Trend indicators

VWAP Indicator: How the Volume Weighted Average Price Is Calculated, Why It Resets Each Session, and What It Did on Hourly Bitcoin

The VWAP formula from StockCharts' documentation, a worked example, what a daily VWAP really shows, and how hourly Bitcoin moved around its session VWAP over 25 days in October 2026.

· Founder of PatternsWizard
Checked 9 min read
Session VWAP, ±1 standard deviation on Bitcoin, hourly, October 26, 2026, 12:00 UTC to October 28, 2026, 23:00 UTC. Illustration of how the indicator reads, not a backtest.
On this page
  1. How VWAP is calculated
  2. VWAP on daily bars: what you actually get
  3. How traders read it
  4. What it did on 25 sessions of hourly Bitcoin
  5. The five sessions on the chart
  6. Where VWAP comes from
  7. Does it work?
  8. Frequently asked questions

The Volume Weighted Average Price (VWAP) is the average price of a trading session, with each trade counted in proportion to its volume. It starts at the open, adds every bar's price times its volume, and divides by the volume traded so far, so at any moment it tells you the average price paid by everyone who traded that day. StockCharts writes that institutional investors developed it to place large orders without disrupting the market, and that it is also used to measure trading efficiency: whether an order was filled at a good price. Because it resets each session, it only makes sense on intraday bars. Below we compute it on hourly Bitcoin, reset at midnight UTC each day, and check what it did over 25 sessions.

Session VWAP on hourly Bitcoin, October 24 to 28, 2026

Session VWAP, ±1 standard deviation on Bitcoin, hourly, October 24, 2026, 00:00 UTC to October 28, 2026, 23:00 UTC. Illustration of how the indicator reads, not a backtest.
  • Price, hourly candles (hollow up, filled down)
  • VWAP, reset at 00:00 UTC each day (solid line)
  • ±1 volume-weighted standard deviation (dashed line)
  • Shaded area between the bands
Session VWAP, ±1 standard deviation on Bitcoin, hourly, October 24, 2026, 00:00 UTC to October 28, 2026, 23:00 UTC. Illustration of how the indicator reads, not a backtest.

How VWAP is calculated

For each bar, take the typical price, the average of the high, low and close, and multiply it by the bar's volume. Keep a running total of those products and a running total of volume from the first bar of the session. VWAP is the first total divided by the second. The bands on our chart are one volume-weighted standard deviation of the typical price around the VWAP, also reset each session.

VWAP on daily bars: what you actually get

StockCharts states that VWAP is not defined for daily, weekly or monthly periods. With one bar per session, the running totals hold a single bar, so the "VWAP" is that bar's typical price. On our daily Bitcoin series the two are identical: 81,187.88 on October 28, 2026 for both. A platform that shows a VWAP line on a daily chart is either doing that, or anchoring the sum to some other start, which is an anchored VWAP. A multi-day average weighted by volume, without a reset, is a volume weighted moving average.

Session VWAP restarts every day; anchored VWAP does not

Session VWAP restarts every day; anchored VWAP does not: a price path over three trading days, Monday, Tuesday and Wednesday, separated by dashed lines. A dashed session VWAP starts again at the price of the first bar of each day. A solid anchored VWAP starts at a chosen low on Monday and runs on through Tuesday and Wednesday without a reset.
Schematic, not market data. Drawn by hand to show the reset, not computed.The session VWAP restarts at the first bar of each day; on Bitcoin, our chart restarts it at midnight UTC. An anchored VWAP keeps one running total from a chosen bar.

How traders read it

ReadingWhat it is used for
Buying below VWAP, selling above itA good fill for an order placed during the day, the trading efficiency use StockCharts describes
Price holding above VWAPBuyers in control of the session; pullbacks to the line watched as entries
Price holding below VWAPSellers in control; rallies to the line watched as short entries
Price far outside the bandsA stretched move, either strong momentum or a candidate for a return to the line

StockCharts adds that VWAP lags more as the session goes on, because each new bar is a smaller share of the total. Our 25 sessions of hourly Bitcoin show it: over the three hourly bars after the reset (01:00 to 03:00 UTC), VWAP moved 0.10% per hour on average, 42% of the average move of the close; on the bars from 19:00 to 23:00 UTC it moved about 0.02% per hour, 11% to 12% of the move of the close. The US stock market open, in the 13:00 to 15:00 UTC block, is the one time the line picks up again, because hourly closes moved twice as much there.

Average hourly move of VWAP and of the close, by time of day

Average hourly move of VWAP and of the close, by time of day. 01:00 to 03:00 UTC: Close 0.24%, VWAP 0.10%; 04:00 to 06:00 UTC: Close 0.20%, VWAP 0.04%; 07:00 to 09:00 UTC: Close 0.18%, VWAP 0.04%; 10:00 to 12:00 UTC: Close 0.19%, VWAP 0.03%; 13:00 to 15:00 UTC: Close 0.39%, VWAP 0.07%; 16:00 to 18:00 UTC: Close 0.18%, VWAP 0.03%; 19:00 to 21:00 UTC: Close 0.17%, VWAP 0.02%; 22:00 and 23:00 UTC: Close 0.21%, VWAP 0.03%. Our computation; the inputs and the method are in the text.
  • Close
  • VWAP
Our computation; the inputs and the method are in the text.Hourly Bitcoin, 25 UTC days from October 4 to October 28, 2026: absolute change from one hourly bar to the next, averaged over the 25 days and over the bars of each block. The 00:00 bar, where VWAP restarts, is left out.

What it did on 25 sessions of hourly Bitcoin

We took the 25 complete UTC days of hourly Bitcoin from October 4 to 28, 2026, noted on which side of VWAP price closed at the end of the first hour, and followed the remaining 23 hourly closes of each day.

Measure, 25 sessionsResult
Days where all 23 later hourly closes stayed on the first hour's side1 (October 22, a day down 3.18%)
Days that ended on the first hour's side8 of 25
Hourly closes within one band (±1 standard deviation)40.70% of 575
Hourly closes within two bands (±2 standard deviations)86.61% of 575

Where the 575 later hourly closes sat against VWAP

Where the 575 later hourly closes sat against VWAP. More than 2 deviations above: 6.6%; 1 to 2 deviations above: 22.4%; 0 to 1 deviation above: 19.0%; 0 to 1 deviation below: 21.7%; 1 to 2 deviations below: 23.5%; More than 2 deviations below: 6.8%. Our computation; the inputs and the method are in the text.
Our computation; the inputs and the method are in the text.Hourly Bitcoin, October 4 to 28, 2026, 23 closes per UTC day after the first one, in volume-weighted standard deviations of the session. Highlighted: inside the ±1 band, 40.7% in all.

On this sample, which side of VWAP the first hour closed on was a poor guide to the rest of the day: 17 sessions ended on the other side. Price crossed the line several times on most days, up to 8 times on October 15 and October 26. Only October 22, a steady fall, kept every later hourly close below VWAP; October 9 and October 14 came close, with 19 of 23 closes on the same side. The crossings bunched early, as the lag predicts: 76 of the 107 crossings came on the hourly bars from 01:00 to 11:00 UTC, 31 on the twelve bars from 12:00 to 23:00.

The five sessions on the chart

The chart above covers October 24 to 28, 2026, 120 hourly bars:

  • October 24: below VWAP after the first hour, 15 of the next 23 closes below, then a close 0.57% above it; the day lost 0.37%;
  • October 25: above after the first hour, only 8 closes above after that, a close 0.51% below VWAP and a day down 0.91%;
  • October 26: above after the first hour, 8 crossings, a close 0.07% above VWAP on a flat day;
  • October 27: below after the first hour, a close 0.78% above VWAP and a day up 0.95%;
  • October 28: below after the first hour, a close 0.34% above VWAP and a day up 0.54%; the last bar, at 23:00 UTC, closed at 81,343.50 with VWAP at 81,068.11.

Last hourly close against the session VWAP, October 24 to 28

Last hourly close against the session VWAP, October 24 to 28. October 24 (below after the first hour): +0.57%; October 25 (above after the first hour): −0.51%; October 26 (above after the first hour): +0.07%; October 27 (below after the first hour): +0.78%; October 28 (below after the first hour): +0.34%. Our computation; the inputs and the method are in the text.
Our computation; the inputs and the method are in the text.Hourly Bitcoin, 2026: distance of the 23:00 UTC close from the session VWAP, which restarts at 00:00 UTC; in brackets, the side of the VWAP at the close of the first hour.

Four of these five days ended on the opposite side from where the first hour left them. Treat this as a description of five days, not a rule.

Our view, untested: on a 24-hour market, VWAP is more useful as a reference price than as a signal line. Use it to judge whether a planned entry is above or below the session's average cost, and pair it with a trend tool on a higher time frame, such as the 200-day moving average. A time-based alternative that ignores volume is the TWAP.

Where VWAP comes from

StockCharts says VWAP was originally developed by institutional investors to place large orders without disrupting the market. The same idea appears in academic work: in March 1988, Berkowitz, Logue and Noser published "The Total Cost of Transactions on the NYSE" in The Journal of Finance (vol. 43, no. 1). Their abstract describes a measure of execution costs, "the volume-weighted average price over the trading day," and says it "yields results that are less biased than measures that use single prices, such as closes." Applied to more than 14,000 trades, it put total costs, commission plus market impact, at 23 basis points on average. StockCharts describes the same benchmark use for individual orders: a buy below VWAP is a good fill.

Does it work?

We have no long-run measurement of VWAP: our bench tests chart patterns detected by rule against random entries, not indicators (see our methodology). The 25 sessions above are one month of one market; they describe how price moved around the line, not an edge. For other trend tools, see the full list of trend indicators.

Every pattern's results on one sheet

The only pattern cheat sheet with tested stats: 152 patterns, what the textbook says and what really happened on 6 markets. Free PDF, sent to your inbox.

The PDF by email, plus our newsletter: new tests and results that change. Unsubscribe anytime. Privacy.

Frequently asked questions

What does the VWAP indicator tell you?

The average price paid for the asset since the session opened, weighting each price by the volume traded there. Price above VWAP means the latest trades are above the day's average cost; below means they are under it.

Who created the VWAP?

No single inventor is documented. StockCharts says institutional investors developed it to place large orders without disrupting the market, and a 1988 Journal of Finance paper by Berkowitz, Logue and Noser used the day's volume-weighted average price as its measure of execution costs on the NYSE.

Can you use VWAP on a daily chart?

Not in its standard form. VWAP resets each session, so on a daily bar it equals that bar's typical price. For a volume-weighted line across several days, use an anchored VWAP or a volume weighted moving average.

When does VWAP reset for crypto?

It depends on the platform, since crypto trades around the clock. Platforms differ; our chart uses midnight UTC. Check the setting before comparing levels.

Is VWAP a good trading signal?

We have no long-run test. Over 25 sessions of hourly Bitcoin in October 2026, the side of VWAP after the first hour matched the side at the close on only 8 days, so on that sample it was a poor directional signal.

Why does VWAP flatten late in the day?

Each new bar adds its volume to a growing total, so it moves the average less and less. StockCharts notes that this lag increases as the session goes on.

Sources

  1. StockCharts ChartSchool, Volume Weighted Average Price (VWAP)
  2. Berkowitz, Logue and Noser, The Total Cost of Transactions on the NYSE, The Journal of Finance 43(1), 1988
, Founder of PatternsWizard

Passionate about finance, quantitative trading and data analysis. Amaury built the PatternsWizard test bench, which detects each pattern on full price history and measures what happened next against random entries.

PatternsWizard tests chart and candlestick patterns on full market history, against random entries. How we test patterns.

179 patterns testedRandom-walk self-test passed

Related guides

All trend indicators articles