Flags, triangles and wedges are all pictures of a pause. Price has made a move, then stopped and coiled while the two sides reprice. The shape tells you a decision is coming and roughly where the line is. It does not tell you which way. Only the break — and the hold — does that.
That distinction is why most pattern trading goes wrong. Traders learn to name the shape and then treat the name as a prediction. A bull flag is not a forecast; it is a boundary drawn around indecision, with a level on each side. What makes the trade is the same thing that makes every trade: a level that breaks, holds, and gives you a price at which you were wrong.
What each shape is actually describing
Strip the vocabulary away and every continuation pattern is one sentence: after a strong directional move, price stopped going anywhere, and the pullback that followed was weaker than the move that preceded it. The differences between the shapes are differences in how the pause is behaving.
| Pattern | What it looks like | What it is saying | Typical resolution |
|---|---|---|---|
| Bull flag | Sharp rise, then a shallow drift down or sideways in a tight channel | Buyers took profit; nobody is aggressively selling | Continuation up |
| Bear flag | Sharp drop, then a shallow drift up in a tight channel | Shorts covered; no real demand appeared | Continuation down |
| Ascending triangle | Flat ceiling, rising floor | A fixed supply level is being tested by progressively more urgent buyers | Usually up, but the ceiling has to go |
| Descending triangle | Flat floor, falling ceiling | A fixed demand level is being tested by progressively more urgent sellers | Usually down, same caveat |
| Symmetrical triangle | Both boundaries converging | Genuine two-sided indecision, narrowing | Direction unknown — wait |
| Rising wedge | Both boundaries rising, ceiling rising slower | Each push higher is achieving less | Often down |
| Falling wedge | Both boundaries falling, floor falling slower | Each push lower is achieving less | Often up |
Notice that the two triangles with a flat side are saying something concrete: there is a specific price where a large, patient order keeps appearing. That is genuinely useful information because it names a level. The symmetrical triangle names two levels and no bias, which is why it deserves less confidence, not more, despite being the shape most often drawn on charts.
The two ingredients that make a continuation pattern worth trading
A shape on its own is a Rorschach test. Two things separate a pattern that resolves from a rectangle you drew because you wanted a trade.
- A real move into it. A flag needs a pole. If the move before the consolidation was slow and choppy, there is no imbalance to continue and the "flag" is just range. The pattern inherits its meaning from the impulse that preceded it.
- A pullback that is proportionally shallow. A bull flag that retraces most of the pole has stopped being a flag and become a reversal in progress. As a working rule of thumb, once the consolidation gives back more than about half the impulse, treat it as a range and re-evaluate rather than as continuation.
Volume adds a third, softer signal: consolidation usually thins out and the break usually does not. Treat that as confirmation, never as a trigger on its own — see what volume actually confirms for why turnout is not the same as direction.
How to set a target without inventing one
The textbook measured move takes the height of the impulse leading into the pattern and projects it from the breakout point. For a triangle, project the height of the widest part of the triangle. This is a reasonable first estimate and a terrible final answer.
The better habit is to draw the measured move and then look at what is between you and it. If a prior swing high, the session high, or a round number sits in the way, that level will resolve first and it is the more honest target. A projection is arithmetic; a level is a place where orders actually exist. When the two disagree, take the level. This is exactly the reasoning behind taking partials at defined levels rather than at a computed price.
Why the same pattern fails constantly
Patterns fail for three specific reasons, and all three are fixable.
- The break was a poke, not a break. Price traded through the boundary intrabar and closed back inside. This is the single most common way a "textbook flag" turns into a loss, and it is a liquidity grab, not a failed pattern.
- The pattern was drawn to fit a bias. Trendlines are hand-placed, and hands are biased. If you had to ignore two wicks to make the line work, the line is not there. Drawing the trendline correctly is a prerequisite for trusting anything built on it.
- The context was wrong. A bull flag on the 5-minute chart forming directly underneath the daily resistance is not a continuation setup; it is a queue at a wall. Checking the higher timeframe first removes most of these.
What the research actually says
It is worth knowing how thin the ground is here, because most pattern content presents folklore as fact. Two pieces of published work are genuinely relevant.
Lo, Mamaysky and Wang built an automated pattern recognizer using nonparametric kernel regression — an attempt to remove the human eye from pattern identification entirely — and applied it to US stocks from 1962 to 1996. Their conclusion was measured: several technical indicators "do provide incremental information and may have some practical value" (Lo, Mamaysky & Wang, "Foundations of Technical Analysis", NBER Working Paper 7613, 2000). Incremental information is a long way from an edge you can trade blind, but it is not nothing.
The broader survey is less flattering and more useful. Park and Irwin reviewed the literature and reported that of 95 modern studies of technical trading, 56 found positive results, 20 found negative results and 19 were mixed (Park & Irwin, "What Do We Know About the Profitability of Technical Analysis?", Journal of Economic Surveys 21(4), 2007). Read that honestly: a clear majority found something, and roughly four in ten studies did not. That is the profile of a tool that works in some hands and some conditions, not a rule of nature. Trade accordingly — with defined risk on every attempt.
Frequently Asked Questions
Which chart pattern is the most reliable?
The question is slightly wrong, because reliability lives in the break rather than in the shape. That said, the bull flag and bear flag are the most tradeable for intraday work, for one structural reason: the consolidation is short and shallow, so the invalidation price sits close to the entry and the position can be sized properly without an oversized loss. A triangle that has been forming for three hours has a wide invalidation and therefore a small position, which changes the arithmetic of the trade regardless of how pretty the shape is.
How do you set a price target from a flag or triangle?
The conventional measured move projects the height of the move that preceded the pattern from the point of the breakout. For a triangle, project the height of the widest part of the triangle from the break. Treat that as a rough distance rather than a promise. In practice it is more useful to check where the measured move lands relative to the next real level on the chart — a prior swing high, a session high, a round number. If the projection sits beyond an obvious level, the level will usually decide the outcome first.
Do chart patterns actually work, or are they just pattern-matching on noise?
The academic evidence is genuinely mixed and it is worth being honest about that. Lo, Mamaysky and Wang built an automated kernel-regression pattern recognizer and applied it to US stocks from 1962 to 1996, concluding that several technical indicators do provide incremental information and may have some practical value. A later survey by Park and Irwin reviewed 95 modern studies of technical trading and found 56 with positive results, 20 negative and 19 mixed. Neither is a licence to trade a shape on sight. Both are consistent with the practical view: patterns describe where a decision is likely to happen, and the decision itself still has to be confirmed.
What is the difference between a wedge and a triangle?
Both narrow, but they lean differently. A triangle has at least one roughly horizontal boundary — a flat top with a rising floor, a flat floor with a falling ceiling, or two lines converging symmetrically. A wedge has both boundaries sloping the same way while still converging, so a rising wedge climbs with a ceiling that rises more slowly than its floor. That slope is the whole point: a rising wedge shows buyers making progressively less headway on each push, which is why it more often resolves downward even inside an uptrend.
Bottom line
Learn the shapes so you can find the boundary quickly, then forget the names and trade the boundary. A flag is a pause with a lid on it; a triangle is a pause with a fixed price on one side; a wedge is a pause where one side is visibly tiring. In every case the trade is the same: mark the line, wait for a close through it, place invalidation where the idea genuinely dies, and size from that distance. For the confirmation rule itself, read break and hold; for telling real breaks from fakes, read what a breakout actually is; and for where all of this sits in a wider toolkit, start with technical analysis basics.