Most reversal patterns are about a moment. A long wick records one session where sellers pushed and failed; an engulfing candle records one session that undid the previous one. The rounding bottom is the opposite kind of evidence. Nothing happens on any single day. Selling simply gets a little less effective each week until, months later, it has stopped being effective at all, and the chart has drawn a bowl without ever drawing a turn.
A rounding bottom is a chart pattern in which price forms a gently curved low over many sessions, usually months, with no sharp reversal candle: a decline that flattens, a low made gradually, then a slow sequence of higher lows back to the level where the move began. It is confirmed by a close above that rim and invalidated by a close below the low of the curve.
Because it forms so slowly, it is the pattern people most often miss on the daily chart and most often imagine on the weekly. Here is what it actually requires, why no candlestick marks its low, and where it misleads.
What the formation looks like, in checkable terms
Bulkowski’s identification guideline is a rounded bowl shape, usually over many months, and he notes that rounding bottoms are difficult to spot unless you look at the weekly scale, where the curve is more apparent. Broken into pieces you can check, the reversal version has four parts:
- A prior decline that brings price to the area of the low. Without it, the curve is a pause, not a turn.
- A low made over many sessions, with no single candle where one side clearly lost. Price drifts, tests slightly lower, drifts again.
- Higher lows that form gradually. Each shallow pullback ends a little above the last one. This is the higher highs and higher lows structure in slow motion: the sequence of higher lows is the pattern’s whole content, and it is only visible once you step back far enough to see several of them.
- A return to the rim, the price where the decline into the bowl started, which is the level the pattern has to get through.
An illustrative timeline, with illustrative prices: a stock declines from 60 to a low region around 40 over three months. It spends the next two months between 39 and 42 with no candle that stands out. Over the following three months the lows step up: 41, 42.5, 44, 46, 49. Eight months after the decline began, price is back at 58 to 60, the left rim, and the bowl is complete. The whole thing is one curve on a weekly chart and a featureless mess on a 15-minute one.
Two features Bulkowski describes are worth knowing because they look like failures and usually are not. Price may shoot up midway through the turn and then retrace most, though usually not all, of that jump back to where it started; he calls this the bump. And at the right rim, price often pauses at the level of the left lip, sometimes forming a small handle before continuing.
Why there is no reversal candle at the low
A single-candle reversal pattern records a session where one side was tested and lost decisively, which is why the invalidation for those patterns is a specific candle’s low. A rounding bottom has no such session. Supply is not defeated in a day; it is absorbed over weeks, by buyers who are patient rather than urgent, against sellers who are tiring rather than capitulating. The chart records that as a curve precisely because there was no moment of decision.
This is why single-candle tools are the wrong instrument here. Searching the low of a rounding bottom for a hammer or an engulfing candle will usually find something, because those shapes appear everywhere, and whatever you find will carry no more information than any other candle in the bowl. The read comes from the sequence of higher lows and from the rim, not from any one bar. If the four prices in a candle are still new to you, how to read a stock chart covers what they record and why the weekly timeframe changes the picture.
Confirmation and invalidation
- Needs: a prior decline into the bowl, and a low that took many sessions to form.
- Confirmed by: a close above the rim. Bulkowski uses a close above the left peak, because price on the right side of the bowl may not pause at a minor high to give you a cleaner level.
- Invalidated by: a close below the low of the curve. That is a new lower low, and the slow sequence of higher lows the pattern was made of is over.
The rim is also a resistance level in the ordinary sense: it is where sellers last overwhelmed buyers, and the pattern is not complete until price has proven it can trade through it. How to draw that level from the chart rather than from wishful thinking is the subject of the support and resistance post.
The traditional measured-move convention, described rather than forecast: take the height from the left lip to the lowest point of the bowl, and project some fraction of it above the rim. Bulkowski’s version multiplies that height by the percentage of his patterns that met the target, 65% in his sample of 990 rounding bottoms (bull-market results, statistics dated August 2020), and adds it to the right lip. His other bull-market figures from the same sample are a break-even failure rate of 4%, an average rise of 48%, and a throwback rate of 64%, meaning that in roughly two thirds of cases price came back to the breakout level within 30 days after breaking out, and he notes that throwbacks hurt performance. His glossary defines average rise as measured from the open the day after the breakout to the highest peak before a 20% decline, which is a generous measure, and none of these figures transfer to a different market or timeframe.
The volume tradition, and what it is
The textbook description gives the rounding bottom a volume shape to match the price shape: heavier on the way in, lightest at the bottom of the bowl, heavier again as price climbs the right side. The cup-and-handle literature, the popular variation of this pattern defined by William O’Neil, states the same convention, with volume expected to fall through the cup and rise sharply as price leaves it.
Treat that as a tradition rather than a tested rule. Bulkowski’s rounding bottom page, which is the most detailed public measurement of the pattern, carries no volume statistic at all. Curved volume is nice to see and is consistent with the mechanism, patient buying absorbing tiring selling, but its absence does not break the pattern and its presence does not confirm it. Price through the rim is the confirmation.
Where it misleads
A range with a curved floor. On a low-volume instrument, several months of sideways drift will often look bowl-shaped if you squint, especially on a compressed chart. The test is the sequence of higher lows: if the lows are not stepping up, it is a range, and the triple bottom post covers what a flat floor of repeated lows means instead. If the turn happens abruptly rather than gradually, with a cluster of prices stranded between two gaps, that is a different structure again, covered in the island reversal pattern.
A bowl so long the prior trend no longer matters. If the decline into the bowl happened two years ago and the curve has taken a year, the “reversal” is not reversing anything that current participants remember. It is a base, which is a different and usually weaker claim.
The pattern that is actually a continuation. This is the one the textbooks skip. Bulkowski reports that price trended upward into the pattern in 67% of the rounding bottoms in his sample, meaning most of them acted as continuation patterns, not reversals. The classic reversal case, a bowl at the end of a decline, is the minority in his data. That does not make the reversal read wrong; it means “rounding bottom” on its own tells you nothing about what came before, and what came before is the part you have to check.
The cup and handle. Bulkowski describes the cup with handle as a popular variation of the rounding bottom, with a small pullback at the right rim before the breakout; it is its own pattern with its own rules, and it deserves its own page rather than a paragraph here.
The rounding top, briefly
The mirror image is a rounding top: a gentle half-moon curve after an upward trend, with the two rims bottoming near the same price. Bulkowski’s measured behaviour is a useful corrective to the name. In his sample of more than 950 rounding tops (bull-market results, statistics dated August 2020), upward breakouts, a close above the pattern’s highest high, ranked 2nd of 39 patterns with an average rise of 55% and a 9% break-even failure rate, while downward breakouts, a close below the lower of the two rims, ranked 3rd of 36 with an average decline of 17% and a 20% failure rate. In other words, a rounding top in his data broke upward well and downward modestly, which is not what the word “top” leads people to expect.
The same logic applies in mirror. It needs an uptrend before it. The bearish read is confirmed by a close below the lower rim and invalidated by a close above the highest high of the curve. Until one of those has printed, it is a curve, and a curve is not a direction.
What this post is not
It is not a way to spot bottoms early. A rounding bottom is only identifiable once most of it has formed, and it is only confirmed once price has cleared a level that was months in the making. The pattern’s value is the opposite of speed: it forces you to name a rim and a low before the read means anything, and the honest figures on it come with a sample and a period attached, or they are not figures.
For the candlestick patterns that do describe a single decisive session, each drawn with its context requirement and invalidation, see the candlestick pattern library. For the mechanics underneath all of it, how to read a stock chart is the place to start.
Sources
Every figure above was checked against these before publishing.