Basics · Volume
Why Volume Matters, and What It Cannot Say
Volume counts transfers, and every share bought was sold. That symmetry rules out most of what is claimed for volume, and what remains is the most dependable material in this reference.
The symmetry
A trade has two sides. If a hundred thousand shares changed hands, a hundred thousand were bought and the same hundred thousand were sold, not two separate quantities but one, counted once.
It follows that there is no such thing as more buying than selling in completed trades, and any sentence of that form is describing something else. What can be true is that buyers were more eager: that shows up as the price moving up, because eagerness is expressed in price rather than in quantity. Volume records how much stock the move required.
That sounds like a technicality and it disposes of a large share of published volume commentary. Once the symmetry is held firmly, "volume shows accumulation" and "buying pressure was heavy" become claims that need a mechanism, and the honest measures are the ones that do not need one.
What "up volume" actually is
Several well-known measures give volume a sign. None of them measures a direction; each imposes one by a rule, and the rules disagree.
The first row of that figure is the clearest case. A session that gapped down heavily and then recovered to close near its own high counts as strong accumulation under one rule and as a decline under the other, and a holder is materially worse off than the day before, which neither rule records.
What volume can support
How much stock a move required. The same two per cent rise on three times the usual volume and on half of it are different events: in the first, a great deal of stock changed hands at those prices, and in the second, very little did. That makes the price move more or less consequential without saying anything about what comes next.
Whether the instrument is unusually busy for itself. This is the workhorse comparison of the whole site, today against the instrument’s own recent average, and it is the only form in which a volume figure is comparable between instruments or across time.
Whether activity was forced. When holders sell because they must rather than because they want to, volume expands unmistakably and every breadth measure records it at once. That is the most reliable pattern in this field, it explains why extreme readings cluster at lows rather than at highs, and the volume-at-turns page works through it.
Why a raw share count is not enough
| Question | The wrong form | The form that works |
|---|---|---|
| Was today busy? | "Two million shares", meaningless without knowing the instrument. | Today divided by the mean of the previous twenty sessions. A number around 1 that means the same thing everywhere. |
| Which instrument is busier? | Comparing share counts, largely a comparison of price levels. | Dollar volume for money, or turnover against free float for how much of the company traded. |
| Is this month heavy? | A monthly total, which belongs to its two or three biggest sessions. | The median session, or the total with the busiest sessions excluded. |
| Is this half-hour heavy? | Comparing it with a flat daily average, which measures the clock. | The same half-hour across recent sessions, the session-shape adjustment. |
Every row is the same idea: a volume figure means something only next to another volume figure. That is why this reference is organised around comparisons rather than levels, and why the pages that survive scrutiny best are the ones that count something and then divide it by something.
What the figure does not contain
Two omissions are worth carrying, because both are invisible in the number and both change what it means.
Venue. A single-exchange figure is a fraction of a security’s total trading, and the exchange a company is listed on is frequently a minority of its own volume. The basics page sets out the composition.
Composition. Around half of consolidated volume in American equities carries a short marking, which sounds alarming and means much less than it appears to. Most of it is market making rather than a directional bet. The short-volume study measures it from FINRA’s published daily files and states what the number can and cannot support, which is the closest thing on this site to a worked example of everything above.
Frequently asked questions
What does a volume figure count?
Shares that changed hands. If a hundred thousand shares traded, a hundred thousand were bought and the same hundred thousand were sold. They are the same shares in the same transactions, counted once. That is the whole definition, and it is the source of both the measure’s value and the most common mistake made with it.
So volume cannot show buying pressure?
Not directly, no. There is no such thing as more buying than selling in a completed trade; the two are the same event seen from two sides. What can be true is that buyers were more eager, which shows up in the price moving up, not in the volume being large. Volume tells you how much stock it took to produce whatever the price did, and that is a genuinely useful and quite different statement.
What about up volume and down volume?
Those are classifications imposed afterwards, not measurements. The whole of a session’s volume in an issue is assigned to one direction, decided by where the close sat relative to something, and there is more than one candidate for that something. The figure on this page runs two standard conventions over the same four sessions and they disagree about half of them, which is what a convention rather than a measurement looks like.
Then what is volume good for?
Three things, all comparative. It says how much stock a price move required, which makes the same move more or less notable. It says whether an instrument is unusually busy for itself, which is the relative-volume comparison this site uses everywhere. And it identifies forced activity: when holders sell because they must, volume expands unmistakably, which is the most reliable pattern in the whole field.
Why is a raw share count almost useless on its own?
Because it has no scale. Two million shares is enormous for one instrument and a quiet morning for another, and the same instrument’s volume is not comparable with its own figure from a decade ago, when its float and shareholder base were different. Every volume statement worth making is a ratio — against the instrument’s own recent average, most often — and the baseline page sets out how to build one.
Is heavy volume bullish?
Neither bullish nor bearish, and the question contains the error. Volume is symmetric, so it has no direction of its own; it makes whatever the price did more consequential, in either direction. A heavy session with a large price rise and a heavy session with a large fall are both sessions in which a great deal of stock changed hands, and the sign comes entirely from the price.
Does the volume figure include everything?
It includes what is reported to the consolidated tape, which in the United States covers exchange and off-exchange executions, but a figure from a single venue is a fraction of that, and the fraction differs by instrument. There are also reporting conventions and corrections, and around half of consolidated volume carries a short marking that means far less than it appears to. The short-volume study measures that from the published files.
Where should someone start with volume?
With a twenty-session average and a ratio. Put the volume panel under the price, divide each session by the average of the previous twenty, and read the result as a multiple: about 1 is ordinary, above 2 is genuinely busy, above 5 means something specific happened and is usually identifiable from the calendar. That single habit accounts for most of the practical value of volume analysis, and it requires no indicator.