The shape, and the honest name for it
A flag is a sharp directional move — the pole — followed by a short, shallow drift against that move, held between two roughly parallel boundaries. A pennant is the same thing with the boundaries converging instead of parallel, so the pause narrows to a point rather than sloping.
That is the entire geometric difference between the two names, and it is worth saying immediately so you can stop worrying about it. Parallel is a flag; converging is a pennant. Nothing else distinguishes them.
Now the part the books lead with and this page will not. A flag is described as a continuation pattern, meaning the move resumes. But look at what is actually on the chart before that happens: a sharp rise, and a small drift downward. That is a pause. It is called a flag once the rise resumes, and if the drift keeps going it was never a flag at all — it was the start of the decline.
A continuation pattern is a pause that was followed by a continuation. Before the continuation, you have a pause.
Eleven sessions up, six sessions sideways
The dashed line is ₹72, the top of the pole and this example's stated test. The fainter line is the drift's lower edge.
Six sessions of drift and the price is ₹6 below the high. Whether that is a flag or the top is not written anywhere on this chart.
What can be said before the resumption
Quite a lot, and all of it checkable. The prior move is real: 50% in eleven sessions, which satisfies any reasonable reading of 07.01.02's requirement. The drift is genuinely shallow — ₹6 given back against ₹24 gained, a quarter of the pole. It has lasted six sessions against the pole's eleven, so it is short relative to the move it interrupts, which is what distinguishes this from an ordinary decline.
Those three sentences are a description somebody else can verify. They are also, precisely, 07.01.03's criteria applied to a different shape: the proportions, the duration, and — next — the volume.
What cannot be said is that the rise resumes. The stated test for this example is a completed daily close above ₹72, fixed now, before the next session exists. If it comes, the pattern definition is satisfied. If price closes below the drift's lower edge instead, the pause was the beginning of something else and no amount of the shape being tidy will have prevented it.
The volume claim, which is the checkable one
The textbook signature is specific: volume is heavy on the pole, dries up through the flag, and expands on the resumption. The reading offered is that the pause is drift rather than distribution — the sellers who would end the move are not there, they are simply not buying either.
Unlike most of this chapter, that is a falsifiable claim about numbers that exist, and 07.01.03 already made the point that it is the one criterion nobody checks. Here it would look like this: the pole's sessions averaging 2,80,000 shares, the six drift sessions averaging 90,000, and whatever the resumption session brings.
Two cautions, both ordinary. Thin volume in a pause is also what a holiday week looks like, and what a stock nobody is interested in looks like — relative to comparable periods is doing all the work, as 04.05.02 said. And the signature being present tells you the example is typical of its class. It does not tell you the class works, which is 07.01.04's department.
- The pole
- ₹48 to ₹72 in eleven sessions — a rise of 50%
- The drift
- ₹72 to ₹66 in six sessions — a quarter of the pole given back
- Drift length against pole length
- six sessions against eleven
- Pole sessions, average volume
- 2,80,000 shares
- Drift sessions, average volume
- 90,000 shares
- Stated test, fixed in advance
- a completed daily close above ₹72
- What the next session did
- not on this chart
Half-mast, and why this page will not use it
The most repeated sentence about flags is that they fly at half-mast — that the move after the pause runs about as far as the move before it, so the pole's height projected from the breakout gives a target. Here that would be ₹24 added to ₹72, giving ₹96.
The arithmetic is easy and the claim underneath it is not measured. 07.01.04 set out the specific problem: when the pattern defines its own target, a taller pattern is graded against a more distant one, and two patterns reported with the same success rate succeeded at different things. Half-mast is exactly that construction.
So the number is written above as arithmetic and is not offered as an expectation. Granthara publishes no target and no success rate for this or any chart pattern, which 07.02.03 promised in as many words.