Chart Patterns: Head and Shoulders, Flags, Triangles and Their Failures

Chart patterns are the larger shapes price traces over weeks and months — structures built from dozens of sessions rather than the few candles a candlestick pattern compresses. A head and shoulders records an uptrend making one final, weaker high before its structure gives way. A double bottom records the same low tested twice and refused twice. A flag records a sharp move pausing for breath. Where candlestick patterns read the mood of the last few sessions, chart patterns read the whole campaign.

This chapter — part of Gale’s technical analysis guide — draws each pattern as a line sketch of the price path (the sketches connect closing prices; if candle anatomy is new, read how to read candlestick charts first). Each pattern gets the same four-part treatment as the candlestick chapter: what it looks like, where it appears, what it suggests, and when it fails.

The framing rule is the same too: a chart pattern describes a structure the market has built and suggests a tendency for how such structures often resolve — it commands nothing, and none carries predictive certainty. The overview chapter on technical analysis of stocks sets out the framework these shapes belong to.

Every chart pattern is a support and resistance structure

Strip the names away and chart patterns are support and resistance structures — nothing more exotic. A double top is resistance refused twice. A head and shoulders neckline is support tested twice and then broken. An ascending triangle is a resistance line under siege from rising support. That is why every confirmation rule in this chapter reduces to the same event: a defined level giving way on a closing basis. The support and resistance chapter explains why those levels form and why their roles reverse after a break; this chapter names the shapes the fight around them leaves behind.

One corollary disciplines everything below: until the defining level breaks, there is no pattern. Three bumps are not a head and shoulders; two matched peaks are not a double top. The name is earned at confirmation, and most half-formed patterns never get there.

Head and shoulders chart patterns

The head and shoulders is the most studied reversal structure in charting, and deserves the attention for an unglamorous reason: it is a plain description of an uptrend failing, drawn slowly enough to watch.

The anatomy: an uptrend makes a peak (the left shoulder), pulls back, rallies to a higher peak (the head), pulls back again, then makes a third peak that fails below the head (the right shoulder). The line through the two pullback lows is the neckline — the support whose break completes the pattern.

What it looks like: three peaks with the middle one highest, resting on a roughly straight neckline through the two troughs; the neckline may slope slightly either way. Where it appears: at the end of an extended advance — the location is part of the definition. What it suggests: the uptrend’s own definition failing. An uptrend is a sequence of higher highs and higher lows; the right shoulder is the first lower high, the neckline break the first lower low — by completion the trend has been violated twice, and the name just makes it memorable. When it fails: most often by never completing — price consolidates through three bumps and then makes a new high, deleting the “pattern” in hindsight. A completed break can also snap back: a close or two below the neckline, then a reclaim, trapping those who acted on the first tick through.

Two refinements matter. The classical volume signature: heaviest into the left shoulder, lighter into the head, lighter still on the right shoulder — an advance on shrinking conviction — then expanding on the neckline break; a quiet break is a weaker statement by convention. And the throwback: after breaking, price frequently returns to the neckline from below and stalls there — old support auditioning for its new role as resistance, which is why the pattern’s meaning survives a partial bounce.

Inverse head and shoulders

What it looks like: the mirror image — three troughs with the middle one lowest, capped by a neckline through the two intervening highs. Where it appears: at the end of an extended decline. What it suggests: the downtrend’s structure failing in the same two steps — a higher low at the right shoulder, then a higher high through the neckline. When it fails: the volume requirement bites harder on this side. Declines can collapse under their own weight, but advances need actual buying, so an inverse neckline break on thin volume fades back into the base more often than its bearish twin — and a sagging right shoulder is often just a downtrend pausing.

Double top and double bottom

The simplest reversal structures: the same level refused twice. The two extremes rarely match to the rupee — they define a zone, not a line — and the pattern only exists once the level between them breaks.

Double top

What it looks like: an M — two peaks at nearly the same high separated by a trough, then a decline through the trough’s level. Where it appears: after an advance, typically with weeks between the peaks; two highs a few sessions apart are ordinary churn, not a top. What it suggests: demand reached the same price twice and was absorbed twice; the second failure shows the first was not an accident. When it fails: constantly, at the labelling stage — every pause in an uptrend prints two nearby highs, and most resolve upward; the trough break, not the peaks, is the pattern. Resistance refused twice can still break on the third attempt.

Double bottom

What it looks like: a W — two troughs at nearly the same low with a peak between, then an advance through the peak’s level. Where it appears: after a decline, and it means more when the second low prints on visibly lighter volume — sellers pressing with less force. What it suggests: supply exhausted at a price twice; the middle peak breaking is the first higher high of a possible new trend. When it fails: the defended low can simply give way on a later test, and W-shapes inside sideways ranges resolve randomly enough that the location requirement does most of the filtering.

Flags and pennants

Flags and pennants are the continuation family. Both require a pole — a sharp, nearly one-way move — followed by a brief, shallow consolidation that drifts against the pole on shrinking volume. The textbook reading is a pause in an ongoing move, not a reversal.

Bull flag

What it looks like: a steep advance, then a small parallel channel drifting gently downward — the flag on its pole. A bear flag is the mirror: a steep decline, then a shallow upward drift. Where it appears: mid-move, after a burst strong enough to stand out; without a genuine pole there is no flag. What it suggests: profit-taking absorbed without damage — a shallow drift on drying volume reads as a market catching its breath, not changing its mind. When it fails: when it stops behaving like one — a consolidation that retraces deep into the pole or drags on for weeks has become something else, often a reversal wearing the wrong label, and a “flag” against the larger trend is usually just a bounce.

Pennant

What it looks like: the same pole, but the consolidation converges into a tiny triangle instead of a channel. Where it appears: exactly where flags do — mid-move, after a sharp thrust, lasting days rather than weeks. What it suggests: the same pause with the range compressing; the coiling itself is the statement. When it fails: the break can come against the pole; because the structure is so small, one wide session can pierce both lines and mean nothing; and a pennant without a pole is just a small symmetrical triangle, with no lean at all.

Triangles

Triangles are consolidations defined by two converging trendlines; drawing them honestly needs at least two touches on each line — anything less is a guess wearing a ruler. The variants differ in which side is doing the work.

Ascending triangle

What it looks like: a flat ceiling tested repeatedly while each pullback bottoms higher — rising support squeezing price against horizontal resistance. Where it appears: often within an uptrend, beneath a well-defined level such as a prior high. What it suggests: buyers stepping in at progressively higher prices while sellers defend one line; the conventional lean is that supply at the ceiling is being eaten through. When it fails: the ceiling holds and the rising support line breaks instead, and false upside breaks at a flat line are common precisely because the line is widely watched.

Descending triangle

What it looks like: the mirror — a flat floor tested repeatedly while each bounce tops out lower. Where it appears: often within a downtrend, above a level that has attracted buyers before. What it suggests: sellers pressing at progressively lower prices against one defended line; the conventional lean is bearish — demand at the floor being worn down. When it fails: floors defended that many times sometimes hold, and in a rising broader market descending triangles break upward often enough that the bearish label deserves its qualifier.

Symmetrical triangle

What it looks like: lower highs and higher lows converging toward an apex — both sides giving ground. Where it appears: anywhere a trend pauses; it carries no directional lean of its own. What it suggests: compression — the range narrows, volume shrinks, and the coil stores the disagreement until one side runs out of patience, in whichever direction that happens. When it fails: near the apex — breaks after price has drifted deep into the point are weak by convention, and apex churn produces whipsaw breaks in both directions within days.

Rising and falling wedges

Wedges look like triangles but both lines slope the same way, and the reading inverts intuition: a wedge leans against its own slope.

Rising wedge

What it looks like: price grinding higher inside two rising, converging lines — each push up smaller than the last. Where it appears: late in advances, or as a laboured correction inside a downtrend. What it suggests: effort rising, progress shrinking — highs still print, but each rally achieves less; momentum thins while price technically climbs. When it fails: strong trends can grind through the apex and keep rising, and the lines are the weakness: wedges drawn generously around any decelerating advance are among the most over-identified structures on charts.

Falling wedge

What it looks like: the mirror — price stepping lower inside two falling, converging lines, each decline shallower than the last. Where it appears: late in declines, or as a drifting correction inside an uptrend. What it suggests: selling pressure fading with each push; the downtrend still printing lower lows but with visibly less force. When it fails: fading pressure is not arriving demand — a falling wedge can keep falling in slow motion for months — and the upside break needs volume for the same reason every bullish break does.

Rounding bottoms and the cup with handle

A rounding bottom is a saucer: a decline that decelerates, flattens and curls upward over months, no single dramatic low — sentiment turning on geological time. Its failure mode is never finishing: the rim rejects the advance and the saucer flattens.

What it looks like: a rounded base that climbs back to its old high (the rim), then a small, shallow dip — the handle — before a break above the rim. Where it appears: on weekly-scale charts, classically in stocks recovering from long corrections. What it suggests: the last impatient holders selling into the rim (the handle), leaving less supply overhead for the breakout. When it fails: the handle keeps deepening into a new decline, and because the structure takes months, hindsight finds beautiful cups on every chart that already went up.

The measured move: target arithmetic as convention

Every pattern above arrives with a textbook target; the arithmetic is worth knowing, filed under convention, not prediction.

The rule is uniform: project the pattern’s height from the point of the break. Head and shoulders: the vertical distance from the head to the neckline, projected downward from the break (the inverse projects upward). Double tops and bottoms: the height between the extreme and the middle level. Triangles and wedges: the height at the widest point. Flags and pennants go grander — project the entire pole from the breakout, the “flag flies at half mast” rule.

This arithmetic is a shared discipline, roughly a century old, for asking “how far might this travel?” in a way two chartists can agree on. It is not a forecast. Measured targets are reached in only a fraction of cases — treat any precise success percentage as sample-dependent — and price that does travel rarely stops on the projected line. The honest use is proportion: comparing the conventional target against the distance to the level that would invalidate the pattern says whether the structure offers more potential than risk by its own textbook’s rules — arithmetic about a convention, not a promise from the market.

The cheat sheet: every pattern in one table

Pattern	Shape	Bias	Confirmation	Failure
Head and shoulders	Three peaks, middle highest	Bearish reversal	Close below neckline	Never completes; break snaps back
Inverse head and shoulders	Three troughs, middle lowest	Bullish reversal	Close above neckline	Thin-volume breaks fade
Double top	Two matched peaks, trough between	Bearish reversal	Close below middle trough	Most twin peaks resolve upward
Double bottom	Two matched lows, peak between	Bullish reversal	Close above middle peak	Defended low gives way later
Bull flag	Pole up, shallow falling channel	Bullish continuation	Break above channel	Deep or prolonged drift; no real pole
Bear flag	Pole down, shallow rising channel	Bearish continuation	Break below channel	Drift becomes a reversal
Pennant	Pole, tiny converging coil	Continuation with the pole	Break in pole’s direction	Break against the pole; no pole
Ascending triangle	Flat top, rising lows	Bullish lean	Close above flat line	Ceiling holds; support breaks instead
Descending triangle	Flat floor, falling highs	Bearish lean	Close below flat line	Floor holds; resistance breaks instead
Symmetrical triangle	Converging highs and lows	Neutral until break	Close beyond either line	Whipsaw breaks near the apex
Rising wedge	Both lines rising, converging	Bearish lean	Break below lower line	Trend grinds through the apex upward
Falling wedge	Both lines falling, converging	Bullish lean	Break above upper line	Slow-motion decline continues
Rounding bottom	Long saucer of lows	Bullish reversal	Close above the rim	Rim rejects; saucer flattens into drift
Cup with handle	Rounded base plus shallow dip	Bullish continuation	Close above the rim	Handle deepens into a new decline

How reliable are chart patterns?

The candlestick chapter’s honesty section applies here, plus an extra problem.

The subjectivity is worse. Candlestick patterns are at least computable — an engulfing body either covers the prior body or it does not. Chart patterns are drawn. Two people with the same chart will disagree about where a neckline sits, whether a third touch counts, whether a consolidation is a flag or the start of a top. Structures this dependent on the drawer’s judgment are this dependent on the drawer’s hopes.

Patterns are catalogued in hindsight. Every named structure was identified by studying charts where the ending was known, which finds real recurring behaviour but flatters it: successful examples got the names; the three bumps that became nothing were never counted. Published attempts to measure pattern statistics find meaningful failure rates for every structure in this chapter, with success rates that shift by market, decade and sample. Suppose a completed pattern resolved as expected in 60 of 100 historical cases — an illustration, not a measured figure. That is a real tendency carrying a 40% failure rate — why the invalidation level matters more than the name.

Most forming patterns never complete. This is where the money is lost: acting on a head and shoulders before the neckline breaks means acting on a pattern that does not yet exist. The completion is the pattern.

Vocabulary is not an edge. SEBI’s study of individual equity F&O traders found roughly nine in ten lost money over the three years examined — a population well stocked with people who know every shape on this page. In thinly traded smallcaps the problem compounds: prices gap through necklines and print false breaks on single large orders; whatever reliability these structures have lives mostly in liquid names.

Where patterns meet the live screens

Every completed pattern in this chapter resolves the same way: price closing through a defined level — a breakout, and breakouts are measurable in advance. Gale.in computes breakout candidates daily from exchange data: the near-breakout stocks screen lists stocks trading within a small distance of their recent highs, exactly where ascending triangles, cup handles and flag consolidations sit in their final, unconfirmed sessions. The screen does not name patterns — it measures the distance to the level, which this chapter has argued is the part that matters.

Trend context comes first: the moving averages chapter covers the 20/50/200-day conventions that separate a flag inside an uptrend from a bounce inside a decline. The screens are this chapter’s vocabulary applied to today’s market, refreshed daily.

What to weigh

• The level is the pattern. Until the neckline or trendline breaks on a closing basis, there is no head and shoulders, no triangle, no flag — only a range with a story attached.
• Structure over name. A pattern earns respect by describing trend structure changing — lower highs appearing, a defended floor giving way. Absent that logic, the shape is decoration.
• Volume as the honesty check. Breaks on expanding participation and breaks on a quiet session are different events wearing the same shape — the rule the candlestick chapter closes on.
• Measured moves set proportion, not prophecy. The target arithmetic is a shared convention for weighing potential against invalidation, nothing more.
• The base rate, minus costs. Completed patterns fail at material rates and forming ones mostly never complete; position size and the predefined exit decide what a wrong read costs.
• The next step. Candlestick patterns read the individual sessions inside these larger structures; the live screens show which stocks press against their defining levels today.

Gale.in is not a SEBI-registered investment adviser or research analyst; this chapter is education, not a recommendation to buy or sell any security.