Three peaks at about the same price, two dips between them, and a line under the dips. Every broker’s education page draws it that way and then moves straight on to where to sell. This post stops at the drawing, because that is where most triple-top reads go wrong: the peaks were not really at one level, there was no trend to reverse, or the pattern had not actually happened yet.
A triple top does not exist until price closes below the neckline. Before that close, what you are looking at is a trading range that has been tested from below three times, and ranges resolve in both directions.
What three peaks at one price actually record
Price rallies to a level and stops. It pulls back, rallies again on whatever the market now knows, and stops at the same place. Then it does it a third time. Each of those three rallies is a separate group of buyers deciding the price is worth paying, and each time enough supply arrives at the same price to absorb them.
That is the whole mechanism: one price where sellers 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 third failure carries more weight than the second only if it comes with less progress. If the third rally is slower, peaks on smaller candles, or gets there on shrinking volume, buyers are running out of conviction at exactly the level where they most need it. Bulkowski’s identification notes record the typical shape: volume across the pattern trends downward 62% of the time, while it is usually high beneath each peak. A third rally 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 peaks at roughly the same level” and leave “roughly” to you. Here are the rules with the ambiguity taken out.
1. There was an uptrend to reverse. Bulkowski lists the price trend as “upward leading to the pattern,” and adds the reason plainly: price must have something to reverse, so a small rise into the pattern tends to produce a small decline out of it. A stock that has been sideways for months and touches the top of its range three times is a range, not a triple top, whatever the shape looks like.
2. Three peaks inside a tolerance band. First measure the pattern’s height: the highest of the three peaks to the lowest of the two troughs. Then measure the spread between the highest and lowest peak. 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 descriptions say only “near the same price,” and it is fine to pick your own. What is not fine is deciding after you have looked. Fix the fraction, then apply the same one every time.
Illustrative prices to make it concrete: peaks at 100.0, 99.6 and 100.3 with troughs near 94 give a height of about 6 and a spread of 0.7, well inside the band. Peaks at 100, 97 and 102 with the same troughs are not one level; that chart is a range with a slope in it, and the reading changes.
3. Two intervening troughs. They do not need to match, but both should sit clearly below all three peaks. If the second is far shallower than the first, price is compressing toward the level, which is a different structure with a different meaning.
4. The neckline through the lower trough. Bulkowski’s confirmation rule uses the lowest valley in the pattern, so the line goes there. Some traders draw it through both troughs when they are close; when they are not, the lower one is the honest choice, because it is the price the pattern must actually reach. A middle peak marginally below the other two is within the definition.
A chart that passes all four is a candidate. It is not yet a triple top.
When does the range become a triple top?
On a close below the neckline, on the same timeframe as the chart you are reading. A wick below 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 “becomes valid when price closes below the lowest valley in the pattern.”
Until that close, the sellers at the peaks have won three rounds and the buyers at the troughs have won two, and nothing on the chart says which group gives way first. The neckline break is the moment the buyers who bought the dips stop buying them.
One more thing the same source records about what happens after confirmation: a pullback rate of 66%. In that sample, two thirds of confirmed triple tops saw price return to the neckline before continuing. A retest of the line 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 above the highest of the three peaks. The level that absorbed three rallies has been taken, and the sensible description of the chart is now a range breakout upward with three defined rejections behind it. Whatever you had called it, it is no longer a triple top.
After confirmation: a close back above the neckline weakens the read, and a close above the highest peak finishes it. Given the pullback figure above, a touch of the neckline from below 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 peak to the lowest trough, and project it downward from the neckline. Bulkowski’s version of the rule multiplies that height by the historical percentage meeting the target before subtracting it, which is already an admission that the full height is more often missed than reached.
The numbers on that page. The results are labelled Important Bull Market Results, are based on 1,964 perfect trades, and carry a statistics-updated date of 8/27/2020. Within that sample: overall performance rank 24 out of 36, a break even failure rate of 25% (the page’s term, defined in its glossary), an average decline of 14%, a pullback rate of 66%, and 49% meeting the price target. The page’s own summary calls the failure rate higher than the author likes and the average decline meagre.
Read those as a description of one dataset, not as odds on the chart in front of you. In that sample a confirmed triple top produced a modest decline on average and reached the full measured move about half the time. The numbers say nothing about your instrument, your timeframe, or the pattern you are looking at, and this post is not a forecast for any of them.
Where the triple top misleads
Three touches in a flat market. The most common false read. A stock that has drifted sideways, tested the top of its box three times and sat back down is a box. There is no trend for the pattern to reverse, and the same three touches will look identical whether the box eventually breaks up or down. Rule 1 exists for this case.
It is not a different animal from a double top. A triple top is a double top whose neckline break did not come, where buyers stepped in at the trough and produced a third rally that failed at the same place. The extra peak adds one more rejection to the evidence and moves the neckline to the lowest trough. It does not change what the pattern means. One detail from Bulkowski’s trading notes: when the third peak sits below the middle one, the page reports a stronger decline on average, because the last attempt at a new high fell short of the previous one.
The failed triple top is an outcome, not an anomaly. The three-peak shape resolves upward often enough that a reader who has not planned for it has only read one of the two outcomes. How often exactly is a question the published evidence does not answer: Bulkowski’s statistics count only patterns that confirmed with a downward breakout, so a three-peak range that broke upward was never counted as a triple top at all. The honest position is that the frequency is unmeasured in the sources available, which is why the invalidation level above the peaks is part of the read from the start.
What the third peak often looks like up close
Zoom into the third rejection and the candles at the peak often form one of two bearish patterns, because they describe the same event at a smaller scale.
The shooting star is a single candle with a small body near the bottom of its range and a long upper wick, forming after a rise. Price was carried well above the open and sold back down before the close, which is exactly what a third rally into supply looks like in one session. It needs the rise before it, and the third leg of a triple top is one; a close above its high invalidates it.
The bearish engulfing is an up candle followed by a down candle whose real body covers it entirely: the session after the peak undid the one before it. Bodies are compared, not wicks, and engulfing a large body means more than engulfing a tiny one. A close above the engulfing candle’s high ends that read.
Either candle at the third peak is a detail inside the rejection, not a triple top on its own. It can show the third attempt failing a few sessions before the neckline does; it cannot show that the neckline will break. The full set of these, with the context each one needs, is in bearish candlestick patterns.
The habit this pattern teaches
The triple top is useful mostly because it forces the discipline every pattern should: a definition you can test, a moment it becomes real, and a price that proves it wrong. Three peaks inside a band you fixed in advance, after a rise, a close below the lowest trough to confirm, a close above the highest peak to invalidate. Anything short of that is a range tested three times, and a range is allowed to go either way. The triple bottom pattern is the same structure upside down, held to the same discipline. A range that narrows instead of repeating its highs is a different read, and the symmetrical triangle pattern treats it the same way: no direction until a close outside the lines says so.
If troughs, 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 shooting star and the bearish engulfing drawn to their real proportions, alongside the rest.
Sources
Every figure above was checked against these before publishing.