What Are Chart Patterns?

Michael Whitman · Trading since 2017, full-time since 2020 · Published 2026-09-05 · Updated 2026-09-26 · Reviewed 2026-09-26

A chart pattern is a named shape formed by price, such as a double top, a flag or a triangle. Each one comes with a line that has to break before the pattern means anything, and that line is a support or resistance level. The name describes the shape; the level is what you trade.

Key takeaways

Every pattern has a line, and nothing happens until that line breaks.
The shape is a description. The level is the trade. Strip the name and you still have everything you need.
A double top is two touches of resistance and a support level below. That is the whole thing.
The measured move — project the pattern’s height from the break — is a convention, not a mechanism.
They are only patterns once complete, which is always after the entry would have been.
Of 287 chart pattern videos studied, the median gets 2,307 views and 79% never pass 50,000 — among the most saturated topics measured.
On SPY since 1993, 52 pairs of near-equal highs broke their neckline no more often than a random walk would have, and 14 of the 23 breaks were back above the top within 63 trading days.

Covered on this page: TradingView.

A chart pattern is a named shape in price, such as a double top, a flag or a triangle, that only counts once price closes through a support or resistance line drawn from it. On SPY since 1993, pairs of near-equal highs broke that line no more often than a random walk would have.

There are dozens of named shapes and one idea underneath all of them. This page covers three patterns and then the thing they have in common, which is more useful than the other twenty-five.

How it forms

Take the best-known one.

A double top is price reaching the same high twice and failing both times. That is the shape, and on its own it is not tradeable — plenty of charts do this and carry on upward.

What makes it a pattern is the neckline — the low between the two peaks. Until price closes below that, there is no double top, only two highs.

The part worth keeping

That is the identical chart described as two levels. A level tested twice and held, and a second level below that broke.

Everything the pattern told you is in that description, and it needed no name. The shape is a way of talking about the levels; the levels are what price is actually reacting to.

This holds for all of them, which is why this page teaches three rather than thirty.

Two more shapes

A flag is a strong move and a tight pause. Underneath: a pullback that held, inside a trend.

A triangle is a range that narrows. Underneath: compression, and the expansion that follows it — the same idea as the quiet-then-loud section on the price action page.

Two names, two shapes, and both resolve to things this site already covered without them. The named chart patterns page writes each of the others as a condition you could check.

Continuation vs reversal chart patterns

The only classification that changes what you do, and it is worth having instead of the names.

A continuation pattern is a pause inside a move — flags, pennants, most triangles. Price arrived going one way and is expected to leave the same way. The case usually made for this group is that it agrees with what is already happening.

A reversal pattern claims the direction is changing — double tops and bottoms, head and shoulders. That is a bigger claim about a rarer event, and it needs more to go right.

The practical rule: ask which one you are looking at before you ask what it is called. A shape that agrees with the market structure around it is a different proposition from one betting against it, whatever the textbook calls each.

And if the shape disagrees with structure, structure is the thing with a specific invalidation price in it. The shape only has an outline.

The measured move

The convention is to take the height of the pattern and project it from the break as a target.

It is a convention, not a mechanism. Nothing about a market makes the next move the same size as the last one — but it gives you a number to aim at before you enter, which is better than deciding mid-trade, and that is the honest case for it. The measured move page takes the convention apart on its own.

A worked example

Price makes a high, pulls back, and makes the same high again. Nothing to do. Two highs is a trading range so far.

You mark the low between them. That is the line, and it is the only price that matters.

Price closes below the line, and the next candle continues. Now the pattern exists — and note the order: the pattern is confirmed by the break, not the other way round.

The stop goes back inside the shape, above the neckline, because if price is back there the break did not happen.

And the target comes from the measured move, decided at the same moment as everything else. Not because the projection is reliable, but because a number chosen now is worth more than one chosen while the position is open — the argument on the entry and exit page.

A real one: SPY, January to April 2026

SPY printed a high of $697.84 on 28 January 2026 and $697.14 on 11 February, 0.10% apart. The lowest low between them was $675.79 on 5 February. That is the neckline.

On 6 March SPY closed at $672.38, below the line, sixteen trading days after the second peak. The height was $697.84 minus $675.79, or $22.05, so the measured move pointed to $675.79 minus $22.05, or $653.74. SPY closed below that on 20 March and bottomed at a $631.97 close on 30 March.

Then it undid the whole thing. On 15 April SPY closed back above $697.84, the higher of the two tops. Every textbook step happened in order — the shape, the break, the target — and six weeks later price stood above the top the pattern was named for.

The original data

Two equal highs on SPY, 1993 to 2026

We tested the double-top picture on every SPY daily bar from 29 January 1993 to 25 September 2026, using Yahoo Finance’s unadjusted high, low and close. The rules, fixed before counting:

A peak is a day whose high is above the highs of the five days before and the five days after (473 of them).
A pair is two peaks in a row, at least ten trading days apart, with a dip of 3% or more from the lower peak to the neckline, the lowest low between them (245 pairs).
Equal means the second peak is within 1% of the first (52 pairs; the other 193 are unequal).
Resolution is whichever comes first after the second peak: a close below the neckline, or a close above the higher peak. All 245 resolved by 25 September 2026.

Of the 52 equal pairs, 23 broke down through the neckline first and 29 closed above the tops first. Seven of the 23 broke before the second peak could even be recognized, which takes five more days.

From the day the second peak was recognizable, 16 of the remaining 45 broke down (35.6%). A random walk starting from that day’s close, with the two lines where they actually sat, would have hit the neckline first 42.5% of the time on average, about 19 of 45. The unequal pairs did the same thing: 63 of 173 broke down (36.4%) against 37.9% for the random walk. Neither gap is distinguishable from chance (two-sided binomial p = 0.37 and 0.75). The equality of the two highs, which is what gives the pattern its name, did not show up in how the pairs resolved.

And a completed break was often undone. After 14 of the 23 equal-pair breakdowns, SPY closed above the higher top again within 63 trading days; for unequal pairs it was 31 of 83. That is a small sample in a market that rose over the period, so read it as a warning about the upside risk after a break, not a rate to plan around. The median equal pair took 10 trading days from the second peak to resolve, with a 4.6% dip between peaks 15 trading days apart. Every pair is listed in the SPY equal-high pairs table.

What the videos teach

Across our study of 24,971 trading videos, 287 have chart pattern in the title. The median one gets 2,307 views, 79% never pass 50,000, and the median length is 9.7 minutes.

That is one of the lowest medians measured for this glossary — below breakouts at 6,187 and far below trend lines at 61,260.

The likeliest reading is saturation. Chart patterns are the oldest, most reproduced content in trading education: 287 videos explaining shapes whose definitions have not changed in decades.

The named shapes do better than the umbrella term. Titles naming a double top number 54 with a median of 14,482 views; head and shoulders 46 at about 3,550; a flag pattern, bull flag or bear flag 51 at 2,702; and triangles 78 at 1,133.

The corpus carries description text for 138 of those 287 — one of the largest samples in this glossary — and across those 138, 3 mention invalidation, failure, or what a bad read looks like.

How this connects

Patterns are descriptions of shape, and every one of them is built from candlesticks on a timeframe you chose.

Most reduce to two families. Compression - triangles, wedges, flags - and reversal shapes like double tops and head and shoulders.

Support and resistance does the real work in all of them. The boundaries are levels, and a level tested repeatedly is the part with a mechanism behind it.

A breakout is how nearly all of them are traded, which means they inherit the false-breakout problem wholesale.

And the trading range is the honest null case. A great deal of price action is neither a pattern nor a trend, and reading shapes into it is where most pattern trading goes wrong.

When it fails

The pattern completes and does the opposite

Two touches of the same high is also what a level looks like just before it breaks. The double top and the breakout are the same picture until one of them resolves — and on SPY, 29 of 52 equal pairs resolved upward.

It was only a pattern once it finished

The same chart stopped at the second peak, before any break. — At the second high. Double top, or a range? Illustrative chart - not real market data.

Every completed pattern is obvious and every incomplete one is ambiguous, and you can only ever act inside the second condition.

The break came back

A close through the neckline is the confirmation, and it is not the end. The 2026 SPY example above reached its measured target and still closed above its top six weeks later. Decide before the break where the pattern is wrong on the other side, not only where it is right.

The pattern was too small to matter

A double top spanning six candles on a five-minute chart is a shape almost nobody saw. The same shape across three months on a daily chart was watched by thousands.

Size and timeframe decide how much is resting behind the line, and a textbook pattern on a chart nobody is watching has nothing behind it at all.

You are pattern-matching a random chart

Humans find shapes in noise reliably, and a chart with thirty named patterns available will always contain one that fits what you already want to do. The level test is the defense: if you cannot name the line and what happens when it breaks, there is no trade in the shape.

Support and resistance is what every pattern on this page resolves to, and it is the more useful thing to learn first.

Breakout is what the neckline break actually is, including why it fails so often.

And trading range is what a double top is before it becomes one. To work a double top step by step, see how to trade a double top.

The practical check

Trade levels, not shapes. Look at where the level is and whether it broke properly — a close through it and then a candle going the same way. If a shape happens to be there too, fine, but whether it is called a bear flag or a pullback should not change a decision; the names are a way of talking about charts, not a way of reading them.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money. Some links on this page earn a commission if you buy through them. It costs you nothing and it does not decide what appears here or in what order — how these pages are made is set out in our methodology.