What is a VWAP mean reversion strategy?

A VWAP mean reversion strategy bets that when price stretches unusually far from VWAP (the volume-weighted average price since the open), it is more likely to come back toward VWAP than to keep going. The stretch is measured with standard-deviation bands around VWAP, most often 2 standard deviations (2σ); the trade goes against the last move, with the stop beyond the stretch and VWAP as the target.

It is the opposite bet to the VWAP pullback. The pullback trades with a trend after price comes back near VWAP; mean reversion trades against a run after price has gone far from it. Below: what the bands measure, the rules as numbers, a real session that reached VWAP and a real session that failed, and how to keep a trend day from ending a prop account with a trailing drawdown.

What VWAP's standard-deviation bands measure

VWAP since the anchor = Σ(price × volume) ÷ Σ(volume), usually anchored to the 9:30 ET cash open. The bands sit k volume-weighted standard deviations (σ) either side: how spread out today's traded prices are around VWAP, weighted by volume.

• A close beyond 2σ means price is unusually far from where most of today's volume traded. It does not mean "too far": the bands widen as the day spreads out, and on a trend day price can ride outside them.
• Platforms differ in anchor time, the price used (close or typical price) and the band formula. Your backtest's bands and your chart's bands must be the same calculation.

VWAP mean reversion vs VWAP pullback

The rules, as a checklist

On 1-minute MNQ bars:

1. VWAP from 9:30 ET on the typical price, with ±1σ and ±2σ bands (TradingView: the built-in Anchored VWAP on the 9:30 bar).
2. Window: new signals 10:00 to 14:29 ET only.
3. Stretch: a bar closes beyond the 2σ band.
4. Entry: the first close back inside the band, toward VWAP.
5. Stop: one tick beyond the stretch's extreme, never closer than 10 points. Target: VWAP at the entry bar, and only if it is at least as far as the stop.
6. Cap: 2 trades a day; after a stop on one side, that side is done for the day. Flat by 15:50.

This is the common 2σ version, taught as a worked example of the rules; that is not evidence it is profitable.

A session where the rules reached VWAP

Past session, selected to show the rules, not a typical result. MNQ1!, 1-minute, 10 August 2026; chart: TradingView snapshot, levels and labels marked by Tarrow Futures. MNQ is $2 a point (CME Group contract specs).

A session where the same rules failed

Past session, selected to show the rules, not a typical result. MNQ1!, 1-minute, 5 August 2026; chart: TradingView snapshot, levels and labels marked by Tarrow Futures.

5 August was a trend day under a falling VWAP, the same morning our VWAP pullback example used as its target day. Every new low looked stretched: after the stop, price closed back inside the lower band five more times by 12:14 ET, and the one-stop-per-side rule took none of them. The session closed at 29,606.75. The wider stop also meant fewer contracts, so the loss stayed inside the example budget; at the first session's size of 3, the same stop would have cost $277.50. We don't offer a filter for spotting trend days in advance. Test any filter before trusting it.

Hypothetical results. The outcomes above are what the written rules would have done on selected past sessions, before commissions and slippage; they are not trades anyone placed. CFTC Rule 4.41: hypothetical or simulated performance results have certain limitations. Unlike an actual performance record, simulated results do not represent actual trading. Also, since the trades have not been executed, the results may have under- or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profit or losses similar to those shown.

How it fits prop firm rules

• Daily loss limit. Two trades at a $200 example budget is a $400 worst day by rule, inside a $1,000 example limit. Size each trade from its own stop: $163.50 at risk on 3 MNQ with the 27.25-point stop, $185 on 2 MNQ with the 46.25-point stop. The position size calculator does it for your numbers.
• Trailing drawdown. A fade with a far target can lose more often than it wins. Example sequence: six stops at $200 (−$1,200) then two targets of +$600 (+$1,200) ends flat, after a $1,200 dip into a $2,000 drawdown. Under an intraday trailing drawdown, open profit on a far target also lifts the floor before a reversal.
• Consistency. One target can be most of a week. Under a 50% rule (example), a $429 best day needs $858 of total profit; check it with the consistency rule checker.

Common mistakes

• Fading a wick. A wick through the band is not a close beyond it. Waiting for the close, then the close back inside, is what puts the stop beyond an extreme that already exists.
• No side limit. On a trend day, "stretched" keeps getting more stretched. Decide before the open how many times you will fade one direction.
• Moving the target with VWAP. VWAP drifts toward price as the day goes on. Fix the target at entry, or test the moving version separately.
• Changing the band after seeing results. If you try 2σ, then 2.5σ, then 3σ on the same sessions and keep the best, you have learned about those sessions, not about the band.

For how this compares with other strategies under the same prop rules, see futures trading strategies ranked by difficulty.

Test it before you trade it

Fix the band and bar size before the test, check your platform's VWAP matches, run the strongest trend days in your data first, backtest with costs, then forward-test on a simulator. Our free guide, Build It, Break It, Then Trade It, shows how.

The full VWAP mean reversion breakdown has the rule card, both real MNQ sessions step by step, and the prop-rule arithmetic. It is free with your email, below.

Charts are TradingView snapshots of past sessions; the levels and labels on them were added by Tarrow Futures. Trade outcomes shown are hypothetical; see the CFTC Rule 4.41 notice after the examples above. Diagrams labelled as illustrations are drawn by Tarrow Futures and are not market data.