Heavy volume that fails to move price is one of the most informative events on a chart. It is also the one most frequently misread as strength, because the volume histogram shows a tall bar and the tall bar looks like conviction.
It is often the opposite. Someone was buying, and someone larger was selling them everything they wanted at the same price.
Two ways to trade
Every trade has two sides, and they are not symmetrical. One participant crosses the spread to get filled now — the aggressor. The other was already resting in the book, waiting to be hit — the passive side. Volume counts both, which is why volume alone tells you so little.
What matters is the relationship between aggression and movement. Aggressive buying that lifts price is ordinary: demand met thin supply. Aggressive buying that does not lift price means the supply was not thin. Something absorbed it.
Why size hides in the passive side
A participant who needs to move a genuinely large position has a problem: crossing the spread repeatedly announces them and moves the price against their own remaining size. So they do the opposite. They rest orders and let the market come to them, replenishing as they are filled.
This is why the largest participant in a move is frequently invisible on the tape as an initiator. They are not initiating. They are the wall the initiators are running into — and the only evidence they exist is that a great deal of aggression produced almost no price change.
Distinguishing absorption from a pause
Not every flat patch is absorption. Price stalls constantly for the boring reason that nobody is doing anything. Three conditions separate the two.
Participation must be elevated
Absorption requires volume meaningfully above the recent norm. Measure it against a rolling baseline of comparable bars, not a fixed threshold — a thousand contracts is enormous at 3am and unremarkable at the open. A tool that uses a fixed number is wrong for most of the session.
The range must be compressed relative to that participation
The signal is the ratio, not either number alone. High volume with a wide range is a normal trend bar. High volume with a narrow range is the anomaly worth naming.
The aggression must be one-sided
If buying and selling are balanced, a narrow range is simply two-way trade. Absorption means the aggression was directional and still failed to produce movement. This is where you need a genuine buy/sell split — not the total volume, and not an inference from whether the candle closed green.
That last point is where most retail tooling quietly fails. Splitting a bar's volume by its close direction is a guess, and on the bars that matter — the ones with heavy two-way trade — it is a bad guess. A bar can close green having absorbed enormous buying.
What it implies, and what it does not
Absorption tells you a large passive participant defended a price. It does not tell you they will win. Walls are pulled. Participants change their minds, get run over, or were hedging something you cannot see.
What it does give you is a located level — a specific price at which someone with size showed their hand. The follow-up is what carries information:
- Price rejects away from the level and the absorbing side is validated.
- Price grinds through it, and the absorber has been overwhelmed — often violently, because the size that was defending is now trapped.
- Price returns later and the level is defended a second time, which is considerably stronger evidence than the first.
The third case is the one worth waiting for. A level defended once is a data point. A level defended twice, with participation both times, is a participant with a mandate.
Feed quality decides whether you can see any of this
All of the above depends on knowing which side was the aggressor. That requires a feed reporting genuine buy and sell volume. On a feed that supplies only total volume, absorption cannot be measured — only guessed at.
This is worth being blunt about, because the industry generally is not: a tool that shows absorption readings on a feed that cannot supply the data is showing you an inference dressed as a measurement. Ours detect the absence and fall back to structure rather than fabricating the number. Boring, and correct.