Three lows at about the same price, two bounces between them, and a line over the bounces. That drawing is on every education page, and it is why the triple bottom is so over-called: the shape is easy to see, and most charts labelled with it fail one of three checks the drawing never mentions. This post is the mirror of the triple top, with the same rules turned upside down and one extra job: giving you those three checks.
A triple bottom does not exist until price closes above the neckline. Before that close, what you are looking at is a trading range that has been tested from above three times, and ranges resolve in both directions.
What three lows at one price actually record
Price falls to a level and stops. It bounces, falls again on whatever the market now knows, and stops at the same place. Then it does it a third time. Each of those three declines is a separate group of sellers deciding the price is worth taking, and each time enough demand arrives at the same price to absorb them.
That is the whole mechanism: one price where buyers have met every attempt so far. The pattern is not a prediction that they will keep doing it. It is a record that they have done it three times, which is more evidence than twice and less than proof. The level itself is ordinary support, drawn the same way as any other.
The third hold carries more weight than the second only if the selling into it was weaker. If the third decline is slower, bottoms on smaller candles, or gets there on shrinking volume, sellers are running out of conviction at exactly the level where they most need it. Bulkowski’s identification notes record the typical shape: volume is usually higher on the first bottom than on the last, trending downward 61% of the time. A third decline that arrives faster and heavier than the first two is a different situation.
The formation rules, stated so you can check them
Most descriptions say “three lows at roughly the same level” and leave “roughly” to you. Here are the rules with the ambiguity taken out.
1. There was a downtrend to reverse. Bulkowski lists the price trend as downward leading to the pattern, with one qualifier worth keeping: the decline should not drop below the first bottom. A reversal pattern needs something to reverse, so a stock that has drifted sideways for months and touched the floor of its range three times is a range, not a triple bottom, whatever the shape looks like.
2. Three lows inside a tolerance band. First measure the pattern’s height: the highest of the two intervening peaks to the lowest of the three lows. Then measure the spread between the highest and lowest of the lows. A working rule you can apply: the spread must be small next to the height, something like a fifth of it or less. That fraction is a house rule rather than a published standard, since the published description says only that the valleys bottom near the same price, and it is fine to pick your own. What is not fine is deciding after you have looked.
Illustrative prices to make it concrete: lows at 50.0, 50.4 and 49.8 with peaks near 56 give a height of about 6 and a spread of 0.6, well inside the band. Lows at 50, 48 and 46 with the same peaks are not one level; that chart is a decline with pauses in it.
3. Two intervening peaks. They do not need to match, but both should sit clearly above all three lows. If the second is far lower than the first, price is compressing toward the level, which is a different structure with a different meaning.
4. The neckline through the higher peak. Bulkowski’s confirmation rule uses the highest peak between the valleys, so the line goes there. Some traders draw it through both peaks when they are close; when they are not, the higher one is the honest choice, because it is the price the pattern must actually reach. A middle low marginally above the other two is within the definition, and the same source notes that a last valley sitting above the second one is associated with better performance in its sample.
A chart that passes all four is a candidate. It is not yet a triple bottom.
When does the range become a triple bottom?
On a close above the neckline, on the same timeframe as the chart you are reading. A wick above it is a price the market visited and did not hold; a close is a price it accepted. Bulkowski’s rule is the same: the pattern confirms once price closes above the highest peak between the valleys.
Until that close, the buyers at the lows have won three rounds and the sellers at the peaks have won two, and nothing on the chart says which group gives way first. The same page is blunt about how often the pattern never gets there: triples are rare, and price often continues down without confirming.
One more thing the source records about what happens after confirmation: a throwback rate of 65%. In that sample, close to two thirds of confirmed triple bottoms saw price return to the neckline before continuing. A retest of the line from above after the break is the normal case, not evidence the break was false.
What ends the read?
Two prices, and it is worth naming both before the chart moves.
Before confirmation: a close below the lowest of the three lows. The level that absorbed three declines has been taken, and the chart is now a range breakdown with three defined holds behind it. Whatever you had called it, it is no longer a triple bottom.
After confirmation: a close back below the neckline weakens the read, and a close below the lowest low finishes it. Given the throwback figure above, a touch of the neckline from above is expected; a close through it is not.
The measured move is a convention, not a forecast
The standard convention: take the height, from the highest intervening peak to the lowest low, and project it upward from the neckline. Bulkowski’s version of the rule multiplies that height by the historical percentage meeting the target before adding it to the breakout price, which builds the shortfall into the target from the start.
The numbers on that page. The results are labelled Important Bull Market Results and are based on more than 2,500 perfect trades; the page gives no date for when the statistics were last updated, so read them as an undated snapshot. Within that sample: overall performance rank 12 out of 39, a break even failure rate of 13% (the page’s term, defined in its glossary), an average rise of 46%, a throwback rate of 65%, and 74% meeting the price target. The page’s own summary calls the failure rate reasonable and the average rise decent.
Read those as a description of one dataset, not as odds on the chart in front of you. Every one of them counts only patterns that confirmed with an upward breakout, so a three-low range that broke down was never counted as a triple bottom at all. The numbers say nothing about your instrument or your timeframe, and this post is not a forecast.
The three checks that disqualify most candidates
Most charts labelled “triple bottom” fail one of these. Run them in order, because the first one removes the most.
1. There was no prior decline. A stock that has been moving sideways, tested the floor of its box three times and bounced is a box. There is no trend to reverse, and the same three touches look identical whether the box eventually breaks up or down. Bulkowski’s trading lessons open with exactly this case: price moves sideways in a series of peaks and valleys, a bullish pattern appears near the floor, the breakout runs to the top of the range and fails. He also warns off triple bottoms that appear after a long uptrend, where the “decline” into the pattern is a pullback in a rise. Rule 1 exists for both.
2. The lows drift lower each time. Three lows at 50, 48 and 46 are not a defended level; they are a descending structure, and each “hold” is lower than the last. That is what a downtrend looks like at close range, and calling it a bottom mistakes the pauses for the floor. A third low above the second is the version the source associates with better performance; a falling one is the opposite.
3. The third bottom is really a retest. This is the double bottom counted twice. Price makes two lows, closes above the neckline, then throws back to the line and holds near it. The throwback rate above says that retest is the normal sequence, and a reader who counts it as a third low has drawn a triple bottom whose neckline broke before the pattern was finished, which is not a thing. If a close above the neckline already happened, the chart is a confirmed double bottom in its throwback, and the confirmation, invalidation and measured move all belong to that pattern.
A triple bottom that passes all three is a double bottom whose neckline break did not come, where sellers stepped in at the peak and produced a third decline that held at the same place. The extra low adds one more hold to the evidence and moves the neckline to the higher peak. It does not change what the pattern means.
What the third low often looks like up close
Zoom into the third hold and the candles at the low often form one of three bullish patterns, because they describe the same event at a smaller scale.
The hammer is a single candle with a small body near the top of its range and a lower wick at least twice the body’s height, forming after a decline. Sellers drove price well below the open and buyers took all of it back before the close, which is what a third decline into demand looks like in one session. It needs the decline before it, and the third leg of a triple bottom is one; a close below its low invalidates it.
The tweezer bottom is two adjacent candles that reach almost exactly the same low after a decline: the level was tested twice in consecutive sessions and held both times. A close above the second candle’s high confirms it, and a close below the shared low ends the read. The dragonfly doji is the hammer with no real body, opening and closing at the high of a long lower wick, and it carries the same message with the same invalidation.
Any of these at the third low is a detail inside the hold, not a triple bottom on its own. It can show the third push lower failing a few sessions before the neckline is tested; it cannot show that the neckline will break. The full set, with the context each one needs, is in bullish candlestick patterns.
The habit this pattern teaches
The triple bottom teaches the same discipline as its mirror: a definition you can test, a moment it becomes real, and a price that proves it wrong. Three lows inside a band you fixed in advance, after a decline, a close above the higher peak to confirm, a close below the lowest low to invalidate. Anything short of that is a range tested three times, and a range is allowed to go either way.
If peaks, closes and necklines are still new vocabulary, how to read a stock chart covers the structure underneath every pattern on this site. The candlestick pattern library has the hammer, the tweezer bottom and the dragonfly doji drawn to their real proportions, alongside the rest. If you want a reversal low with no sharp lows to count at all, the rounding bottom is the slow, curved cousin.
Sources
Every figure above was checked against these before publishing.