VWAP reversion strategy

On ES futures, the VWAP reversion strategy lost $99,597 after costs and ran a $100,000 account down to its margin on April 9, 2025. The run stops there.

Mean reversionIntraday1-minute barsBacktested Nov 6, 2026

VWAP reversion backtest, the result

Contract

E-mini S&P 500

  • −$99.6K

    Net after costs

    After $78,660 in fees and slippage

  • 2 of 10

    Years that made money

    January 2016 to April 2025

  • −$104K

    Largest drawdown

    Peak to trough, 1 contract

  • −$36

    Average trade

    2,758 trades

VWAP reversion on ES, net after costs by month, from January 2016 to April 2025: −$99,597; buying and holding one ES contract: +$236,212

Net after costs by month, January 2016 to April 2025.

VWAP reversion rules

Trade back to the session's VWAP when price has stretched 2 standard deviations away from it.

A drawing of the rules on a made-up day, not a result.

  1. The VWAP is the Globex session's, from its 18:00 New York open, over each one-minute bar's typical price (high + low + close) / 3, with the volume-weighted standard deviation of that price around it.
  2. In the day session, a one-minute close above VWAP + 2 deviations sells short and one below VWAP - 2 deviations buys.
  3. The position is closed when a one-second close reaches the VWAP, or at 15:45 New York.
  4. A new entry may follow an exit the same day.
Decides on
1-minute bars
Holds
Nothing overnight
Parameters
bands 2, open minute 570, exit minute 945, session minute 1,080

Results by year

It made money in 2 of 10 years on ES. Its best year was 2019, +$6,197, and its worst 2022, −$25,812. The account was closed out on April 9, 2025, so the years stop there.

Net after costs by year on ES, with trades, win rate, largest drawdown and buying and holding the same contract
YearNetTradesWin rateDrawdownBuy and hold
2016−$3,62728762%−$4,529+$11,497
2017−$4,75031161%−$6,741+$21,936
2018−$16,21833363%−$18,066−$8,589
2019+$6,19727365%−$5,161+$35,436
2020+$4,63822355%−$19,360+$27,761
2021−$14,12727257%−$19,724+$51,361
2022−$25,81231462%−$34,649−$45,939
2023−$5,42633065%−$17,801+$38,435
2024−$15,14932765%−$21,130+$42,947
2025, Jan to Apr−$25,3228864%−$29,706+$37,010

VWAP reversion backtest results

On ES it lost $20,938 before costs, then paid $78,660 in fees and slippage.

Net = before costs − fees − slippage
ES: −$20,938 − $9,710 − $68,950 = −$99,597
Every figure of the VWAP reversion backtest, by contract
FigureESNQ
Net after costs−$99,597−$99,047
Before costs−$20,937−$36,715
Fees$9,710$9,217
Slippage$68,950$53,115
Trades2,7582,649
Trades a year290312
Win rate62%64%
Profit factor0.870.91
Average trade−$36−$37
Average trade, 95% interval−$70 to −$4−$88 to $12
Sharpe ratio-0.70-0.51
Largest drawdown−$104,232−$100,677
Time in the market8%9%
Buying and holding the contract+$236,212+$401,004

The whole interval is under zero: after costs, the average trade lost money.

VWAP reversion code and Pine Script

The strategy as the backtest ran it, and the Pine Script our compiler wrote from it for TradingView.

VWAP reversion
vwap_reversion.py
1import math2 3from atlantive import BUY, NEW_YORK, SELL, Bars, Session, Strategy4 5 6class VwapReversion(Strategy):7    """Fade a one-minute close two standard deviations from the session VWAP, New York day session.8 9    The VWAP is the Globex session's, from its 18:00 New York open: the volume-weighted mean of each10    one-minute bar's typical price (high + low + close) / 3, with the volume-weighted standard11    deviation of that price around it. In the day session, a one-minute close above VWAP + 212    deviations sells short and one below VWAP - 2 deviations buys. The position is closed when a13    one-second close reaches the VWAP (as of the last finished minute), or at 15:45. A new entry14    may follow an exit the same day. The clock is one-second bars; the minutes are built from them.15    """16 17    data = "bars"18 19    class Params:20        bands: float = 2.021        open_minute: int = 57022        exit_minute: int = 94523        session_minute: int = 108024        clip: int = 125 26    class State:27        clock: Session(zone=NEW_YORK)28        bars: Bars(interval_ns=1_000_000_000)29        session_key: int30        origin: float31        sum_v: float32        sum_pv: float33        sum_ppv: float34        vwap: float35        dev: float36        ready: bool37        bar_key: int38        bar_high: float39        bar_low: float40        bar_close: float41        bar_volume: float42        bar_minute: int43        bar_session: int44 45    def on_event(self, view, ev, orders):46        self.clock.update(ev)47        if not self.bars.update(ev):48            return49        minute = self.clock.prev_minute50        second = self.clock.prev_second51        day = self.clock.prev_day52        high = float(self.bars.prev_high)53        low = float(self.bars.prev_low)54        close = float(self.bars.prev_close)55        volume = float(self.bars.prev_volume)56        key = day * 1440 + minute57        session = day58        if minute >= self.p.session_minute:59            session = day + 160        if key != self.bar_key:61            if self.bar_key != 0:62                # The minute bar_key closed: fold it into its session's VWAP, then judge its close.63                if self.bar_session != self.session_key:64                    self.session_key = self.bar_session65                    self.origin = self.bar_close66                    self.sum_v = 0.067                    self.sum_pv = 0.068                    self.sum_ppv = 0.069                    self.ready = False70                typical = (self.bar_high + self.bar_low + self.bar_close) / 3.0 - self.origin71                self.sum_v += self.bar_volume72                self.sum_pv += self.bar_volume * typical73                self.sum_ppv += self.bar_volume * typical * typical74                if self.sum_v > 0:75                    mean = self.sum_pv / self.sum_v76                    var = self.sum_ppv / self.sum_v - mean * mean77                    self.vwap = self.origin + mean78                    self.dev = 0.079                    if var > 0:80                        self.dev = math.sqrt(var)81                    self.ready = True82                in_session = self.bar_minute >= self.p.open_minute and self.bar_minute + 1 < self.p.exit_minute83                if self.ready and in_session and view.position == 0 and self.dev > 0:84                    if self.bar_close > self.vwap + self.p.bands * self.dev:85                        orders.ioc(side=SELL, qty=self.p.clip)86                    elif self.bar_close < self.vwap - self.p.bands * self.dev:87                        orders.ioc(side=BUY, qty=self.p.clip)88            self.bar_key = key89            self.bar_minute = minute90            self.bar_session = session91            self.bar_high = high92            self.bar_low = low93            self.bar_volume = volume94        else:95            self.bar_high = max(self.bar_high, high)96            self.bar_low = min(self.bar_low, low)97            self.bar_volume += volume98        self.bar_close = close99        # Exits on every one-second close.100        late = second + 1 >= self.p.exit_minute * 60 or minute < self.p.open_minute101        if view.position < 0:102            if close <= self.vwap or late:103                orders.ioc(side=BUY, qty=self.p.clip)104        elif view.position > 0:105            if close >= self.vwap or late:106                orders.ioc(side=SELL, qty=self.p.clip)
1//@version=62// VwapReversion, exported from Atlantive.3// Chart simulation: decisions use completed bars; market orders fill at the next bar open.4// Native fills, fees, data gaps and session boundaries can differ. Re-test on the target symbol.5// Use standard time-based candles and the strategy's timeframe, not synthetic chart prices.6// Pine v6 trims old trade records at its order limit; the strategy keeps calculating.7// Margin is explicitly disabled to retain the native signal simulation's sizing.8strategy("VwapReversion", overlay=true, initial_capital=100000.0, commission_type=strategy.commission.cash_per_contract, commission_value=1.68, default_qty_type=strategy.fixed, default_qty_value=1, calc_on_every_tick=false, process_orders_on_close=false, slippage=0, margin_long=0, margin_short=0)9 10if timeframe.in_seconds() != 111    runtime.error("VwapReversion was exported for the 1 second timeframe; set the chart to 1S.")12 13bands = input.float(2.0, "bands")14open_minute = input.int(570, "open_minute", minval=-9007199254740991, maxval=9007199254740991)15exit_minute = input.int(945, "exit_minute", minval=-9007199254740991, maxval=9007199254740991)16session_minute = input.int(1080, "session_minute", minval=-9007199254740991, maxval=9007199254740991)17clip = input.int(1, "clip", minval=-9007199254740991, maxval=9007199254740991)18 19_q_safe_int(x) =>20    if math.abs(x) > 900719925474099121        runtime.error("an integer left the exact chart range; this run cannot be compared faithfully")22    int(x)23_q_int(x) => _q_safe_int(x < 0 ? math.ceil(x) : math.floor(x))24_q_iadd(a, b) => _q_safe_int(a + b)25_q_imul(a, b) =>26    if a != 0 and math.abs(b) > 9007199254740991 / math.abs(a)27        runtime.error("an integer multiplication left the exact chart range; this run cannot be compared faithfully")28    _q_safe_int(a * b)29 30_q_high = _q_int(math.round(high * 1000000000.0))31_q_low = _q_int(math.round(low * 1000000000.0))32_q_close = _q_int(math.round(close * 1000000000.0))33// Set to the native instrument quantity step; futures contracts usually use 1.34_q_volume_unit = input.float(1.0, "Volume units per quantity step", minval=0.000000001)35_q_volume = _q_int(math.round(volume / _q_volume_unit))36 37var int session_key = 038var float origin = 0.039var float sum_v = 0.040var float sum_pv = 0.041var float sum_ppv = 0.042var float vwap = 0.043var float dev = 0.044var bool ready = false45var int bar_key = 046var float bar_high = 0.047var float bar_low = 0.048var float bar_close = 0.049var float bar_volume = 0.050var int bar_minute = 051var int bar_session = 052// clock: completed chart bar's opening timestamp (America/New_York).53_q_clock_prev_minute = hour(time, "America/New_York") * 60 + minute(time, "America/New_York")54_q_clock_prev_second = _q_clock_prev_minute * 60 + second(time, "America/New_York")55_q_clock_prev_weekday = (dayofweek(time, "America/New_York") + 5) % 756_q_clock_prev_day = int(timestamp("UTC", year(time, "America/New_York"), month(time, "America/New_York"), dayofmonth(time, "America/New_York"), 0, 0) / 86400000)57// Current clock: this chart bar's scheduled close, known without a future bar.58// A native next event after a data gap can carry a later time.59_q_clock_minute = hour(time_close, "America/New_York") * 60 + minute(time_close, "America/New_York")60_q_clock_second = _q_clock_minute * 60 + second(time_close, "America/New_York")61_q_clock_weekday = (dayofweek(time_close, "America/New_York") + 5) % 762_q_clock_day = int(timestamp("UTC", year(time_close, "America/New_York"), month(time_close, "America/New_York"), dayofmonth(time_close, "America/New_York"), 0, 0) / 86400000)63_q_clock_seen = 164int minute_l = na65int second_l = na66int day = na67float high_l = na68float low_l = na69float close_l = na70float volume_l = na71int key = na72int session_l = na73float typical = na74float mean = na75float var_l = na76bool in_session = false77bool late = false78 79minute_l := _q_int(_q_clock_prev_minute)80second_l := _q_int(_q_clock_prev_second)81day := _q_int(_q_clock_prev_day)82high_l := _q_high * 1.083low_l := _q_low * 1.084close_l := _q_close * 1.085volume_l := _q_volume * 1.086key := _q_int(_q_iadd(_q_imul(day, 1440), minute_l))87session_l := _q_int(day)88if minute_l >= session_minute89    session_l := _q_int(_q_iadd(day, 1))90if key != bar_key91    if bar_key != 092        if bar_session != session_key93            session_key := _q_int(bar_session)94            origin := bar_close95            sum_v := 0.096            sum_pv := 0.097            sum_ppv := 0.098            ready := false99        typical := (bar_high + bar_low + bar_close * 1.0) / 3.0 - origin100        sum_v := sum_v + bar_volume101        sum_pv := sum_pv + bar_volume * typical102        sum_ppv := sum_ppv + bar_volume * typical * typical103        if sum_v > 0104            mean := (sum_pv * 1.0) / sum_v105            var_l := (sum_ppv * 1.0) / sum_v - mean * mean106            vwap := origin + mean107            dev := 0.0108            if var_l > 0109                dev := math.sqrt(var_l)110            ready := true111        in_session := bar_minute >= open_minute and _q_iadd(bar_minute, 1) < exit_minute112        if ready and in_session and _q_int(strategy.position_size) == 0 and dev > 0113            if bar_close > vwap + bands * dev114                _q_qty1 = _q_int(clip)115                if _q_qty1 <= 0116                    runtime.error("market order quantity must be a positive whole unit")117                _q_pos1 = _q_int(strategy.position_size)118                if _q_pos1 < 0 or _q_pos1 > 0 and _q_qty1 < _q_pos1119                    runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")120                strategy.order("q1", strategy.short, _q_qty1)121            else if bar_close < vwap - bands * dev122                _q_qty2 = _q_int(clip)123                if _q_qty2 <= 0124                    runtime.error("market order quantity must be a positive whole unit")125                _q_pos2 = _q_int(strategy.position_size)126                if _q_pos2 > 0 or _q_pos2 < 0 and _q_qty2 < -_q_pos2127                    runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")128                strategy.order("q2", strategy.long, _q_qty2)129    bar_key := _q_int(key)130    bar_minute := _q_int(minute_l)131    bar_session := _q_int(session_l)132    bar_high := high_l133    bar_low := low_l134    bar_volume := volume_l135else136    bar_high := math.max(bar_high, high_l)137    bar_low := math.min(bar_low, low_l)138    bar_volume := bar_volume + volume_l139bar_close := close_l140late := _q_iadd(second_l, 1) >= _q_imul(exit_minute, 60) or minute_l < open_minute141if _q_int(strategy.position_size) < 0142    if close_l <= vwap or late143        _q_qty3 = _q_int(clip)144        if _q_qty3 <= 0145            runtime.error("market order quantity must be a positive whole unit")146        _q_pos3 = _q_int(strategy.position_size)147        if _q_pos3 > 0 or _q_pos3 < 0 and _q_qty3 < -_q_pos3148            runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")149        strategy.order("q3", strategy.long, _q_qty3)150else if _q_int(strategy.position_size) > 0151    if close_l >= vwap or late152        _q_qty4 = _q_int(clip)153        if _q_qty4 <= 0154            runtime.error("market order quantity must be a positive whole unit")155        _q_pos4 = _q_int(strategy.position_size)156        if _q_pos4 < 0 or _q_pos4 > 0 and _q_qty4 < _q_pos4157            runtime.error("this order would add to or partially reduce a position, so its trade ledger would not be faithful")158        strategy.order("q4", strategy.short, _q_qty4)

How it was tested

The published rules, never tuned on the data, with real costs on every fill. Every strategy here is tested the same way.

Market data
CME Globex 1-minute bars, Jan 2, 2016 to Jul 31, 2026
Rules
As published, with the published parameters, never tuned on this data
Size
1 contract on every trade
Contract
The most traded one, rolled as the market moved
Fees
The broker's $0.50 a side, plus CME's and the NFA's fees when the window opens
Slippage
1 tick a side on ES, 2 in NQ's regular session and 3 outside it, 1 on other markets, where it was not measured
Stops and entries
Tested on 1-second closes and filled at the next bar
Contracts traded
ES: 43 contracts, ESH6 to ESU6NQ: 43 contracts, NQH6 to NQU6
Costs paid
ES: $9,710 in fees, $68,950 in slippageNQ: $9,217 in fees, $53,115 in slippage
Warm-up
None: each contract reads only its own bars
Flagged days
ES: NoneNQ: None Days the data holds too few bars for. Flagged, not removed.

Could you pass
a prop firm challenge
with the VWAP reversion strategy?

A backtest says whether a strategy makes money. Our AI runs its trades through the rules of every prop firm account, then tells you which one it would pass and at what size.

No card needed. Free credits every month.

  • Most money: Blue Guardian Futures Direct 50K, $5,084
  • Best payout odds: Funded Futures Family Straight to Funded 50K, 98%
  • Fastest pass: Phidias Express to Live 50K, 11 days
Blue Guardian Futures Direct 50K
Most money

$5,084

Median net across 2,000 simulated runs

Evaluation
None
First payout
97%
Fees paid
$1,521

Median net across 2,000 simulated runs

Best 5%$13,143Worst 5%−$412
Day 060120180
Example: ES opening range breakout

Questions about the VWAP reversion strategy

What is the VWAP reversion strategy?

Trade back to the session's VWAP when price has stretched 2 standard deviations away from it. Atlantive backtested it on CME futures from January 2016 to July 2026 after costs.

Is the VWAP reversion strategy profitable on ES futures?

Not in this backtest: on ES futures from January 2016 to April 2025 it lost $99,597 after costs, over 2,758 trades, positive in 2 of 10 years; on NQ it lost $99,047. Buying and holding one ES contract over the same window made $236,212. The average trade was −$36, its 95% interval −$70 to −$4. These are hypothetical results of a backtest, not of trading.

What are the VWAP reversion strategy's rules?

The VWAP is the Globex session's, from its 18:00 New York open, over each one-minute bar's typical price (high + low + close) / 3, with the volume-weighted standard deviation of that price around it. In the day session, a one-minute close above VWAP + 2 deviations sells short and one below VWAP - 2 deviations buys. The position is closed when a one-second close reaches the VWAP, or at 15:45 New York. A new entry may follow an exit the same day. Its parameters are the published ones, never tuned on the data it was tested on.

How was the VWAP reversion strategy backtested?

On CME Globex data from January 2016 to July 2026, one contract, on 1-minute bars, rolled to the most traded contract as the market moved, with the broker's and CME's fees and the slippage measured on each market. Stops and entries were tested on 1-second closes and filled at the next bar. Nothing was fitted to the data.

Can I run the VWAP reversion strategy in TradingView?

Yes. The page carries its Pine Script, written by Atlantive's compiler from the same code the backtest ran, for TradingView's strategy tester. A chart's fills and fees can differ from the backtest's, so test it again on the symbol you trade.

Can I change the VWAP reversion strategy and test it again?

Yes. Rerun it in Atlantive: a new chat opens with the strategy's rules in the box, for you to read, change and send. The agent writes the code, backtests it on the market and window you name, and reports the result with its costs.

Rerun it in Atlantive. Change the market, the dates or a rule, then send it.

A new chat opens with this in the box

Backtest a VWAP reversion strategy on ES, 2016-01-02 to 2026-07-31, one contract on a $100,000 account with default costs: VWAP from the 18:00 New York session open with 2-standard-deviation bands on one-minute typical prices; in the day session fade a one-minute close outside the bands, exit when price reaches the VWAP (one-second closes) or at 15:45.

Never sent for you

May we measure this visit?

With your consent, Google Analytics and PostHog count visits to these pages. You can separately allow Google Ads to measure visits from our ads. Visit measurement starts only after you choose. Cookie policy