VWAP — Volume-Weighted Average Price
The answer
VWAP is the average price of every share traded in the session so far, with each price weighted by the volume that traded at it. It resets to nothing at the next open.
It matters for a reason that has almost nothing to do with charts. VWAP is the score institutions are graded on. A fund manager who bought 400,000 shares is judged by whether the average price paid beat the day's VWAP. That is why price reacts near the line. It is not a prediction. It is a target that thousands of orders are aiming at.
Why this costs you money
Cluster 01 article 05 explained that there is no single price, only a price at a size, and that the difference between the price you see and the price you get is a real cost called slippage. VWAP is the tool that turns that cost from a feeling into a number.
Here is the number most retail traders have never calculated.
Take your last 20 intraday buy orders. For each one, find where your fill price sat relative to VWAP at that moment. Most retail buys sit above VWAP. Most retail sells sit below it.
That is not an accident and it is not bad luck. It is the direct result of how retail orders get placed. You see a move, you decide, you press buy. The move is what made you press the button, so you bought after the move, which means you bought above the session's average. Then you sell when the position frightens you, which is after an adverse move, which means you sold below the session's average.
Say your buys average 0.35% above VWAP and your sells average 0.30% below it. That is roughly 0.65% per round trip, on top of brokerage, on top of taxes, on top of the spread. A trader doing 4 round trips a week is paying that 200 times a year.
This cost is invisible because it never appears on a contract note. It is not a fee. It is the gap between the price you took and the price that was available across the session. VWAP is the only easy way to see it.
The second cost is different and it is a misreading. Traders are taught "above VWAP is bullish, below VWAP is bearish", and they use it as an entry rule. Consider what that rule says. It says buy when price is above the day's average and sell when it is below. That is a rule that guarantees you buy above average and sell below average, which is exactly the behaviour the first cost describes.
How it works
The formula is short and every part of it matters.
For each bar in the session:
- Take the typical price: (high + low + close) ÷ 3.
- Multiply that by the bar's volume. This gives you the value traded in that bar.
- Add it to a running total of value.
- Add the bar's volume to a running total of volume.
- VWAP = running total of value ÷ running total of volume.
That is all of it. There is no smoothing, no period setting and nothing to tune.
Three consequences follow, and they are the whole indicator.
1. It is cumulative from the open. VWAP at 10:00 is built from 15 minutes of data. VWAP at 15:15 is built from the whole day.
2. It becomes stiffer as the day goes on. The denominator only grows. By the afternoon, a single 5-minute bar is a small fraction of the day's volume, so it barely moves the line. VWAP in the first 30 minutes is jumpy and close to price. VWAP in the last hour is almost a fixed horizontal level. These are 2 different things wearing the same name, and reading them the same way is the most common technical error with this tool.
3. It resets. At the next open, the running totals return to 0. Yesterday's VWAP does not exist today. This is not a flaw. It is the definition, and it is the honest limit stated in section 4.
Anchored VWAP
Anchored VWAP changes 1 thing: where the running totals start. Instead of the session open, you choose the starting bar yourself — a results announcement, a gap day, a major low, the day an index rebalance took effect.
The output answers a specific and honest question: what is the average price paid by everybody who has traded since that event?
That question has a real meaning. If price is above the anchored VWAP from a company's results day, the average buyer since those results is holding a gain. If price is below it, the average buyer since then is holding a loss. Brian Shannon popularised the technique and the reasoning behind it.
Anchored VWAP does not reset daily, which removes the biggest limitation of the intraday version. What it adds is a choice: you picked the anchor, and a different anchor gives a different line. That choice is where the discipline has to live.
What it tells you, and what it does not
VWAP tells you where the current price sits relative to the average price paid during a defined window. That is a fact, and it is a fact that a large number of professional orders are actively responding to.
Six limits.
It says nothing about anything beyond the session. Intraday VWAP is destroyed every evening. A stock closing above VWAP tells you nothing about tomorrow, because tomorrow's VWAP has not started yet. Anybody applying intraday VWAP to a swing decision is using a number that will not exist when their trade is still open.
It is a lagging average, not a forecast. VWAP contains only what has already traded.
It is not the same line on every platform. VWAP depends on which trades are counted. Different data feeds include different venues and different session definitions, so 2 charts can show 2 different VWAP lines for the same stock. This matters far more in the United States than in India, for reasons in the market sections below.
It has no meaning on illiquid instruments. If 90% of a day's volume traded in 3 trades, the volume-weighted average is the average of 3 decisions.
Price crossing VWAP is not a signal. On a normal day price crosses VWAP several times. Every crossing is a "signal" under the simple rule, and most of them lead nowhere.
The reaction near VWAP is a crowding effect, not a law. Price often stalls or turns near VWAP because execution algorithms are programmed to buy below it and sell above it. That behaviour is real, but it exists only while enough capital is being benchmarked that way. It is not a property of prices. It is a property of how the industry currently measures its own staff.
The decision rule
Treat VWAP as an execution tool first and a chart line second.
If you have already decided to buy, use VWAP to decide when within the day. Buying below the session VWAP is a measurably better average than buying above it, over many trades.
If price is above VWAP and has been all session, the buyers have controlled the day. That is a description, not a forecast. Use it to size and to time, not to decide direction.
For anything held longer than 1 session, use an anchored VWAP from a real event, and state the event out loud before you draw it. If you cannot name why the anchor matters, you are fitting a line to a chart you have already interpreted.
Unless the instrument is thin, or the day is an expiry or index rebalance day. On those days VWAP is dominated by mechanical flow that has no opinion about value, and reading it as market sentiment is reading a settlement procedure.
Try this now
Five minutes, and it produces a number in rupees or dollars that is currently missing from your trading records.
- Open your broker's trade history. Find your last 10 intraday trades, or if you do not trade intraday, your last 10 trades of any kind. Write down the date, the time, the side (buy or sell) and the fill price for each.
- For each trade, open a 5-minute chart of that instrument for that date, and add VWAP.
- Find the bar containing your fill time. Read the VWAP value at that bar.
- For each trade, calculate:
(your fill price − VWAP) ÷ VWAP × 100. Write it as a percentage. For a buy, a positive number means you paid above the session average. For a sell, a negative number means you sold below it. - Count. How many of your buys were above VWAP? How many of your sells were below it? Then average the absolute percentages.
What you should see. Most readers find that 6 or more of their 10 buys sat above VWAP, and most of their sells sat below it. The average distance is usually between 0.2% and 0.6%, and on volatile instruments it can be larger.
Now do the arithmetic that makes it real. Take your average distance, double it for the round trip, and multiply by the number of round trips you did last year. A trader averaging 0.4% each way over 150 round trips gave up 120% of a single position's value across the year, spread thinly enough that no single trade ever looked expensive.
That is the cost of buying after the move that convinced you. It is not brokerage, it is not tax, and no statement itemises it.
If your numbers come out near 0, or your buys sat mostly below VWAP, you already place orders with patience. That is worth knowing too, and almost nobody has measured it either way.
Three real cases
1. SEBI's pricing rules for preferential issues and qualified institutions placements (India, current law) — VWAP written into securities regulation When an Indian listed company issues new shares to a selected investor, it cannot choose any price it likes. The SEBI (Issue of Capital and Disclosure Requirements) Regulations set a floor price computed from the volume weighted average price of the shares over defined past periods before the relevant date . The principle is what matters here. The regulator did not use the closing price, and did not use the average of daily closes. It used volume weighting, because a price at which 100 shares traded and a price at which 500,000 shares traded are not equally informative. VWAP is not a chart indicator in this context. It is a legal definition of fair value.
2. Tesla's addition to the S&P 500, effective 21 December 2020 (United States) — what happens when the benchmark is not VWAP Tesla was added to the S&P 500 with effect from 21 December 2020, and index funds tracking the S&P 500 had to hold it. Those funds are measured against the index, and the index uses closing prices. So the funds needed to trade at the close, not at the day's average. An extraordinary volume traded in the closing auction on 18 December 2020, reported at the time as one of the largest closing auctions in United States market history. The lesson is the one this article is built on. Institutional behaviour follows the benchmark, and the benchmark is a choice. Change the measurement from VWAP to the close, and the flow moves from across the day to a single moment.
3. Brian Shannon and anchored VWAP (United States, 2008 onward) — the extension that removes the daily reset Brian Shannon, a United States trader and author, is widely credited with popularising anchored VWAP: starting the volume weighted calculation from a chosen event rather than from the session open. The value of the idea is not that it predicts. It is that it converts a vague question into an arithmetic one. "Are the people who bought this after the results announcement in profit or in loss?" has an exact answer, and the anchored VWAP from that day is the answer. Nothing else on a standard chart answers it.
The question that resolves it
A novice looks at price sitting above VWAP and asks: is this bullish?
An expert asks: what time is it?
VWAP at 09:45 is an average of 15 minutes and it moves with price, so being above it means almost nothing. VWAP at 15:10 is an average of the whole session and it is nearly fixed, so being above it means the day's buyers, across 6 hours, are collectively in profit.
The same line, the same relationship, 2 completely different amounts of information. The clock is the variable nobody reads.
What would make this wrong
The central claim is that VWAP matters because institutional execution is measured against it, and that retail fills sit systematically on the wrong side of it.
You would falsify the second half with the exercise above, on your own trades. If your buys sit at or below VWAP more often than above it, this article does not describe your behaviour. Believe your own count.
You would falsify the first half by watching whether price reactions near VWAP persist as institutional benchmarking changes. If large funds moved to measuring execution against implementation shortfall or arrival price instead of VWAP — which some already do — the crowding around the VWAP line would weaken. The effect is a fact about the industry, not a fact about prices, so it can decay.
Four honest limits.
First, this article does not claim VWAP predicts. It claims VWAP describes a real average and marks a level that a large amount of programmed order flow responds to. Those are different claims and only the second is defensible.
Second, the fill measurement measures cost, not skill. A trader who buys above VWAP and holds for 3 years has paid a cost that rounds to nothing. The measurement matters in proportion to how often you trade.
Third, anchored VWAP is chosen. Two traders anchoring at 2 different events get 2 different lines and 2 different conclusions. Nothing in the arithmetic protects you from picking the anchor that supports what you already believed.
Fourth, no evidence here shows that trading around VWAP is profitable. It shows that executing below it is cheaper than executing above it, which is a statement about cost.
In India
VWAP is more deeply built into Indian market structure than into any other major market, and most Indian retail traders do not know it.
The official closing price is a VWAP. NSE and BSE do not use the last traded price as the day's close. The closing price is computed as the volume weighted average price of trades in the last 30 minutes of the session. This means the number every Indian newspaper, every mutual fund net asset value and every stop-loss reference uses is itself a VWAP. The index closing value is computed the same way, from the weighted average of the index over the last 30 minutes.
That makes expiry days a VWAP problem. Indian index options are cash settled against the closing index value, and that value is a 30-minute weighted average . On expiry, an enormous notional value depends on where that 30-minute average lands. Positions are defended and attacked inside that window. What you see on the chart between 15:00 and 15:30 on an expiry day is frequently not opinion about value. It is settlement pressure on an average.
VWAP is a legal price. As described in case 1, SEBI uses volume weighted average price to set floors for preferential allotments and institutional placements. Buyback and takeover regulations also reference past average prices .
Indian VWAP is cleaner than American VWAP. Almost all Indian equity volume executes on NSE and BSE during a single continuous session from 09:15 to 15:30. There is no significant extended-hours trading and far less venue fragmentation. So the VWAP your broker shows is close to the VWAP everyone else sees.
The liquidity warning is severe below the large caps. NSE cash turnover is concentrated in a relatively small number of names. On a mid-cap or small-cap, a session's VWAP can be the weighted average of a handful of large orders, and price will not respect it, because no algorithm is being graded against it.
In the United States
VWAP is standard, universal, and messier than it looks.
Volume is fragmented. United States equities trade across many exchanges and a large share executes off-exchange at wholesalers and alternative venues . The consolidated tape captures most of it, but your charting platform may not use the consolidated tape. Two United States traders can be looking at 2 genuinely different VWAP lines on the same stock at the same moment. In India this problem barely exists.
Extended hours are usually excluded. United States shares trade before 09:30 and after 16:00 Eastern time. Most platforms exclude that volume from the regular-session VWAP. On a results day, a large amount of trading happens outside regular hours and never enters the VWAP calculation, while the price change it caused does appear on the chart.
The close is an auction, not an average. The United States primary listing exchanges run a closing auction that sets the official closing price. This is the opposite design choice from India's 30-minute weighted average, and it changes where institutional flow concentrates. Index funds, which are measured against closing prices, execute in that auction. Active managers, who are frequently measured against VWAP, execute across the day. The 2 benchmarks pull flow to 2 different places in the session.
VWAP algorithms are a standard product. Every large broker offers an execution algorithm that slices a large order across the day in proportion to expected volume, with the explicit goal of matching or beating VWAP. When you see steady, evenly spaced participation in a stock through the middle of the day, that is often what it is. There is no view in it. There is a schedule.
Where they differ, and what that tells you
The difference is where the average is used, and it produces 2 different practical instructions.
1. India makes VWAP official; the United States makes the auction official. The Indian close is a 30-minute volume weighted average. The United States close is the price from a single auction. So in India, the last 30 minutes are a contested average that many settlements depend on. In the United States, a single moment carries that weight.
What that tells you is where the mechanical flow sits. In India, be careful interpreting 15:00 to 15:30, especially on expiry days. In the United States, be careful interpreting the final seconds and the auction print, especially on index rebalance dates.
2. The Indian VWAP line is more trustworthy as a number; the United States VWAP line is more trustworthy as a level. Indian VWAP is computed from a single session on 2 exchanges, so it is consistent across platforms. But less Indian institutional money is formally benchmarked to intraday VWAP than in the United States, so the crowding effect that makes price react at the line is weaker. In the United States the number is fuzzier because of fragmentation, and the reaction is stronger because a very large amount of algorithmic execution is graded against it.
3. The practical instruction therefore differs. In India, use VWAP mainly as your own execution scorecard, and treat the last 30 minutes as a special period governed by settlement mechanics. In the United States, use VWAP as an execution scorecard as well, but expect genuine price reaction at the line during the middle of the session, and check what your data feed is actually including before you trust the exact value.
Carry this
- VWAP is the average price paid so far in the session, weighted by volume. It resets every day, so it says nothing about tomorrow.
- Price reacts near VWAP because institutional execution is measured against it, not because the level has power.
- VWAP early in the day and VWAP late in the day are 2 different measurements. Check the clock before you read the line.
- Use anchored VWAP for anything held longer than a session, and name the event you anchored to before you draw it.