Indicators · Volume
Accumulation and Distribution Line
A running total of volume weighted by where each session closed inside its own range. It never looks at the previous bar, which is both what makes it different from on-balance volume and where it goes wrong.
What it is calculated from
Each session contributes its volume, scaled by where the close sat within the high-low range:
multiplier = ((close − low) − (high − close)) ÷ (high − low)
A/D = previous A/D + (multiplier × volume)
The multiplier runs from −1, for a close exactly at the low, to +1 at the high, and is zero for a close at the midpoint. So a session with enormous volume closing mid-range contributes nothing at all, which is either the measure’s cleverest feature or its strangest, depending on what you think a mid-range close means.
That disagreement rate is the reason to know which measure a chart is showing. Two indicators described in the same language, computed from the same bars, differ about the sign of a large minority of sessions, because they are asking different questions.
Two questions about one bar
| Measure | The question | Blind to |
|---|---|---|
| Accumulation/distribution | Where in its own range did this session close? | The previous session entirely, so a heavy gap down that recovers into its own high counts as strong accumulation. |
| On-balance volume | Did this session close above the previous one? | Everything inside the bar, so a session that fell all day and closed a cent up assigns all its volume to the up side. |
| Volume price trend | By how much did it close above the previous one? | The same inner structure, but at least the size of the move is weighted rather than ignored. |
| Money flow index | Where did it trade on average, weighted by volume? | Nothing structurally, and it is bounded, so it cannot be read as a total at all. |
Reading the first two rows together is the practical use of this family. Their blind spots are complementary: one cannot see gaps, the other cannot see the inside of a bar. When both lines agree, the volume reading is unambiguous; when they disagree, the reason is always identifiable in the bars, and it is usually a gap or a run of mid-range closes.
Where the measure fails
Gaps, decisively. The multiplier is computed entirely within the session, so an instrument that gaps down five per cent and then closes near the high of its new range contributes strongly positively, while a holder is materially worse off. On a gap-prone instrument, this line can rise through a sustained decline without any error in the arithmetic.
The undefined bar. When high equals low (a halted session, or an extremely thin one that printed a single price) the denominator is zero. Implementations treat the multiplier as zero, which is the sensible choice and worth knowing, because a series of such bars quietly contributes nothing while the price may be moving in steps between them.
The interpretation. The names carry a claim: accumulation means large buyers taking stock, distribution means them letting it go. Nothing in the formula can see who traded. Every share bought was sold, and a close near the high tells you where the session ended, not who was on which side of it. This is the same objection the up/down volume page makes about assigning a direction to volume: the classification is a convention, and it is useful precisely as long as it is not mistaken for a measurement.
What is built on top of it
The line is rarely read raw. Its best-known derivative takes the difference between a short and a long exponential average of it, the same construction MACD applies to price, which turns a cumulative total with a meaningless level into an oscillator around zero.
That transformation solves a real problem and introduces the usual one. It solves the level problem: an oscillator has no dependence on when the accumulation began, so two instruments can be compared. It introduces the smoothing problem: two averages mean two periods, two seeds and two smoothing constants, each of which is a choice, and the resulting series inherits everything the moving average page says about why two platforms disagree.
The reading attached to it is the standard one for any zero-line oscillator (above zero is accumulation, below is distribution, crossings are events), and it carries the standard caution. A crossing is a statement that a short average has passed a long one, which is true by construction and is not a forecast. If you use the oscillator form, recompute its thresholds as percentiles of the instrument's own history rather than borrowing levels, because the scale depends on the instrument's volume as well as on its price.
Reading it honestly
Direction and divergence, and nothing else. A line that keeps rising while price stalls says the closes have been holding near the highs of their ranges on decent volume, which is a real description of recent sessions. A line that falls while price rises says the opposite: the advance has been made by sessions that closed poorly within their own ranges.
Both statements are about the past, both are checkable against the bars, and neither carries a date. That is the standard this reference holds every measure to, and this one meets it as long as the names on it are read as labels rather than as findings, which is why the library groups it with the other cumulative volume lines and recommends reading two of them rather than trusting one.
Frequently asked questions
What does the accumulation/distribution line measure?
Where each session closed within its own high-low range, weighted by that session’s volume, accumulated into a running total. A close near the high contributes positively, a close near the low negatively, and a close in the middle contributes almost nothing however large the volume. The premise is that the position of the close inside the range says something about who won the session.
How is it different from on-balance volume?
On-balance volume asks a question about the previous session — did the close finish above or below it — and then assigns the entire volume to that direction. This measure asks a question about the current session only: where in its own range did it close. The two can disagree about the same bar, and on the synthetic series on this page they disagree about the sign of a substantial minority of sessions.
Why do two indicators share the name?
Because two different people built one. The version computed here, using the close’s position within the high-low range, is Chaikin’s. An earlier and simpler measure by Larry Williams compares the close with the previous close and divides by the range, which gives different values and a different line. Charting packages label both "accumulation/distribution", so the first thing to establish about any such chart is which formula produced it.
What is the formula?
The multiplier is ((close − low) − (high − close)) ÷ (high − low), which runs from −1 for a close exactly at the low to +1 for a close exactly at the high. Multiply that by the session’s volume and add it to the running total. When high equals low the multiplier is undefined and implementations treat it as zero, which is a small detail that matters on halted or extremely thin sessions.
Does the level of the line mean anything?
No, for the same reason no cumulative line’s level means anything: it depends entirely on when the accumulation started. Only the direction and the divergences against price can be read. Anyone quoting a level is quoting an arbitrary start date, and comparing the levels of two instruments is comparing two arbitrary start dates.
Is the premise sound?
Partly, and it is worth being precise about which part. That a close near the high means buyers held the session is a reasonable reading of one bar. That accumulating this quantity produces a measure of institutional buying is a much larger claim, and nothing in the calculation can distinguish a buyer from a seller; every share bought was sold. The line is a summary of where closes sat, weighted by volume, and that is the honest description of it.
What is the most common practical failure?
Gaps. The measure never looks outside the current bar, so a session that gaps down heavily and then recovers to close near its own high contributes strongly positively, despite the instrument being far below where it was. On a gap-prone instrument the line can rise through a decline. On-balance volume has the opposite blind spot, which is why the two are worth reading together rather than choosing between.
Is the money flow index related?
It is a cousin. The money flow index also weights price by volume, but it uses the typical price rather than the position of the close in the range, and it puts the result on a bounded scale rather than accumulating it. If you want a bounded version of this idea, that is the one to use, and its own page sets out where it misleads.