Triple Top and Triple Bottom: 76 Years of the S&P 500 Scanned
A triple top is a chart pattern of three peaks at roughly the same price, confirmed when price closes below the lowest point between the first and third peak. A triple bottom is the mirror image: three lows at about one level, confirmed by a close above the highest point between them.
Three failed attempts at the same high look like a clear message. The trouble is that a market can fail at a level three times and then go straight through it on the fourth, and until the low side breaks there is no way to know which story is unfolding. This page defines the pattern exactly, walks through the S&P 500’s triple top of late 2021, and counts every case since 1950.
How it forms
Price rallies to a high, pulls back, and rallies again to about the same price. That much is a double top. The market then pulls back a second time and makes a third run at the same area, and fails again.
The two dips between the peaks matter as much as the peaks. The lower of the two marks the level that has held twice. Chartists call a line through that low the support or neckline, and it is the line that decides the pattern.
Confirmation is a close below that line. Until it happens, three equal highs describe a trading range with a ceiling, and on the S&P 500 that ceiling gave way more often than the floor did. After it, buyers who defended the lows twice have given way, and the pattern is complete.
The triple bottom is the same process upside down. Three lows at about one price, two rallies between them, and confirmation when price closes above the higher of the two rally peaks.
The usual projection is the pattern’s height. Measure from the highest peak down to the neckline, then subtract that distance from the neckline. This is the same convention as a measured move: a way to fix a number in advance, not a forecast.
The exact definition used here
The scan runs on S&P 500 daily closes, so it looks only at where each day ended. A peak is a close that is the highest of the 11 sessions centered on it, five before and five after. A trough is the same thing for lows.
A triple top is three peaks in a row that satisfy all of these:
- the highest of the three closes is no more than 2% above the lowest;
- at least 10 sessions separate each peak from the next, and the whole shape spans 120 sessions or fewer;
- the lowest close between the first and third peak, the neckline, is at least 3% below the lowest peak.
It confirms on the first close below the neckline within 60 sessions of the third peak, provided no close above the highest peak came first. Two tops may not overlap. The outcome is the close 20 sessions after the confirming close. The triple bottom uses the mirror of every rule.
A worked example
The S&P 500 made three closing peaks between November 2021 and January 2022. It closed at 4,704.54 on 18 November, 4,712.02 on 10 December, and a record 4,796.56 on 3 January.
Check it against the rules. The spread from lowest to highest peak is 4,796.56 divided by 4,704.54, or 1.96%, inside the 2% limit. The lowest close between the first and third peak was 4,513.04 on 1 December, 4.1% below the lowest peak, so the dip was deep enough.
The neckline broke on 20 January 2022, with a close of 4,482.73. That was 12 sessions after the third peak, well inside the 60-session window, and no close above 4,796.56 came first.
Now the projection. The height is 4,796.56 minus 4,513.04, or 283.52 points. Subtracted from the neckline, that gives 4,229.52. The index first closed at or below it on 23 February 2022, at 4,225.50.
Twenty sessions after the break, on 17 February, it closed at 4,380.26, down 2.29% from the confirming close. The index went on to a 2022 low close of 3,577.03 on 12 October. It is also the most recent confirmed triple top in the scan, which is why it is the example.
The original data
Across 19,305 S&P 500 daily closes from 3 January 1950 to 25 September 2026, the rules above found 55 triple tops. Only 19 of them confirmed with a close below the neckline. The other 36, which is 65%, did not: in 35 of them the index closed above the highest peak first, and in one it never closed below the neckline within 60 sessions.
Of the 19 that confirmed, 8 closed lower 20 sessions later, or 42.1%. Across every 20-session window in the same 76 years, the index closed lower 38.2% of the time. The median change 20 sessions after a confirmed break was a gain of 1.03%.
The projection was reached within 60 sessions in 7 of the 19 confirmed tops. A level that is hit about a third of the time is a reference point, not a plan.
Triple bottoms confirmed more often, 14 of 21. Of those 14, 8 closed higher 20 sessions later, which is 57.1%, against 61.8% of all 20-session windows. The projected level was reached within 60 sessions in 9 of the 13 that have had 60 sessions since. The most recent one confirmed on 4 August 2026.
Put together, the pattern on this index usually ended in a break upward, not downward. When it did confirm, what came next was close to what any random 20-session window produced. With 19 and 14 cases, neither gap from the base rate is large enough to separate from chance.
The topic draws little search video. In the 24,971-video corpus this site studies, counted once per video id, 3 titles mention a triple top or triple bottom, from 3 channels, at a median of 1,050 views. Every shape, break and outcome is in the chart pattern event file. The index closes behind the base rates are in the S&P 500 daily close file.
Why three peaks is rare on an index
Three closes within 2% of each other, months apart, need a market that goes sideways at the top. Across 76 years the scan found 1,069 swing highs on the index, and only 55 sets of three in a row met the rule.
Loosen the tolerance and the count jumps. At 3% instead of 2%, the same scan finds 80 triple tops instead of 55, and 47 triple bottoms instead of 21. The 2% band is a choice: tight enough that the three highs look level on a chart, wide enough that an index has a chance to meet it.
Single stocks may meet the rule more or less often; this page did not test them. The same code applies; only the file changes.
When it fails
The most common failure is the breakout. In 36 of the 55 shapes, price never confirmed, and in 35 of those it closed above all three peaks instead. Anyone who sold short at the third high, expecting the pattern, was on the wrong side of a new high.
The second is a confirmed break that snaps back. Price closes under the neckline, and within days it is back above it. That is a false breakout, and it was common here: 11 of the 19 confirmed tops closed back above the neckline within five sessions. A stop placed just above the neckline sits where those snap-backs went.
The third is where the stop has to go. The natural invalidation for a triple top is above the highest peak, which can sit far from the neckline. In the 2022 example the gap was 283.52 points, about 6% of the index. A stop that wide forces a small position, or a large loss when it is hit.
A fourth is calling it early. The third peak only looks like a peak after price has dropped away from it. Treating every approach to an old high as “the third top” means selling into strength many times for each pattern that completes.
And a fifth is ignoring the base rate. The index closed lower over 20 sessions in 38.2% of all windows, pattern or not. A bearish reading that lands at 42.1% on 19 cases has not shown it adds much to that.
Related
The double top page covers the two-peak version and why equal highs draw price toward them. Support and resistance explains the levels this pattern is built from and what happens when one gives way. And measured move sets out the height-projection convention used above, and its limits.
Treat the third test of a level as a question, not an answer. Mark the lowest dip between the peaks, set an alert just under it, and do nothing until a close gets through. On the S&P 500, most three-peak shapes broke out above the peaks instead.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.