Why does a sharp move so often pause in a flag?
A flag is the tightest, shortest version of the pause you met in the last lesson: a steep run (traders call it the pole), then a short, narrow drift that leans slightly against the direction of the run. A pennant is the same thing where the drift converges into a small triangle instead of sloping.
The behaviour inside the cloth
Why does the drift lean against the move? Because the selling in a pause is mostly profit-taking, and profit-taking is mechanical rather than opinionated. Someone who bought at $20 and watched it hit $26 sells a slice — not because they've turned bearish, but because the position got large. That trickle of supply nudges the price gently lower even while nobody is actively negative.
Meanwhile the drift is narrow because buyers who missed the run are waiting underneath, and they don't have to wait long before the trickle reaches them. The two forces are small and roughly matched, so the range stays tight. On the volume pane you see the same signature every time: heavy on the pole, thin in the flag.
Duration matters to the definition. Most traders call a drift a flag when it lasts days to about three weeks. Stretch it to three months and the tight-band story stops holding — new information has arrived, new participants have formed opinions, and you're looking at a genuine range, not a pause.
A worked example
Illustrative numbers: a stock closes at $20.10, then over five sessions runs to $26.40 on volume roughly three times its 50-day average. That's the pole. Over the next eight sessions it drifts from $25.90 down to $24.80, with each session's range under 2%, and volume falls to about half average.
Draw two parallel lines around those eight sessions and you have a textbook flag. Describe it neutrally: a 31% repricing, followed by 6% of orderly giving-back on light participation.
The honest problem with flags
Flags are the most-drawn pattern in technical analysis, and the reason is uncomfortable: the pole is objective, the flag is not. You can measure a 31% move in five sessions and its volume — anyone would get the same numbers. But where the flag's boundaries sit, how many sessions count, how much counter-drift is "too much" — you chose all of that. On a noisy chart, a large fraction of ordinary pullbacks can be framed as flags if you're willing to move the lines.
That's the tell of a weak pattern: the parts you can measure are the parts everyone agrees on, and the parts that carry the alleged meaning are the parts you drew. Unit 4 turns this into an actual test.
Try it now
- Measure the pole on the schematic above: its % size and how many sessions it took. Anyone reading the same bars gets your numbers. Now move the two boundary lines a little — inward, outward, one session longer — and notice that the pole does not change and the flag does. That asymmetry is the lesson.
- Now find a real pole: any move of roughly 15%+ inside two weeks in the year below. Note its % size, and the volume multiple against its recent average: the line over the volume bars under the chart is the 50-session average. Then look at the 5–15 sessions after it and write the drift's high, low and width in %.
- Failure hunt: find a pole where the "flag" gave back the entire move within a month. Then ask yourself honestly — before the giving-back, would you have drawn that flag exactly the same way?