Chart patterns: continuation and reversal

Chart patterns are where technical analysis is at its most seductive and its most dangerous. Seductive because a clean head-and-shoulders or a textbook triangle looks like a secret code the market has handed you. Dangerous because that exact feeling — I've spotted the code — is precisely the trap. So we'll do this honestly: every pattern in this chapter is just a picture of supply and demand doing something, and every one of them fails a great deal of the time. Learn the supply-and-demand story behind each, hold the shape lightly, and you can use patterns sensibly. Memorise shapes as prophecies and you'll lose money with a ruler in your hand.

Two families: continuation and reversal

Patterns sort into two broad families based on what they suggest about the existing trend.

• Continuation patterns suggest the prevailing trend is merely pausing — catching its breath — before resuming in the same direction. Flags and many triangles fall here. The story is: the trend is intact, this is just a rest.
• Reversal patterns suggest the prevailing trend may be ending and about to turn the other way. Head-and-shoulders and double tops/bottoms fall here. The story is: the side that's been winning is finally exhausted.

Crucially, the same shape can mean different things depending on the trend it appears in and whether the expected breakout actually happens. A pattern is a hypothesis about what comes next — and like all hypotheses, it's only worth anything if you've decided in advance what would prove it wrong.

Flags: the trend catches its breath

A flag is the simplest continuation pattern. After a sharp, strong move (the 'flagpole'), price drifts sideways or gently against the move in a tight range (the 'flag') before, often, breaking out to continue the original direction. The supply-and-demand story is intuitive: after a powerful surge, some buyers take profits and the move pauses, but no real selling pressure builds — the pause is shallow and orderly. When the brief profit-taking is absorbed, the original force reasserts itself.

Triangles: a coiling spring of indecision

Triangles form when the trading range narrows over time — the highs and lows converge toward a point, like a spring coiling tighter. The story is a tightening battle between buyers and sellers, with neither able to push price far before the other pushes back, until the range gets so tight that something has to give.

• An ascending triangle has a flat top (resistance) and rising lows — buyers getting more aggressive while sellers hold a fixed ceiling. Often (not always) resolves upward.
• A descending triangle has a flat bottom (support) and falling highs — sellers pressing while buyers hold a fixed floor. Often (not always) resolves downward.
• A symmetrical triangle has both converging — pure indecision, and genuinely ambiguous about which way it'll break.

Head-and-shoulders: a reversal story

The head-and-shoulders is the most famous reversal pattern, and it tells a clean supply-and-demand story about a trend running out of buyers. After an uptrend, price makes a high (left shoulder), pulls back, makes a higher high (the head), pulls back to a similar level, then makes a lower high (right shoulder). The line connecting the pullback lows is the neckline. The story: each successive rally is finding fewer willing buyers — the higher high of the head couldn't be matched by the right shoulder — and when price finally breaks below the neckline, the buyers have decisively given up.

The mirror image, the inverse head-and-shoulders, appears after a downtrend and tells the opposite story — sellers progressively running out of ammunition — suggesting a possible turn upward. In both cases what you're really reading is a weakening of the dominant side, drawn as a sequence of pushes that each fail a little more than the last. That underlying logic — momentum fading across successive attempts — is far more important than the pretty silhouette.

Double tops and double bottoms

A double top is two distinct peaks at roughly the same level with a dip between — price tried twice to push above a ceiling and failed both times. The story is a clear, twice-confirmed resistance: the second failure at the same level suggests buyers can't break through, and a fall below the intervening low signals they've given up. A double bottom is the mirror — two failed attempts to break below a floor, hinting sellers are exhausted.

These are really just support and resistance dressed up with a name. A double top is the market testing a resistance zone twice and being rejected both times — exactly the supply overhang we discussed in the levels chapter, now happening across two visible attempts. Seeing the pattern this way keeps you grounded: it's not a mystical 'M' shape, it's resistance holding, observed twice.

Base-rate humility: patterns fail, often

Now the most important section in this chapter, and the one the pattern-selling industry never wants to dwell on: chart patterns fail a large fraction of the time, and even the 'good' ones are right only somewhat more than chance. There is no pattern that 'works' most of the time in any dependable way, and anyone claiming otherwise is selling something. The honest framing is the one from the very first chapter — patterns shift probabilities, and a pattern that's right, say, slightly more often than not is still genuinely useful, and will fail close to half the time.

This is also why pattern trading lives or dies on risk management, not pattern-spotting. Since you know a fair share of patterns will fail, the only way to come out ahead is to lose small when they fail and gain more when they work — a pre-set stop-loss on every trade, sized so that the failures are survivable. A trader who's right 55% of the time and rigorously cuts losses can prosper; a trader who's right 55% of the time but lets the 45% run into disasters will be wiped out. The edge, if it exists at all, is in the discipline around the pattern, not the pattern itself.

Using patterns honestly

Put it all together and a sane approach to chart patterns emerges. Treat a pattern as a hypothesis, not a forecast. Demand confirmation — a decisive close beyond the relevant level, ideally with supporting volume (the subject of the next sub-module) — before acting, never anticipating the breakout. Decide your invalidation price before you enter, so a failed pattern costs you a small, pre-agreed amount rather than an open-ended disaster. And weigh patterns by context: the same shape means more when it aligns with the larger trend and forms at a significant level than when it appears in a vacuum. Do all that, and patterns become one modest input among several. Skip any of it, and you're just drawing pictures and praying.