Volume weighted average price (VWAP)

An average in which each price is weighted by the quantity traded at it. Multiplying price by quantity for every trade, adding those products and dividing by the total quantity gives the average price the market actually paid, which is a different figure from the simple average of the same prices whenever the heavy trading happened away from the middle of the range. On a chart it is normally computed from the session open and reset at the start of each session, and an anchored variant starts the calculation from a chosen bar instead.

Two properties follow from the arithmetic rather than from any theory. It is cumulative, so each new bar is one contribution among all those already counted, and the line therefore moves less and less as a session progresses. And it is a description of what has traded, not a projection: the figure at any moment covers the period behind it only. Charting implementations commonly weight the average of a bar's high, low and close rather than every individual trade, which is an approximation adopted because bar data is what a chart holds.

Its main institutional use is as an execution benchmark: an order worked through a day is compared with the day's figure to judge whether it was filled better or worse than the market's own average, and algorithms exist to slice an order so that it tracks that benchmark. Three limits bound the measure. On an instrument whose volume is a tick count rather than a traded size, the result is weighted by price updates and is not a volume weighted average at all, and two platforms will draw different lines for the same instrument. An order large enough to be a meaningful share of the day's trading moves the benchmark it is being measured against. And practitioners disagree about the benchmark itself, since measuring execution against the crowd's average rewards trading in line with the crowd, which is not the same thing as trading well.