VWAP in trading: what it is, how to read it and futures strategies
VWAP is the session's volume-weighted average price. Learn how to read it, which setups have clear rules and what the evidence really says.
VWAP (Volume Weighted Average Price) is the session's average price weighted by the volume traded at each price. If price trades above it, the average buyer of the day is in profit; if it trades below, they are underwater. That is why it is used as an intraday bias reference in futures like NQ and ES.
What works in practice is simple: always use the same VWAP, read which side price is on and where the line is pointing, classify the day as trending or range-bound and only then pick the setup. Most VWAP losses come from skipping the third step.
In this guide you get the concept, the step-by-step read, the setups with rules (entry, stop, target and time window), the typical mistakes, what the evidence says and how it fits with a prop firm's rules.
Free tool: if you use NinjaTrader 8, you can download the VWAP indicator we built at El Trader Financiado for the community. See the indicator and download it.
What VWAP is
VWAP adds up, from the start of the session, every price multiplied by its volume and divides it by the total volume. It resets every session. Unlike a moving average, it does not give every bar the same weight: zones where a lot was traded pull the line harder.
Three variants you will see on any chart:
- Session VWAP. The classic. It starts with the session and resets every day.
- Deviation bands (±1σ, ±2σ, ±3σ). They measure how far price has moved from VWAP. They help you see extension, not predict reversals.
- Anchored VWAP (AVWAP). Same calculation, but starting from a point you choose: a major high or low, a sharp reversal or an event such as a Fed meeting.
One thing worth being clear about: VWAP was born as an institutional execution benchmark. A VWAP algorithm spreads a large order across the day following the volume curve. It does not buy "at the line". The idea that "institutions defend VWAP on pullbacks" is widely repeated, but no study backs it.
How to read VWAP step by step
1. Pick which VWAP you watch and don't change it
In CME futures there are two options, and they can be many points apart:
- RTH VWAP: starts at the cash open, 09:30 New York time.
- Full-session (ETH) VWAP: starts with Globex at 18:00 New York time.
There is no consensus on which is better. Some prop-firm-oriented blogs require the full-session one; others argue that in ES you should watch RTH because the thin overnight session skews the ETH line. What is a common rule: always use the same anchor and label which one it is.
2. Read the side of price and the slope
- Price above a rising VWAP: buyer bias.
- Price below a falling VWAP: seller bias.
- Price crossing a flat line: indecision.
3. Classify the day
- Trend day: price stays on one side of VWAP for hours.
- Range or "chop" day: price crosses VWAP again and again. One author puts it this way: "more than three crosses is a chop day". It is a rule of thumb, not measured data.
When price crosses VWAP repeatedly, the line stops giving information. The sensible move is to stand aside.
4. Pick the setup based on the day
- On a trend day: pullbacks in the direction of the bias.
- On a range day: fades at the ±2σ/±3σ bands, and only with exhaustion signals.
- Reclaim: only in a market rotating with no trend (not in chop with many crosses), when a 5-min candle closes clearly on the other side of VWAP with above-average volume. You enter on the first retest from the new side.
Anchored VWAP: where to anchor
Brian Shannon, the author who has done the most to popularize AVWAP, admits that choosing the anchor is "the most subjective part" of the analysis. Anchor on major highs and lows, on sharp reversals or on events, and treat the AVWAP as a zone to watch, not an automatic signal.
The most-cited adaptation to intraday futures uses three anchors:
- The weekly open (Sunday 18:00 New York time).
- Yesterday's high or low.
- The latest macro event: FOMC, NFP or CPI.
Shannon also describes the "VWAP pinch": two AVWAPs, one from a relevant high and another from a low, converging with price compressed between them. The expectation is a close outside and a pullback that fails to reclaim the line. It is a discretionary method and Shannon does not publish mechanical rules or win-rate statistics.
On the bands: phrases like "price lives inside ±1σ most of the day" have no published statistics behind them. The intraday distribution is not normal, so be wary of any percentage attached to the bands.
VWAP strategies: rule-based systems
There are two groups to keep apart. On one side, two mechanical systems with published results, both on ETFs rather than futures. On the other, futures setups from educators, with reasonable rules but unaudited.
Published mechanical systems (ETFs)
| System | Entry | Stop | Exit | Role of VWAP |
|---|---|---|---|---|
| "VWAP: The Holy Grail" (Zarattini and Aziz) | Long if the 1-min close is above session VWAP; short if below | The opposite cross | Flat before the close | Direction filter |
| "Beat the Market" / Noise Area (Zarattini, Aziz and Barbon) | Cross of bands calculated from the open, checked only at :00 and :30 | The greater of the band and VWAP (on longs) | End of day | Dynamic stop |
Notice what they have in common: VWAP is not the entry on a bounce. It is a filter or a stop inside a trend-following system.
Futures setups (educator rules, unaudited)
| Setup | Condition / filter | Entry | Stop | Target | Time window and management |
|---|---|---|---|---|---|
| Trend pullback | Price 30–60 min on one side of VWAP and VWAP sloping | First 5-min close moving away after a rejection wick | Beyond the wick | Prior extreme, ±1σ or POC/VAH/VAL | Don't trade the first 15 min |
| NQ bounce | Pullback within 5 points of VWAP + rejection candle, volume spike or delta | After confirmation | 8–12 points in NQ; 4–6 in ES | Prior high or 1σ band | 09:30–12:00 NY; maximum 2 attempts per session |
| Reclaim | Market rotating with no trend (not chop); clear 5-min close on the other side of VWAP with above-average volume | First retest from the new side | On the other side, past the retest extreme | Opening range, PDH/PDL or volume level, minimum 1.5:1 | Carries more weight if it holds 15 min after the open |
| ±1σ breakout | Trend day, break of band 1 with volume and no return to VWAP | Pullback to band 1 | On the other side of VWAP | ±2σ or PDH/PDL | Invalidated if it crosses VWAP again within 2–3 bars |
| ±2/3σ fade | Range days only, with exhaustion: falling volume, absorption, delta divergence | First close back toward VWAP | Beyond the extreme | Band 1 and then VWAP | Never on trend days or news days |
Example of the ES pullback with the author's numbers: an 8.25-point stop is $412.50 with an ES and $41.25 with an MES.
The same blog that proposes the NQ bounce publishes a table of its own win rates, with no sample size or period. That is not evidence and we do not reproduce it here.
On the fade, the author himself warns: if NQ rallies 200 points on an FOMC day and hits the second band, fading it "can cost you an entire evaluation".
Common VWAP mistakes
- Shorting above a rising VWAP because price "looks high". It is the most repeated way to get run over.
- Mixing sessions. Watching RTH VWAP one day and ETH the next changes every level.
- Trading chop. If price crosses the line again and again, there is no signal.
- Chasing price far from VWAP instead of waiting for the pullback. The stop ends up huge or badly placed.
- Trading without a hard stop. VWAP breaks constantly. The stop goes in at entry and size is set so that losing is routine.
- Trading the first 15 minutes, while VWAP is still adjusting to the opening volume (an author's rule).
- Treating VWAP as an exact price. Give the line some room around it.
- Loading more than three AVWAPs on the chart. You end up seeing levels everywhere.
What to combine it with
The best-defended confluence is VWAP on top of the prior day's POC, VAH or VAL: two independent references that coincide. We explain it in the volume profile guide.
Other common combinations:
- Prior-day and overnight levels (PDH/PDL, ONH/ONL) and the opening range. You have them in the guide to ICT killzones and session levels.
- Time window. The 09:45 to 12:00 New York window that VWAP blogs ask for coincides with the NY AM killzone. No source has measured that combination.
- Order flow confirmation (absorption, delta). We cover it in order flow trading.
- Entry zones such as a fair value gap or a supply and demand zone that lines up with VWAP.
An honest warning: no source quantifies how much these confluences improve results. The only test of a similar filter (VWAP cross plus the prior day's VWAP, on Indian Nifty futures between 2011 and 2019) was profitable before costs and stopped being profitable after commissions and slippage.
What the evidence says
What is proven and what is not:
- Proven: VWAP is a solid benchmark of execution quality in the academic literature.
- Evidence, but on ETFs: it works as a direction filter or stop in trend systems. Zarattini and Aziz's "Holy Grail" system advertised a Sharpe of 2.1 on QQQ. An independent, public replication brought it down to a Sharpe of about 0.7 out of sample. In addition, one of the authors runs an academy that sells the paper.
- No evidence in favor: there is no solid proof that mechanical fades of the bands work.
The futures setups in the table above are educators' rules. There is no public, auditable backtest of them on NQ or ES with real costs.
How VWAP fits a prop firm
The practical advantage of VWAP in an evaluation is that the stop per setup has a stable size, which lets you size risk before entering.
| Stop | NQ ($20/point) | MNQ ($2/point) | ES ($50/point) | MES ($5/point) |
|---|---|---|---|---|
| 5 points | $100 | $10 | $250 | $25 |
| 10 points | $200 | $20 | $500 | $50 |
| 12 points | $240 | $24 | $600 | $60 |
| 40 points (failed fade on a trend day) | $800 | $80 | — | — |
How to read it:
- Daily loss. With two bounce attempts of 10–12 points in NQ, the worst day costs about $400–$480 with one full contract. If your daily limit is larger, it fits. With MNQ, it fits with plenty of room.
- The fade is what breaks accounts. A fade with the stop "above the session high" on an NQ trend day can cost 30–50 points ($600–$1,000 per NQ). It is exactly the kind of loss that ends an evaluation.
- Consistency. A trend system with a low win rate strings together long losing streaks and concentrates profit in a few days. That clashes with consistency rules and the trailing drawdown. A higher-win-rate pullback fits better psychologically, even though its edge is not proven.
- News. Many prop firms restrict trading around news such as CPI, NFP or FOMC; check your firm's rules. Also, after the news, session VWAP is distorted by the volume spike. The alternative cited is an AVWAP anchored to the event itself.
- Size. Starting with micros (MNQ, MES) while you learn keeps losses small.
You can compare each firm's daily loss limits and drawdown in the comparator.
Free NinjaTrader VWAP indicator
At El Trader Financiado we built a free VWAP indicator for NinjaTrader 8, with deviation bands and anchored VWAP, made for the community and available in English and Spanish. The download, settings and installation are on its page: VWAP indicator for NinjaTrader.
Frequently asked questions
What is VWAP in trading?
It is the session's average price weighted by volume. It shows whether the average buyer or seller of the day is winning and is used as an intraday bias reference.
Should I use RTH VWAP or full-session VWAP in NQ and ES?
There is no consensus. RTH starts at 09:30 New York time and full-session at 18:00. What matters is always using the same one and knowing which one you are looking at.
How do you use VWAP bands?
As a measure of extension. On trend days, a pullback to band 1 is an entry in the trend's direction. On range days, bands 2 and 3 can serve for a fade with exhaustion. There are no published statistics on how much time price spends inside each band.
When should you not fade VWAP?
On trend days and news days. If price has been on one side for hours and VWAP is sloping, fading at band 2 is the mistake that breaks the most accounts.
Where do you anchor anchored VWAP?
On major highs and lows, sharp reversals or events. In intraday futures, the weekly open, yesterday's high or low and the latest FOMC, NFP or CPI are widely used.
Does VWAP really work?
As a direction filter or stop in trend systems there is evidence on ETFs, with performance decay out of sample. As a mechanical bounce or fade signal, public backtests lose after costs.
