In plain terms — VWAP answers: “At what average price did the quantities included in this calculation trade?” A large trade carries more weight than a small one.
VWAP stands for Volume-Weighted Average Price. It may appear on a chart, serve as an execution benchmark, or become the target of an algorithm.
VWAP is not investors' actual average cost. The same instrument can change hands repeatedly, starting positions are unknown, and the data scope may omit trades. Trading above or below VWAP does not prove who is profitable or who “controls” the session.
Formula and meaning
For individual trades with price (p_i) and quantity (q_i):
VWAP = Σ (pᵢ × qᵢ) / Σ qᵢ
Three trades — 100 units at 10.00, 200 at 10.20, and 50 at 9.80 — have traded value of 3,530 and volume of 350. VWAP is about 10.09. Their simple price average is 10.00 and ignores that the 10.20 trade is the largest.
Bar-based charts often use a representative price such as ((high + low + close) / 3), multiplied by bar volume. This is an approximation from aggregated data, not the only possible trade-by-trade benchmark definition.
Session, venue, and data quality
A VWAP needs a declared scope: instrument, venues, eligible trades, start and end, time zone, corrections, and volume source. Changing filters or window changes the result. FINRA, for example, publishes Treasury VWAP statistics using explicit criteria for reported trades.
The regular equity session is a common anchor; continuously traded markets need a convention. A daily crypto VWAP may reset at midnight UTC, exchange time, or another point. Two valid values can differ if this is not stated.
Benchmark and algorithm are different
The VWAP benchmark measures trades in the chosen scope. A VWAP algorithm uses volume forecasts or profiles to distribute a parent order and seek a result near that benchmark. It is best efforts: price, liquidity, limits, participation, and residual quantity prevent a guaranteed match.
Implementations differ. Coinbase Prime documents times, expiry, limit, and participation rate; Interactive Brokers calls its VWAP a best-efforts algorithm. These are product examples, not universal rules.
Reading the chart without a story
Price above VWAP means only that current price is above the weighted average for the scope; below means the opposite. Distance, crossover, and return may be descriptions, filters, or risk references, but are not self-sufficient signals.
Cumulative VWAP generally becomes less responsive as volume accumulates. Bands and anchored VWAP are extensions with their own methods; they do not automatically belong to the base definition.
Advanced: benchmark risk and TCA
Transaction cost analysis compares a fill with a VWAP aligned by side, quantity, currency, venue, and time. Choosing the benchmark after seeing the result creates selection bias. A large order may also contribute to the volume and prices that form the benchmark, so the reference is not always independent of the execution.
Review should combine VWAP with fill ratio, residual quantity, commissions, spread, costs, estimated impact, and opportunity cost. “Better than VWAP” on completed quantity can hide a large unexecuted remainder.
Limits and common errors
VWAP is backward-looking over included trades and does not forecast the next price. Incomplete or incomparable volume distorts it. Bar formulas lose trade sequence within each bar. A limit can prevent an algorithm from following its profile. Comparing VWAPs from different sessions, feeds, or currencies creates false precision.
Sources
- U.S. SEC, Release No. 34-44724 — VWAP calculation and methodological variation.
- FINRA, About Treasury Aggregate Statistics — reported-trade VWAP, filters, and exclusions.
- Coinbase Developer Documentation, Prime — Order Types — documented VWAP order parameters and behavior.
- Interactive Brokers, VWAP order type — one best-efforts VWAP algorithm.
- ESMA, Supervisory Briefing on Algorithmic Trading in the EU, 2026 — VWAP as an algorithm example and associated controls.