VWMA: A Rolling Volume-Weighted Average

Specify the window and price input

For a window of n observations, VWMA is the sum of price multiplied by volume over those observations, divided by their volume sum. TradingView documents VWMA and its configurable length and price source. The window can use closes or another declared price series; the selected input is part of the definition.

A three-bar VWMA is not a three-hour average unless every bar represents one hour. A daily chart with length 20 uses twenty trading observations, not necessarily twenty calendar days. Missing bars and session filters can change the represented interval. Record the chart timeframe alongside the length rather than quoting “20-period” as a complete setting.

A rolling example with unequal volume

Consider three hypothetical closes: 10 with volume 100, 12 with volume 300, and 11 with volume 100. The numerator is 1,000 + 3,600 + 1,100 = 5,700. Total volume is 500, giving VWMA of 11.40. The simple mean is 11.00, so the heavier activity at 12 pulls the weighted result upward.

A fourth close arrives at 10 with volume 100. A three-bar calculation removes the first observation and now uses 12, 11 and 10 with volumes 300, 100 and 100. Its value remains 11.40. Although the newest close is lower than the previous one, replacing an identical old price-volume observation leaves the weighted average unchanged.

Next, a fifth close is 10 with volume 100. The unusually heavy 12-price observation leaves the window. The result becomes (1,100 + 1,000 + 1,000) / 300 = 10.3333. The large movement in the indicator reflects the removal of old data as well as the new close. That distinction matters when interpreting sudden changes in a rolling line.

What weighting does and does not establish

Volume weighting assigns greater arithmetic influence to observations with more recorded activity. It does not identify whether buyers or sellers are informed, which participants initiated transactions, or whether activity is representative of a broader market. A large auction print, a futures roll or a feed anomaly can affect the line substantially.

A bar's close multiplied by its entire volume assumes that close represents all activity within the bar. It is therefore not the same as calculating an exact transaction-weighted mean. A broker's tick-count series creates another interpretation again. If the whole selected window has zero volume, there is no valid weighted denominator; silently substituting an arbitrary price disguises missing information.

Compare the right alternatives

Session VWAP keeps observations from a reset boundary, while anchored VWAP keeps them from a chosen event. VWMA forgets observations when they age out. When comparing them, hold the symbol, session, source price and volume definition constant so the comparison isolates the intended difference.

Before treating a crossover as a rule, export the window that caused it and identify whether new prices or dropped observations drove the change. Then write the rule and costs into a reproducible backtest. A line crossing does not by itself specify order timing, stop distance, position size or acceptable loss.