How is a poor high different from a double top?

A double top is a bar chart pattern: two swing highs at a similar price with a dip between them. A poor high is a statement about profile structure — repeated TPOs at the extreme with no excess above them. They often appear together, but one is drawn from highs and lows, the other from how the market distributed time at price.

How many TPOs make a high poor?

The common convention is at least two periods printing the same extreme with no single prints above them; some readers also require those periods to be non-consecutive. There is no authority to appeal to, so fix your version, write it down, and apply it identically to every session you log.

Can a poor high repair itself?

Yes. A later session that probes above the flat top and gets rejected leaves the excess that was missing, finishing the auction higher. The original extreme then stops being a poor high — one reason retrospective counts of this pattern are so easy to get wrong.

Glossary

Poor High / Poor Low

Also called: poor low · unfinished auction · flat top · flat bottom

A poor high is a profile extreme made of two or more equal prints with no excess above it. The auction stopped without leaving a rejection tail, so the high is unfinished and often revisited. Its mirror image is a poor low. Both are judged only once the session has closed.

Excess is what a finished auction leaves

When an auction genuinely runs out of room it leaves a mark: price probes beyond value, is rejected fast, and no later period returns. The result is a tail of single prints at the extreme — evidence that the market tested a price and refused it.

A poor high has no tail. Two or more periods sit flat on the same top price and then the market goes elsewhere. Nothing was rejected; the buying simply stopped in the middle of what it was doing. The extreme records an interruption rather than a decision.

Why it invites a revisit

An unfinished auction leaves unanswered the question the market usually settles: how high is too high. Until that is answered the price stays a candidate. This is the structural argument, and the reasonable half of the concept.

The mechanical half is more concrete. A flat top is a visible number that everyone reading a market profile can see, so buy stops from shorts sit just above it and protective stops from longs just beneath. When price returns, the orders firing there are part of what makes the return look decisive — trapped traders on both sides of three ticks.

The number to be suspicious of

The figure circulating in order flow teaching is that the reading only holds on roughly 20% of trending days. It may well be right. It has also never been replicated on a public dataset, the people quoting it do not share a definition, and the sessions where it failed appear in nobody's slide deck.

Make it falsifiable instead. Define the pattern mechanically — at least two periods printing the extreme, no more than one tick of excess above them — then bound revisited in sessions and ticks, and count both outcomes on your own instrument. A definition you can test is worth more than a percentage you cannot.

A worked example

In a synthetic ES session, the last two 30-minute periods both print up to 5 344.50 and no higher, leaving two letters on each of the top two prices. There is no excess: the profile ends flat. The previous session of the same synthetic series ended the other way — one letter at 5 351.25 with three ticks of single prints beneath it, a clean rejection tail.

Three sessions later, trade returns and prints 5 345.25, three ticks above the flat top, then leaves without settling there. Under a definition written in advance — revisited within five sessions by at least one tick — that counts. Under a definition demanding acceptance above the level, it does not. Which definition you wrote decides what you just saw.

The trap

The poor high is judged at the close and traded during the session. Those are not the same object. Mid-afternoon every high is flat, because the tail that would disqualify it has not printed yet. Selling into a poor high before the close is selling into a shape that may not exist by the time it can be named.

The second trap is that a revisit is not a trade. The market can come back three sessions later, having gone twenty points the other way first, and the reading was technically correct throughout. Without a price that ends the idea and a bound on how long you will wait, the concept can never be wrong — and a concept that can never be wrong teaches you nothing.

Frequently asked

How is a poor high different from a double top?: A double top is a bar chart pattern: two swing highs at a similar price with a dip between them. A poor high is a statement about profile structure — repeated TPOs at the extreme with no excess above them. They often appear together, but one is drawn from highs and lows, the other from how the market distributed time at price.
How many TPOs make a high poor?: The common convention is at least two periods printing the same extreme with no single prints above them; some readers also require those periods to be non-consecutive. There is no authority to appeal to, so fix your version, write it down, and apply it identically to every session you log.
Can a poor high repair itself?: Yes. A later session that probes above the flat top and gets rejected leaves the excess that was missing, finishing the auction higher. The original extreme then stops being a poor high — one reason retrospective counts of this pattern are so easy to get wrong.