Continuation Patterns

"Continuation" is not a prediction

Reversal patterns flip a trend. Continuation patterns — triangles, flags, pennants, wedges, rectangles — allegedly pause a trend before it resumes in the same direction. Every chart-pattern textbook you'll read treats "continuation" as near-deterministic.

Bulkowski's data blows that up. Most so-called continuation patterns are really coin-flips with a modest bias — and a few are worse than random. Murphy hedges appropriately:

Triangles are usually continuation patterns, but sometimes act as reversal patterns. — John Murphy

Bulkowski makes the hedge specific. For each pattern he counts R (reversal) versus C (continuation) outcomes:

Look at that last row. A descending triangle — the textbook's bearish continuation pattern — resolves as a reversal two-thirds of the time when it breaks down in a bull market. The classification "continuation pattern" is a statistical preference, not a rule.

The patterns — gallery

Toggle the five patterns above. The trendlines are drawn in. Each narrative ends with Bulkowski's actual performance rank and breakeven failure rate. Notice the stats — they're not what textbooks lead you to expect.

1. Symmetrical Triangle — the overrated classic

Two converging trendlines; no directional slope bias; requires at least five trendline touches (three on one line, two on the other) per Bulkowski, or the pattern isn't really a triangle.

Murphy's two-thirds rule is the rare classical rule Bulkowski's data confirms:

Prices should break out in the direction of the prior trend somewhere between two-thirds to three-quarters of the horizontal width of the triangle. If prices remain within the triangle beyond the three-quarters point, the triangle begins to lose its potency, and usually means that prices will continue to drift out to the apex and beyond. — Murphy

Bulkowski's median breakout occurs at 72–75% of the way to the apex — right in Murphy's range. If a triangle drifts past 75% without breaking, it's tired. Take it off your watch list.

The headline stats (upward breakout, bull market):

• Performance rank: 36 of 39 — nearly the worst
• Breakeven failure rate: 25%
• Average rise: 34% (vs 42.4% all-pattern average)
• % meeting measure-rule target: only 58%

Bulkowski's own verdict:

Performance is, well, awful. Performance ranks near the bottom of the list of chart patterns.

The symmetrical triangle is the poster child of continuation patterns in every textbook. It's also one of the worst-performing patterns in a 150,000-sample database. The lesson: pattern recognition without pattern rating is a random-walk generator.

2. Ascending Triangle — a legitimate bullish edge

Flat resistance top, rising higher-lows. The shape tells a clean story: buyers are willing to pay progressively more while sellers hold a fixed price ceiling. Eventually, sellers run out.

Stats (upward breakout, bull market):

• Performance rank: 16 of 39 — upper half
• Breakeven failure rate: 17%
• Average rise: 43% — above the all-pattern average
• Volume trend: downward during the pattern (as expected)
• % meeting target: 70%
• Throwback rate: 64%

Breakout direction: 63% up / 37% down. The bullish bias is real.

The trap: downside breakouts of ascending triangles are statistically bad trades. Bulkowski's gun-to-my-head quote:

I would not consider shorting a stock showing an ascending triangle with a downward breakout unless the situation was compelling (like someone held a gun to my head).

If an ascending triangle breaks down, the price action is telling you something surprising. Step aside — don't try to profit from the surprise.

3. Descending Triangle — trade only in bear markets

Mirror of ascending: flat support bottom, falling lower-highs. Supposed to be bearish.

In bull markets, price breaks UP 53% and DOWN 47% — the "bearish" pattern is essentially neutral. Worse, the downside breakouts that do occur often resolve as reversals back into uptrends (34% C / 66% R).

In bear markets, the pattern earns its reputation — but in bull markets, trading a descending triangle short is the beginner mistake. Bulkowski's rule:

Short only in bear markets.

4. Flag — "fly at half-mast"

A sharp move (the "pole") followed by a tight parallel-channel consolidation (the "flag"), resolving in the direction of the pole. Classical Murphy:

Flags and pennants are said to "fly at half-mast" from a flagpole.

Project the flagpole height from the breakout point — that's your target. Simple, visual, everybody loves flags.

The cold Bulkowski stats (upward breakout, bull market):

• Breakeven failure rate: 44%
• Average rise: 9%
• % meeting target: 46%

Forty-four percent of bull-market flag breakouts fail to advance even 5% before reversing. The "half-mast" measure rule hits fewer than half the time. Bulkowski's fix:

Try computing the measure rule target but use something like 75% of the flagpole height to get a more conservative target.

Flags are so visually obvious that they trade on their reputation, not their statistics. If you use them, size small and take partial profits well before the measure-rule target.

5. Pennant — the pattern Bulkowski says doesn't work

Same idea as a flag but the consolidation is a tiny symmetrical triangle instead of a parallel channel. Duration 1–3 weeks (Bulkowski's sample averages 8 days).

Bulkowski tested the "half-mast" claim and published a flat verdict:

Pennants serve as half-staff patterns… Does it work? No.

Stats (upward breakout, bull market):

• Breakeven failure rate: 54% — coin flip
• Average rise: 7%
• % meeting target: only 35%

More than half of upward-pennant breakouts fail to advance 5%. Most of the ones that do fail to reach their projected target. If you're still using pennant signals, you're trading on the aesthetics, not the edge.

6. Rectangle — an elite continuation pattern

Horizontal support + horizontal resistance. Price chops within a range for a while, then resumes.

Rectangle Top (trade range inside an uptrend, breaks up) stats:

• Performance rank: 4 of 39 — top 10%
• Breakeven failure rate: 15%
• Average rise: 51%
• % meeting target: 78%
• Throwback rate: 66%

Compare this to the symmetrical triangle (rank 36). Same conceptual idea — horizontal consolidation — but dramatically different stats. Why? Because a rectangle requires unambiguous horizontal lines, which implies a real supply/demand equilibrium at specific prices. A triangle's converging lines are softer; they can be drawn many ways.

The catch: if the rectangle breaks the wrong way, rank drops to 32/36. Trend context determines everything. A rectangle-top that breaks down is not a strong short — it's a weak trade because the shape lost its context.

7. Wedges — the counter-intuitive ones

Murphy's key rule, worth memorizing:

A rising wedge is bearish, and a falling wedge is bullish.

Both lines slope the same direction, but one line slopes more steeply, so the shape converges. In a rising wedge, higher highs come in slower than higher lows — buyers are running out of steam even as price drifts up. In a falling wedge, lower lows come in slower than lower highs — sellers are exhausting themselves.

Falling Wedge (upward breakout, bull market):

• Performance rank: 31/39
• Breakeven failure rate: 26%
• Average rise: 38%
• % meeting target: 62%

The bullish bias is real, but the magnitude is modest — mid-tier.

Rising Wedge (downward breakout, bull market):

• Performance rank: 36 of 36 — dead last
• Breakeven failure rate: 51%
• Average drop: 9%

Bulkowski on rising wedges:

Can you imagine a pattern where half of them (51%) will fail to see price drop more than 5%? That's alarming.

The "rising wedge is bearish" rule is correct directionally but useless practically — the edge is too small and the failure rate too high. Trade something else.

8. Cup and Handle — the elite pattern Bulkowski does cover

Popularized by William O'Neil; Bulkowski simplified O'Neil's strict identification rules and tested it. The pattern: a rounded U-shaped base (the "cup"), followed by a smaller pullback (the "handle"), then a breakout.

Stats (upward breakout, bull market):

• Performance rank: 3 of 39 — top tier
• Breakeven failure rate: 5% (rank 2 of 39 — fewest failures of any pattern)
• Average rise: 54%
• % meeting target: 61%

Bulkowski on the handle requirement:

A cup-with-handle without a handle is like peanut butter without jelly.

Interesting oddity: for most patterns, tall versions outperform short versions. For cups, it's the reverse — shorter cups outperform taller ones, per Bulkowski's data. No one knows why.

We haven't built a Cup-and-Handle interactive for this lesson — the shape needs too many bars to render crisply in our gallery. Recognize it on real charts: 7+ week U-shaped base, handle should pull back under 1/3 of the cup's depth.

The volume myth, again

Every chart-pattern textbook says breakout volume matters. Bulkowski tested it across every pattern. The verdict, as with reversal patterns, is underwhelming:

Heavy breakout volume? Light breakout volume? The cup-with-handle pattern showed no preference in my tests.

For descending triangles, gaps on breakout help (+48% rise with gap vs +36% without). For ascending triangles, gaps help a tiny bit — Bulkowski's words: "isn't large enough to twist our knickers." For symmetrical triangles and most others, volume is loosely correlated with performance but not predictive enough to act on.

The volume pattern inside the consolidation (volume drying up) is statistically real across most continuation patterns — over 80% of symmetrical triangles show declining volume through the pattern. But breakout volume filtering is mostly noise.

Quick check

What you now know

• "Continuation" is a bias, not a rule. Bulkowski's R-vs-C percentages are 55/45 to 65/35 at best; a descending triangle in a bull market resolves as a reversal 66% of the time.
• Symmetrical triangle: rank 36/39. The textbook's favorite is one of the worst performers. Performance is "well, awful" per Bulkowski. Two-thirds rule (Murphy) survives the data.
• Ascending triangle: rank 16/39, +43% rise, 70% hit target. Legitimate bullish edge for upside breakouts. Short-only in bear markets.
• Flag: 44% breakeven failure in bulls. Measure rule hits 46% of the time. Halve the flagpole projection.
• Pennant: "Does it work? No." 54% breakeven failure, 7% average rise. Don't trade it standalone.
• Rectangle Top: rank 4/39, +51% rise. Elite pattern — horizontal lines imply real S/R, not soft trendlines.
• Cup and Handle: rank 3/39, 5% failure. The most reliable continuation pattern Bulkowski measures.
• Rising wedge: 36/36 — dead last. Bearish direction is correct; magnitude is terrible. Falling wedge is better but mid-tier.
• Volume story is mostly lore. Breakout volume is weakly predictive at best. The only robust volume pattern is declining volume during the consolidation (80%+ of symmetrical triangles).

Next: Fibonacci Retracements (coming soon) — 38.2 / 50 / 61.8 levels, the math, the psychology, and Bulkowski's tests on whether price really reacts at them.