TARROW™FUTURES Free breakdown

Home / Resources / Strategy library / VWAP mean reversion

Strategy library · Education only

VWAP mean reversion: fading the stretch back to the line.

When price stretches far from VWAP, some traders bet on the snap back. Here is how the stretch is measured with standard-deviation bands, how this differs from the VWAP pullback, two real MNQ sessions, and why a trend day is the risk to plan for.

Unlocked. Your link is on its way by email. Jump to the full breakdown ↓

The full breakdown · free with your email

Unlock the full VWAP mean reversion breakdown

The rule card, a real session it reached VWAP and a real one it stopped out, and the one-stop-per-side rule that kept the trend day to one loss. The free overview is below; we also email you the link, so it is one tap away next time.

What is a VWAP mean reversion strategy?

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

Illustration, not market data: two panels with a VWAP line and bands. Left, the VWAP pullback: price pulls up to the 1 standard deviation band under a falling VWAP and the trade goes short, away from VWAP. Right, VWAP mean reversion: price closes beyond the 2 standard deviation band and the trade goes back toward VWAP.
Same line, opposite bets. The pullback trades with the trend from near VWAP; mean reversion trades a stretch beyond the 2σ band back to VWAP. Illustration drawn by Tarrow Futures, not market data.

Mean reversion vs the VWAP pullback

Our VWAP pullback breakdown trades the other way. It waits for a trend to pull back toward VWAP and enters with the trend. Mean reversion waits for price to run away from VWAP and enters against the run. Same line, same kind of bands, opposite bets:

VWAP pullbackVWAP mean reversion
BetThe trend continuesThe stretch snaps back
Price at entryNear VWAP (the 1σ band)Far from VWAP (just back inside 2σ)
DirectionWith the trendAgainst the last move
TargetA multiple of the riskVWAP itself
The day it fearsA trend that gives wayA trend day that keeps stretching

That last row is the whole story. On a strong trend day, price can close outside the outer band again and again, and each fade is a stop. Most of this strategy's rules are about how many times you let that happen.

What the bands measure, in one line

VWAP = Σ(price × volume) ÷ Σ(volume) since the anchor, usually the 9:30 ET cash open. The bands sit a number of volume-weighted standard deviations (σ) either side: the spread of the day's traded prices around VWAP, weighted by volume. A close beyond 2σ means price has moved unusually far from where most of today's volume traded. "Unusually far" is not "too far": the bands widen as the day spreads out, and on a trend day price can ride outside them for an hour.

The idea in five lines

  • Stretch: a 1-minute bar closes beyond the 2σ band (a wick is not enough).
  • Entry: the first close back inside the band, toward VWAP.
  • Stop: one tick beyond the stretch's extreme, never closer than 10 points.
  • Target: VWAP at the moment of entry, and only if it is at least as far away as the stop.
  • Limits: two trades a day, and after a stop on one side, no more trades on that side.
Illustration, not market data: a bar closes above the upper 2 standard deviation band (1); bars stay outside and make an extreme; the first close back inside the band is the entry (2). The stop sits one tick past the extreme, at least 10 points away, and the target is VWAP at the moment of entry.
Stretch, close back inside, stop past the extreme, target VWAP. If VWAP is closer than the stop, there is no trade. Below the lower band everything is mirrored. Illustration drawn by Tarrow Futures, not market data.

Why the trend-day rule matters in a prop account

A fade loses most on the days a trend never stops. A daily loss limit treats the fifth faded rally exactly like the first, so the rule set needs a line that says "the market is not reverting today, stop trying". This one is simple: one stop on a side ends that side for the day, and two trades end the day. The real failed session in the full breakdown shows it at work: after the first stop, price closed back inside the band five more times, and the rule took none of them.

The stop also moves a lot from day to day, because it sits beyond whatever extreme the stretch made. On the two MNQ sessions in the full breakdown it was 27.25 and 46.25 points, $54.50 and $92.50 per micro contract before costs. Size has to come from the stop, every time.

What is in the full breakdown

  • The rule cardAnchor, bands, window, stretch, entry, stop, target, cap and the one-stop-per-side rule, in numbers.
  • A real session where it reached VWAPMNQ, 10 Aug 2026: a stretch above the band and the trip back, step by step.
  • A real session where it failedMNQ, 5 Aug 2026: a trend day under a falling VWAP, seen from the fade's side.
  • Prop-rule fitSizing from the stop, the run of stops to budget for, and a far target under an intraday trailing drawdown.
  • A test-it-yourself planFixing the band before the test, matching your platform's VWAP, and checking trend days first.

The full breakdown · free with your email

VWAP mean reversion: rules, real charts, prop-rule fit

Preview below. Unlock it at the top of the page ↑ and it opens here.

1. The rule card

Teaching example · not a recommendation

VWAP mean reversion from the 2σ band, 1-minute bars

  1. VWAP. Anchored at the 9:30 ET bar, on the typical price ((high + low + close) ÷ 3), with volume-weighted bands at ±1σ (context only) and ±2σ (the rule band). On TradingView: the built-in Anchored VWAP placed on the 9:30 bar, bands on, multipliers 1 and 2.
  2. Window. New signals only from the 10:00 bar to the 14:29 bar ET. The first half hour is skipped: the bands are built from too few bars to mean much.
  3. Stretch. A 1-minute bar closes beyond the 2σ band: above the upper band sets up a short, below the lower band a long.
  4. Entry. The first 1-minute close back inside the 2σ band. Enter at that close, toward VWAP.
  5. Stop. One tick (0.25) beyond the stretch's extreme: the highest high for a short, the lowest low for a long, from the stretch bar to the entry bar. Never closer than 10 points to the entry.
  6. Target. VWAP's value when the entry bar closes, rounded to the tick, and fixed from then on. If VWAP is closer than the stop, there is no trade.
  7. Exit. Stop, target or 15:50 ET, whichever comes first. A bar that touches both counts as the stop.
  8. Cap and side rule. At most 2 trades a day. After a stop on one side, no more trades on that side that day.
  9. Size. Contracts = the money you allow per trade ÷ ((stop in points + your cost allowance in points) × the point value), rounded down. If that is zero, there is no trade.

This is the common 2σ version of the strategy, taught as a worked example of the rules; that is not evidence it is profitable. MNQ is $2 a point (CME Group contract specs). The examples below use a $200 budget per trade and a 2-point cost allowance (example figures).

2. A session where the rules worked

Real chartSelected example, chosen to show the rulesReached VWAP

MNQ 1-minute chart, 10 August 2026, 9:45 to 10:40 ET, with an anchored VWAP from 9:30 and its 1 and 2 standard deviation bands: a 10:06 close above the upper 2 sigma band, a high of 29,900, a 10:17 close back inside at 29,873.00 (short), the stop at 29,900.25 and the VWAP target at 29,801.50 traded at 10:25
MNQ1! · 1-minute · 10 Aug 2026, 9:45–10:40 ET · TradingView Anchored VWAP from 9:30 with ±1σ and ±2σ bands · Chart: TradingView (snapshot; levels and labels marked by Tarrow Futures). A past session, selected to show the rules, not a typical result.
StepWhat the chart showsNumber
Stretch10:06: a 1-minute close above the upper 2σ band arms a shortclose 29,858.00
ExtremePrice stayed outside the band until 10:16; the highest high came at 10:1529,900.00
Entry10:17: the first close back inside the band. Short at the close29,873.00
StopOne tick above the extreme29,900.25, 27.25 pts
TargetVWAP when the entry bar closed (71.5 pts away, more than the stop, so the trade is allowed)29,801.50
Size$200 ÷ ((27.25 + 2) × $2) = 3.4, rounded down (example budget and cost allowance)3 MNQ
OutcomeVWAP target traded at 10:2571.5 × $2 × 3 = +$429

The morning ran up hard enough to close outside the outer band at 10:06 and stayed outside for ten minutes. The rule did not fade the first close outside: it waited for a close back inside, so the stop could sit above a high that had already been made. After the entry, the highest price was 29,885.00 at 10:20, 15.25 points short of the stop. The target was far, about two and a half times the risk. That is by design: the entry is far from VWAP and the stop is close to the extreme. Risk at 3 MNQ was 3 × 27.25 × $2 = $163.50. Dollar figures are before commissions and slippage. Hypothetical: what the written rules would have done that day, not a trade anyone placed (CFTC Rule 4.41 notice below the next example).

3. A session where the same rules failed

Real chartSelected example, chosen to show the rulesStopped out

MNQ 1-minute chart, 5 August 2026, 9:45 to 12:20 ET, with an anchored VWAP from 9:30 and its bands: a 10:29 close below the lower 2 sigma band, a 10:33 close back inside at 29,899.25 (long), the stop at 29,853.00 hit at 10:37, and five later closes back inside the band marked as skipped while price kept falling under a falling VWAP
MNQ1! · 1-minute · 5 Aug 2026, 9:45–12:20 ET · TradingView Anchored VWAP from 9:30 with ±1σ and ±2σ bands · Chart: TradingView (snapshot; levels and labels marked by Tarrow Futures). A past session, selected to show the rules, not a typical result.
StepWhat the chart showsNumber
Stretch10:29: a close below the lower 2σ band arms a longclose 29,915.25
Entry10:33: a close back inside the band, by less than half a point. Long at the close29,899.25
StopOne tick under the stretch's low (29,853.25, at 10:32)29,853.00, 46.25 pts
TargetVWAP when the entry bar closed (82 pts away)29,981.25
Size$200 ÷ ((46.25 + 2) × $2) = 2.07, rounded down2 MNQ
OutcomeStopped at 10:37. Longs were then off for the day, and no short signal came46.25 × $2 × 2 = −$185
The lesson: a fade needs a rule that admits it is a trend day

This was the same morning our VWAP pullback breakdown used as its target day: price stayed under a falling VWAP for hours. For a fade, that is the worst kind of day, because every new low looks stretched. After the stop, price closed back inside the lower band five more times by 12:14 ET (10:42, 10:46, 10:56, 12:04 and 12:14). Without the one-stop-per-side rule each of those was another long; with it, the first stop ended the longs. The session closed at 29,606.75, near its low of 29,593.00.

The size rule did the rest. The stop was 46.25 points, much wider than the first example's 27.25, so the same $200 budget allowed 2 MNQ instead of 3, and the loss stayed inside it. At the first day's size, the same stop would have cost 3 × 46.25 × $2 = $277.50 before costs.

Can you tell a trend day from a reverting day in advance? Traders propose many tells: the slope of VWAP, the opening range, the news calendar. We do not offer one. Test any filter before trusting it.

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.

4. How it fits prop firm rules

Bar chart at a $200 example budget: a 27.25-point stop sized to 3 MNQ risks $163.50; a 46.25-point stop sized to 2 MNQ risks $185; the same 46.25-point stop at 3 MNQ would risk $277.50, over the budget.
Size from the stop, every trade. Contracts = $200 ÷ ((stop + 2 pts of example costs) × $2), rounded down: 3 MNQ at 27.25 points, 2 at 46.25. Example figures, not results.

Daily loss limit: the cap sets the worst day

Example figures: a $200 budget per trade and a cap of 2 trades make the worst day by rule $400 plus costs, inside a $1,000 daily limit. Size every trade from its own stop: 3 MNQ at a 27.25-point stop, 2 at 46.25, and none once the stop is wider than 98 points (one micro at (98 + 2) × $2 is already the whole $200).

Trailing drawdown: budget for the run of stops, not the day

A fade with a far target can lose more often than it wins and still depend on a few long trips back to VWAP. Example sequence, not results: six stops at $200 is −$1,200; two targets at +$600 is +$1,200. Back to flat, but the account went $1,200 into a $2,000 drawdown on the way.

Bar chart of an example sequence of eight trades: six stops of minus $200 then two targets of plus $600. The running total falls to minus $1,200 before returning to $0.
Budget for the run of stops. Six stops at $200 then two targets at $600 end flat, after a $1,200 dip into a $2,000 drawdown. Example figures, not results.

Under an intraday trailing drawdown it is worse: a far target lets open profit lift the floor a long way before a reversal. Example: 3 MNQ that reach +40 points open (+$240) from a balance high and then stop out at 27.25 points (−$163.50) shrink the room by $240 + $163.50 = $403.50.

Consistency rule: one target can be the week

Example figures: under a 50% consistency rule, a +$429 best day needs at least $429 ÷ 0.50 = $858 of total profit before it fits. Rules and percentages differ by firm; check yours.

Close-out time

The card's 15:50 ET exit is ours. If your firm requires you to be flat earlier, its time becomes the time exit, and the last entry time should move with it. Check your firm's current rulebook.

Free toolPosition size calculator: contracts from your limit and your stop Free toolDrawdown room calculator: where your floor sits now Free toolConsistency rule checker: does your best day fit? BlogWhy open profit raises an intraday trailing floor

5. Test it yourself before you trade it

  1. Write the band down and do not touch it during the test. Band width (2σ, 2.5σ, 3σ) and bar size (1-minute or 5-minute) are part of the strategy. Pick them before you look at results; changing them afterwards and re-running on the same sessions tells you about those sessions, not about the strategy.
  2. Check your VWAP matches your platform's. Anchor time, the price used (close or typical price) and the band formula all differ between platforms. A close "back inside" by half a point, as on 5 August, depends on it.
  3. Test the side rule on trend days first. Find the strongest trend days in your data and check that the cap and the one-stop-per-side rule kept each one to a normal loss.
  4. Decide a news rule. Data releases at 10:00 ET and the FOMC statement at 14:00 ET fall inside the window. Write down whether you trade through them.
  5. Backtest with costs, then count the worst run. Commission both ways and slippage on every fill (a fade enters a fast market; a late fill is never better than the backtest price). Put the worst run of stops against your trailing drawdown at your size, then forward-test on a simulator for weeks before any paid evaluation.
Free 32-page guideBuild It, Break It, Then Trade It: backtest and stress-test a strategy, step by step Strategy libraryThe other side of the same line: the VWAP pullback

Know a trader learning this strategy? Send them this breakdown. It's free for them too.

Disclosures

Education only. This page explains how a widely known, public strategy is defined and how you could test it yourself. It is not advice, not a recommendation to trade any contract, size or strategy, and not a signal: it shows no setup for any future session. Charts are real past sessions, chosen to show the rules (including a day they failed); they are selected examples, not typical results. The charts are TradingView snapshots (MNQ1!, regular trading hours, one-minute unless the caption says otherwise); diagrams labelled as illustrations are drawn by Tarrow Futures and are not market data. The level, range, entry, stop and target marks and labels on them were added by Tarrow Futures. Bars may differ slightly from those on your platform. Trade outcomes on this page are what the written rules would have done on those days, before commissions and slippage. They are not trades Tarrow Futures placed. Dollar figures for sequences and rules are example figures, not results. Prop firm rules differ and change; your firm's current rulebook decides. Futures trading involves substantial risk of loss. Automation does not remove risk.

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.