TL;DR. Volume-Weighted Average Price (VWAP) is the average price a security trades at during a session, weighted by the size of each trade rather than the passage of time. Institutions use it as an execution benchmark — “did I buy this block above or below the day’s VWAP?” — and discretionary traders use it as an intraday reference line for trend and mean-reversion. It resets each session, hugs the price where the most volume actually changed hands, and behaves very differently from a simple or exponential moving average. Get the formula, one worked example, and the places where VWAP quietly misleads.
What VWAP actually measures
VWAP is defined precisely: for a defined intraday window, take every executed trade, multiply its price by its size, sum those products, and divide by the total volume in that window. The result is the average dollar-weighted price at which shares changed hands, not the average of quoted prices. Because larger prints pull VWAP toward themselves, the metric summarises where the day’s capital actually transacted, not merely where the tape flickered.
The formula, written out:
VWAP = Σ(Pi × Qi) / Σ Qi
Here Pi is the price of trade i, Qi is its share (or contract) quantity, and the sum runs over every trade in the chosen window — typically a single regular-hours session on the primary listing exchange. The idea was first put into practice as an execution benchmark on the NYSE in 1984 by James Elkins at agency broker Abel Noser, and it was academically formalised four years later in a widely-cited Journal of Finance paper by Berkowitz, Logue, and Noser on the total cost of NYSE transactions (see the Wikipedia entry on VWAP for the original citations).
Why the “volume-weighted” part matters
An analogy: if a stock trades 100 shares at $50 and then a 10,000-share block prints at $49.50, the time-weighted midpoint of those two prices is $49.75. The volume-weighted midpoint is about $49.505 — the block dominates because that is where almost all the capital changed hands. Time-based moving averages ignore that distinction. VWAP does not.
A worked example
Imagine an illiquid stock trades exactly four times in the first thirty minutes of the day:
| Time | Price ($) | Shares | Price × Shares ($) | Cumulative VWAP ($) |
|---|---|---|---|---|
| 09:30 | 100.00 | 500 | 50,000 | 100.00 |
| 09:45 | 101.00 | 1,000 | 101,000 | 100.67 |
| 09:55 | 99.50 | 8,000 | 796,000 | 99.75 |
| 10:00 | 100.20 | 500 | 50,100 | 99.80 |
Notice what the 8,000-share print at $99.50 does. It is a below-average price on a low count of transactions, yet it drags VWAP from $100.67 down to $99.75 in one tick, because it represents the day’s dominant order. Anyone who bought earlier at $101 is now filled “above VWAP” even though only one print occurred at that level. That is exactly what makes VWAP a useful execution benchmark: it captures where the crowd’s money actually transacted, not where a quote briefly touched.
How institutions use VWAP
The first and by far the most important use of VWAP is as an execution-quality benchmark. A pension fund that needs to buy 500,000 shares of a mid-cap without moving the tape does not just hit the offer at 09:31 — that would create market impact and get the fund a terrible average fill. Instead, the fund’s broker slices the parent order into hundreds of child orders spread across the day, targeting participation with the natural volume profile.
At close of business the broker’s fill is compared to the session VWAP:
- Beating VWAP (buy). Average fill price is below VWAP — the algo bought cheaper than the day’s money-weighted average. Good.
- Missing VWAP (buy). Average fill price is above VWAP. The algo either mistimed participation or chased liquidity when it should have been patient.
- Guaranteed VWAP. Some brokers will offer a client the day’s VWAP as a fixed price and take the execution risk themselves; the broker’s desk then trades throughout the day trying to beat that promise.
Because VWAP is the shared benchmark, everyone can grade a trade the same way. That standardisation matters. Regulators write best-execution rules — FINRA Rule 5310 in the U.S. — that require broker-dealers to seek the “most favorable” terms reasonably available for client orders (see FINRA Rule 5310). VWAP slippage is one of the yardsticks compliance and best-ex committees actually apply to the resulting trades.
How discretionary traders use VWAP
Retail and prop-desk traders lean on VWAP for a different reason: it is a natural intraday reference line that a lot of participants are watching, which makes it partly self-fulfilling.
- Trend read. Price sustained above the session VWAP is treated as a buyers-in-control tape; sustained below is sellers-in-control. Crossovers get attention.
- Mean-reversion anchor. Intraday scalpers fade extended moves back to VWAP on the assumption that price gravitates toward where the day’s volume clustered.
- Entry/stop framework. A common approach is to buy pullbacks to VWAP in an uptrending name and stop out if price closes an entire bar below it.
- Anchored VWAP. Popularised by trader Brian Shannon, this variant lets the user drop the VWAP anchor at a specific event bar — earnings, a Fed meeting, an all-time high — and then track how price has behaved relative to the volume-weighted mean since that catalyst. It is one of the most-used forms of VWAP in modern discretionary trading.
VWAP vs SMA vs EMA: they are not the same tool
Newer traders often blur VWAP, the simple moving average (SMA), and the exponential moving average (EMA). The three sit in the same visual family — a smoothed line on top of a candle chart — but they measure different things and behave differently.
| Feature | VWAP | SMA | EMA |
|---|---|---|---|
| Weighting | By trade volume | Equal per bar | Exponential in time |
| Resets | Each session (or at anchor) | Rolling window, never resets | Rolling, never resets |
| Timeframe | Intraday native | Any | Any |
| Sensitive to block prints? | Yes — by design | No | No |
| Primary use | Execution benchmark, intraday anchor | Trend smoothing over N bars | Faster trend smoothing |
| Late-day behaviour | Becomes “sticky” — a new trade barely moves it | Responds evenly to each bar | Responds most to the latest bars |
Where VWAP quietly misleads
VWAP is a good tool, not a magic line. It fails or misleads in four specific situations:
- Late in the session, VWAP is stubborn. By 15:30 the cumulative denominator (total volume so far) is huge, so a single tape print barely moves the calculation. A late-day breakout can look weak simply because VWAP no longer reacts.
- Illiquid names distort it. In a thinly traded stock, one 20,000-share block can pin VWAP for hours. What looks like an “algo defending VWAP” is often just an artifact of one dominant print.
- News-driven gaps make it stale. If a stock gaps 10% on earnings, the pre-gap VWAP tells you nothing about the new regime; discretionary traders re-anchor VWAP to the gap open in these cases.
- It is not a strategy, it is a reference. Buying every pullback to VWAP in a downtrending stock is a losing pattern. VWAP tells you where the day’s money transacted; it does not tell you which direction the trend runs.
The last point deserves emphasis. VWAP works as a decision aid when it is combined with a directional framework — trend structure, sector context, relative strength, an underlying catalyst. Trading VWAP by itself, in isolation from anything else, tends to look great on backtests of trending days and lose money on chop.
Related concepts to learn next
If VWAP made sense, four adjacent ideas are worth reading up on: (1) TWAP, the time-weighted average price, which is the sibling execution algorithm that ignores volume; (2) market impact and implementation shortfall, the broader family of execution-cost measures that VWAP slippage is only one part of; (3) Level 2 quotes and the order book, which show where volume is queued to trade rather than where it has already traded; and (4) volume profile (also known as market profile), which slices the day’s traded volume by price level instead of by time and often flags the same “value area” that VWAP hovers inside.
Sources
- FINRA Rule 5310, “Best Execution and Interpositioning” — finra.org/rules-guidance/rulebooks/finra-rules/5310
- Wikipedia, “Volume-weighted average price,” for the formula, 1984 Abel Noser origin, and the 1988 Berkowitz–Logue–Noser Journal of Finance citation — en.wikipedia.org/wiki/Volume-weighted_average_price
- SEC Investor.gov, glossary of investing basics (context for benchmark and best-execution language) — investor.gov/glossary
- ecmsource, “Moving Averages Explained: SMA, EMA, 50/200 Cross” (companion piece) — ecmsource.com moving averages
Disclosure: This article is for informational purposes only and is not investment advice.