversal patterns mark a turn, a point where the prevailing trend runs out of buyers or sellers and hands control to the other side. Head and shoulders, double and triple tops and bottoms, and wedges live in this group.
The distinction matters because it tells a trader which way to lean before the pattern even completes. A continuation pattern in a strong uptrend has the trend behind it. A reversal pattern is betting against everything that came before, which is harder, and the reliability numbers reflect that difficulty in places worth paying attention to.
The patterns ranked by reliability
The figures below come from testing across 10 years of data and more than 200,000 individual patterns. A pattern counted only when it was complete, meaning a full break of the relevant support, resistance, or trendline. Success meant price went on to reach the target the pattern projected. Read as a ranking, the list says more than any single pattern description.
The most reliable: head and shoulders
Head and shoulders sits at the top. The standard version reaches its target about 83% of the time, and the inverted version, which signals a bullish reversal off a bottom, comes in slightly higher at roughly 83.4%. Three peaks with the middle one highest, a neckline connecting the lows, and a break below that neckline complete the pattern.
The discipline that separates a tradeable head and shoulders from a hopeful one is the neckline. The pattern is not confirmed when the right shoulder forms. It is confirmed when a strong candle closes through the neckline, ideally on rising volume. Calling the reversal early, while price is still building the right shoulder, is the most common way traders turn an 83% pattern into a losing trade.
Rectangles, triples, and doubles
A tight cluster of high performers follows. Rectangles reach their target around 78% to 79.5% of the time. Triple tops and bottoms land near 77.6% to 79.3%, and double tops and bottoms sit at roughly 75% to 78.6%.
These three are related in a way that carries a warning. A rectangle is essentially a double or triple top or bottom that failed to confirm. Price touched the same level two or three times, the swing point between the touches never broke, and the stock resumed its original trend instead of reversing. That relationship is the reason anticipating a double or triple top before the confirming break is dangerous. The reversal looks obvious right up until the moment price turns the formation into a rectangle and continues the way it was already going.
Channels and triangles
Channels and triangles cluster together around 72% to 73%. Ascending and descending channels run roughly 72.9% to 73%, and ascending and descending triangles sit a hair below at about 72.7% to 72.9%. Both are continuation patterns in most cases, a stock consolidating between converging or parallel lines before pushing on in the direction it was already traveling.
One measurement detail changes how these are traded. Most pattern targets are projected from the breakout point. Flags, pennants, and channels are the exception: their targets are measured from the outer edge of the pattern instead. Getting that wrong understates the move and leaves profit on the table.
Flags, and the pennant problem
Flags drop the reliability a noticeable step, to roughly 67%. That is still a usable continuation pattern, particularly after a sharp, near vertical run where a tight flag forms right on the edge of the move. The higher and tighter the flag, the better the odds it breaks in the direction of the trend.
Pennants are the cautionary tale. They get taught right beside flags in every textbook, they look almost identical, and they perform far worse. A pennant reaches its target only about 55% of the time. The small difference in shape, a converging triangle instead of a parallel channel, hides a real gap in odds, and a pennant breaks against the prior trend nearly as often as it breaks with it. A trader who treats the two as interchangeable is trading a near coin flip while believing the odds are in their favor.
Why the bullish versions edge out the bearish
Across nearly every pattern, the bullish version slightly outperforms its bearish mirror. The double bottom reaches its target 78.55% of the time against 75.01% for the double top. Triple bottoms beat triple tops, inverted head and shoulders beats the standard version, and bearish rectangles are the only notable place the pattern runs the other way. The edge is small but consistent, and it lines up with a market that, over long stretches, has drifted upward more often than down.
Reliability is not the same as profit
Here is the trap inside the ranking. A pattern that reaches its target 83% of the time tells a trader nothing about how much they keep when it works or lose when it does not. Hit rate and profitability are different measurements, and the second is the one that pays.
Consider a reversal setup. Entering early, before the new trend confirms, offers a tight stop and therefore small risk, but the probability of the swing actually developing is only around 40%. Waiting for a strong breakout into the new trend lifts that probability toward 60%, yet the stop is now far away and the reward relative to risk shrinks. Better odds get paid for with worse risk-reward. There is no setup that hands a trader high probability and a tight stop at the same time, because the institution taking the other side of the trade needs something in it too. This trade-off, mapped out in detail by Al Brooks, is the single most useful idea a pattern trader can internalize.
The same logic explains why roughly 60% of reversal trades end as small wins and losses that cancel each other out. The money is made on the minority of trades that run, which is why a defined target and a sensible reward-to-risk ratio matter more than chasing a high win rate.
A target and a stop for any pattern
The portable method that works across every pattern above has three steps. Measure the height of the pattern to set the target, projecting from the breakout point for most patterns and from the outer edge for flags, pennants, and channels. Place the stop where the pattern is clearly broken, not at an arbitrary dollar amount. Then check the math before entering, and skip the trade if the reward does not justify the risk.
A worked example makes it concrete. Suppose a stock forms a bull flag after a $5 run, and the flag itself spans about $1 of range. The target projects roughly $5 above the breakout, measured from the flag’s lower edge. A stop sits just below the flag’s support, say $2.50 of risk. That is a 1:2 risk-reward ratio, and it is a trade worth taking. Move the stop or the target so the ratio falls below 1:1, and the same pattern stops being worth the capital regardless of how reliable it is.
Confirmation, and the cost of jumping early
The identical pattern wins or loses based on when a trader acts. A signal is never final until the bar that creates it closes, because a formation that looks complete mid-bar can vanish before the close and leave an early entrant on the wrong side. A single setup on its own is rarely enough either. The bar to clear is strength in the direction of the intended trade plus at least two independent reasons to take it.
The retest is one of the cleanest confirmation tools. After price breaks above a resistance level, it often falls back to that level before continuing. If the old resistance now holds as support and price bounces, the breakout has been confirmed and the entry carries far less risk than chasing the initial break. The opposite habit, jumping on the first push through a level, is what gets traders caught. Waiting for a retest or for visible weakness in the prior trend, and timing the entry with the higher timeframe move rather than against it, turns a guess into a setup. Linda Raschke’s idea of waiting for the market to take out the low of a higher timeframe wick before committing belongs to this same patient school.
When a pattern fails, that is the trade
Failed patterns are not noise to be ignored. They are signals in their own right, and often better ones than the pattern that failed.
The mechanics are about other traders. When a breakout or a reversal fails, everyone who entered on it is suddenly in a losing position. They placed protective stops on the far side of the formation, and as price turns against them, those stops trigger one after another, each forced exit pushing price further the way it was already moving. The phrase for it is that the stops were run. The traders who entered the obvious pattern become fuel for the move against them. This trapped-traders framing was pioneered by Al Brooks, and most variations of it trace back to his work.
The two attempts rule sits underneath this. Markets tend to test a level twice before committing to a direction. A second push to a new low that fails and reverses is the structure of a double bottom: the bears who sold the breakdown are trapped, they cover, and their buying joins fresh buyers to drive a strong move up. A trader who sells the first failed breakout, or buys the second failed breakdown, is positioned to profit from other people’s stops rather than from a prediction about the stock.
Context decides everything
No pattern means anything in isolation. The first read of any chart is whether the stock is trending or stuck in a range, and the answer reshapes how every formation should be traded. The market spends a large majority of its time, by some estimates around 70%, going sideways rather than trending, and ranges are where pattern traders lose the most money because they expect trending payouts from a market that cannot deliver them. Inside genuine chop, where bars overlap heavily and neither side holds control, pattern reliability collapses and the only sensible play is to fade the edges of the range.
Volume is the next filter. A breakout on heavy volume is real; the same move on light volume is a warning. Rising price with strong volume confirms buying interest, falling price with strong volume confirms selling pressure, and a drop on light volume can mark a possible bottom rather than the start of a slide. Layer in confluence, where a pattern lines up with a prior support level, a round number, or a long-term moving average, and the odds improve again. The best setups have several of these factors stacking at once.
One modern wrinkle is worth keeping in mind. Most trading volume now comes from algorithms, which has two consequences. Psychological levels ending in .00 have lost some of the reliability they once had, and the cleaner and more textbook a pattern looks, the more machines recognize and act on it, which can make a well-formed pattern more significant rather than less.
Match the pattern to the time frame
Pattern size and time frame have to agree. Scalpers work on 1 to 15 minute charts, day traders on 5 minute to 1 hour charts, swing traders on the 1 hour, 4 hour, and daily, and position traders on daily, weekly, and monthly charts. The large reversal patterns, head and shoulders in particular, are most effective on the higher time frames, where each bar represents more conviction and the neckline break carries more weight. A head and shoulders on a 1 minute chart is a different animal from one on the daily, and treating them the same is a mistake.
The honest limitation
Price action patterns deserve one piece of plain honesty. The stories traders tell about why a pattern works, about bulls exhausting or sellers stepping in, are after-the-fact rationalizations, and there is no proof they are correct even when the trader using them makes money. Markets are anonymous, so no one can verify who was buying or selling at any given bar. The traders these patterns failed to save are invisible, which makes survivorship bias a constant risk in any claim about what works. Published data on retail accounts is sobering: regulatory disclosures put the share of losing accounts somewhere around 75%, well above the breezy success stories that fill trading forums.
None of that argues against trading patterns. It argues for trading them honestly. A pattern is not a prediction. It is a probability with a defined risk attached, and a trader who treats the most reliable formations as edges rather than certainties, sizes for the losses, and protects capital with a stop on every position is doing the only thing the data actually supports. The shape on the chart is the easy part. What separates traders is the reliability they respect, the confirmation they wait for, and the discipline to treat a failed pattern as the next opportunity rather than a personal insult.