Early access is open. Free sample datasets are live now. Download a sample

Learn

VWAP: How It's Calculated and How Traders Use It

7 min read · Updated 19 November 2026 · By the marketdata.co data team

VWAP is the average price paid for every unit traded since a starting point, weighted by volume. Execution desks grade fills against it, and day traders treat it as the line between buyers and sellers being in control. Both uses depend on calculating it correctly, and the details (where it resets, which price you weight) change the number more than most people expect.

Details

The calculation

VWAP = sum of (price × volume) / sum of volume, from the anchor point to now.

With tick data, use each trade's price and size. Bars need a price that represents each bar. Two common choices:

  • Typical price = (high + low + close) / 3. Most charting platforms use this.
  • The bar's own VWAP, if your data has it. Our bar schemas include a vwap field built from the trades inside each bar, which is more accurate than the typical price.

For stocks, VWAP resets each session, usually at the 9:30 ET open. Crypto has no session, so you pick an anchor, often 00:00 UTC.

Worked example on real bars

The first five rows of our free BTC-USDT 1-minute file, 31 October 2026, with a running VWAP built from the typical price:

timestamp high low close volume typical price running VWAP
15:40 77598.76 77598.75 77598.76 6.02156 77598.76 77598.76
15:41 77628.53 77598.75 77628.53 8.09918 77618.60 77610.14
15:42 77628.53 77576.34 77576.35 7.45285 77593.74 77604.47
15:43 77611.6 77576.34 77610 15.32788 77599.31 77602.33
15:44 77610.57 77609.99 77610.56 6.91454 77610.37 77603.60

Source: Binance. Typical price and running VWAP computed by us.

The 15:43 bar had twice the volume of its neighbours, so it pulled the VWAP toward its own price more than the others did. That is the whole idea: busy minutes count more.

Across all 500 bars (15:40 to 23:59 UTC), VWAP anchored at 15:40 comes to about 77,635.73 using the bars' own vwap field, and 77,635.55 using typical price. The two methods differ by under 20 cents on a $77,000 asset here, but on thin or gappy bars the gap is bigger. Its last close was 77,582.02, about 54 dollars below VWAP. Read the usual way, sellers had the upper hand over that stretch.

How traders use it

Execution benchmark. A fund buying a large position over the day wants an average price at or below VWAP. Beating it means the order was worked well.

Intraday bias. Price above VWAP means the average buyer since the open is in profit; below means they are underwater. Many day traders only take longs above it and shorts below it.

Pullback level. In a trend, price often returns to VWAP before continuing. A break and hold on the other side is read as a change in control.

Details

Compute VWAP and bands in pandas

"""Example: anchored VWAP with standard deviation bands on 1-minute bars."""
import pandas as pd

df = pd.read_csv("btc-usdt-1m-bars.csv", parse_dates=["timestamp"]).set_index("timestamp")

price = df["vwap"]  # or (df["high"] + df["low"] + df["close"]) / 3
pv = (price * df["volume"]).cumsum()
v = df["volume"].cumsum()
df["avwap"] = pv / v

var = ((price - df["avwap"]) ** 2 * df["volume"]).cumsum() / v
df["band_up"] = df["avwap"] + 2 * var ** 0.5
df["band_dn"] = df["avwap"] - 2 * var ** 0.5

print(df[["close", "avwap", "band_up", "band_dn"]].tail(3).round(2))

For stocks, group by trading date first so VWAP resets every session: df.groupby(df.index.tz_convert("America/New_York").date) and apply the same cumulative sums inside each group. The bands are a volume-weighted standard deviation; price outside them is stretched compared with the session so far.

Watch out

Common mistakes

Not resetting at the session. A VWAP that runs across days for a stock blends yesterday's trading into today. Reset at the open, or use an explicit anchored VWAP.

Including pre-market volume by accident. Extended-hours bars are thin and can drag VWAP before 9:30. Decide whether you want them, and filter on the session field in our stock bars.

Using close instead of a bar price. Weighting only the close ignores where trading happened within the bar. Use the bar's vwap or at least the typical price.

Treating VWAP as support. It's an average of past trades, not a wall. Late in the session it barely moves, so price can cross it easily.

Details

Get the data

Bars with a true per-bar VWAP. Every bar in our schema carries vwap and trade_count, so you don't have to approximate with the typical price.

FAQ

Frequently asked questions

How is VWAP calculated?

Multiply each trade's price by its volume, add those up from the anchor point, and divide by total volume over the same period. With bar data, use each bar's VWAP or typical price as the price. For stocks the anchor is normally the session open.

Why is my VWAP different from my charting platform?

Platforms differ in the price they weight (typical price vs close vs tick-level), whether they include extended hours, and where they reset. Match those three settings and the numbers should agree closely.

What is anchored VWAP?

It is VWAP started from a point you choose, such as an earnings day or a major high, instead of the session open. The line then shows the average price of everyone who traded since that event.

Does VWAP work for crypto?

Yes, but you must choose the anchor because crypto trades around the clock. 00:00 UTC is common, and some traders anchor to the weekly open or a major event instead.

Keep reading