Flags and pennants: trading the pause after a sharp move

A flag is a short pause after a sharp move. Price shoots up (or down) almost in a straight line, then drifts sideways or slightly against the move for a few days to a few weeks, in a small, tidy channel. On a chart the steep move looks like a flagpole and the pause like a flag hanging from it. A pennant is the same idea with a small converging triangle instead of a parallel channel.

Both are classed as continuation patterns. The expectation is that after the pause, the move resumes in the direction of the pole. The traditional saying is that flags "fly at half mast": the pause comes roughly halfway through the move, so the second leg is about as long as the first.

This guide covers what makes a valid pole and flag, the difference between flags and pennants, how to measure the target, and three real examples on US stocks. It ends with a test of 66 flags our rules found on 60 large US stocks and ETFs, which gave a much less flattering picture than the textbooks.

Anatomy of a flag and a pennant

The pole

The pole is what makes a flag. It is a nearly vertical move, often triggered by news such as an earnings report, an upgrade or a sector-wide rally, with volume well above normal. On daily charts, practitioners usually look for at least 10% to 15% in a few sessions. Our test required at least 12% in no more than 12 sessions, close to close.

The flag

After the pole, trading calms down. A bull flag drifts slightly lower or sideways in a narrow channel with parallel boundaries. A bear flag drifts slightly higher or sideways after a sharp drop. Three features separate a flag from an ordinary pullback:

• It is short. A few days to a few weeks on a daily chart. A pause that lasts months is a different pattern, such as a triangle or a rectangle.
• It is shallow. It retraces well under half the pole. A deeper retracement suggests the move has lost its force.
• Volume dries up. In the classical description, volume falls during the flag and expands again on the breakout. Heavy volume inside the flag means real selling, which is a warning.

Pennants

A pennant replaces the parallel channel with converging lines, a miniature symmetrical triangle. Everything else is the same: a steep pole, a brief pause, low volume, and a breakout in the direction of the pole. In practice the two are often grouped together, and our test does not try to tell them apart.

High and tight flags

Bulkowski and William O'Neil both describe a rare variant, the high and tight flag, in which the pole is an extreme rise, roughly a doubling within about two months, followed by a shallow flag. It is associated with momentum stocks and does not appear in our two-year sample of large companies.

The breakout and the target

The signal is a close beyond the flag in the direction of the pole: above the top of a bull flag, below the bottom of a bear flag. The failure point is the far side of the flag: a close below a bull flag's low, or above a bear flag's high.

Some traders measure the target from the flag's low instead of the breakout, which gives a slightly closer objective. Others take partial profits at half the pole and let the rest run. Neither is "correct". What matters is deciding before the trade.

Three real flags

Alphabet, 2025: a bull flag that worked, slowly. GOOGL rose from $244.15 on October 21, 2025 to $291.59 on October 30, a 19% pole in seven sessions. It then paused for six sessions between $275.19 and $288.35. On November 10 it closed at $290.10, above the flag. The target, one pole length above the breakout, was about $337.54. The path was rough. On November 14 the stock traded down to $270.70 intraday, below the flag's low, but it closed above it. The target traded on January 13, 2026, 43 sessions later.

Broadcom, 2024–2025: a powerful pole, a failed flag. AVGO jumped 24% on December 13, 2024, the session after its quarterly earnings report, and kept going: from $175.99 to $251.88 in three sessions. It then pulled back into a flag between $217.58 and $242.47, and on December 26 it closed at $245.36, above the flag. The measured target was an ambitious $321, 31% higher. Instead, the stock drifted, and on January 27, 2025, a day of heavy selling across AI-related chip stocks, it fell 17% and closed far below the flag's low. The bigger the pole, the more ambitious the target, and the further price can fall back.

Oracle, 2025: a bear flag. ORCL fell from $322.54 on October 16, 2025 to $269.25 on October 22, a 16.5% drop in four sessions. It bounced into a small flag between $271.35 and $287.19, then broke down on October 30 with a close at $256.89. The pole projected a target near $203.60, 21% below the breakdown, and the stock reached it on November 21, sixteen sessions later.

How often did flags work? A test on 60 large US stocks

We scanned two years of daily bars (October 30, 2024 to October 29, 2026) for the same 60 widely held US stocks and ETFs used in our other pattern tests.

The rules. A pole was a close-to-close move of at least 12% in 3 to 12 sessions, ending at a 5-session high (or low, for bear flags). The flag was the next 5 to 25 sessions, as long as its range stayed under half the pole, it gave back less than half the pole, and it did not exceed the pole's extreme. The breakout was the first close beyond the flag's high (bull) or low (bear). A hit meant the pole-length target traded before any close beyond the flag's opposite side (a failure), within 80 sessions.

The benchmark. For a driftless random walk, the chance of reaching a target before a stop is the stop distance divided by the sum of the two distances. Because flags have tight stops and distant targets, this benchmark is low, under 30%.

Most flags failed before reaching their target. Two-thirds of resolved bull flags and four-fifths of bear flags closed beyond the flag's far side first. Bull flags did slightly better than the benchmark and bear flags worse, a pattern consistent with a rising market during the test period.

Half the move was common. Price covered at least half the distance to the target in 20 of the 38 bull flags, before any failure. That supports the practice of taking partial profits well short of the full measured move.

Common mistakes

• Calling any pullback a flag. Without a steep pole, a pause is just a pause.
• Ignoring depth and duration. A "flag" that gives back 60% of the pole over two months is a failed move, not a continuation.
• Chasing the pole. Buying into the vertical move itself means taking the full risk of the first pullback. The flag is the pattern's lower-risk entry.
• Holding for the full pole no matter what. As the study shows, the full target is reached less often than the tidy diagrams suggest.
• Forgetting event risk. Many poles start with earnings. The next scheduled event, visible on the earnings calendar, can end the flag either way.

Flags compared with similar patterns

How Finabase helps you spot flags

Finabase does not label flags automatically, but its charts and tools make them easy to find:

• The Scanner. The Scanner lists stocks with large gaps and new 52-week highs or lows on the latest session. Those are where poles start. Check them again a week or two later for a tidy pause.
• Drawing tools. On any symbol's chart, use the trend line tool for the flag's two boundaries, or the rectangle tool to box a flat flag. Signed-in users' drawings are saved per symbol.
• Volume pane. Volume shows under price by default, so you can check that it surged on the pole and dried up in the flag.
• ZigZag study. A 5% ZigZag, as on this two-year ORCL chart, makes the poles stand out as long single legs. See the ZigZag indicator: filtering swings, and why the last leg repaints guide.
• Momentum context. Indicators such as Momentum indicator (MOM): price change over n bars and the Relative Strength Index (RSI): how to read and use it show how stretched the pole was. A pole that drives the RSI far into overbought territory leaves less room for an immediate second leg.

Frequently asked questions

What is the difference between a bull flag and a bear flag?

A bull flag forms after a sharp rise and usually drifts slightly down or sideways. It breaks out upward. A bear flag forms after a sharp drop, drifts slightly up or sideways, and breaks down. The rules are mirror images.

How long can a flag last?

On daily charts, most practitioners look for one to five weeks. Our test allowed 5 to 25 sessions. A longer pause usually belongs to another pattern category.

Do flags work on intraday charts?

Traders use them on every timeframe, from one-minute to weekly charts. The logic is the same, but intraday flags are smaller, more numerous and more affected by noise and trading costs.

Is a pennant more reliable than a flag?

There is no consistent evidence either way in the sources we know of, and our test did not separate them. Treat them as the same pattern with a slightly different shape.

Summary

A flag or pennant is a breather after a burst: a steep pole, a short and shallow pause on fading volume, and a breakout in the direction of the pole. Its appeal is the tight stop just beyond the flag and the large measured target, one pole length. Our two-year test on large US stocks showed the catch: the full target was reached in only about a third of bull flags and a fifth of bear flags before the pattern failed, though about half went at least halfway. Use flags to find low-risk entries after strong moves, keep expectations for the second leg modest, and respect the flag's far side as the point where the idea is wrong.