Anchored VWAP Strategy Backtest: Trader Dale's Weekly VWAP Pullback on 22 Years of EUR/USD

00The verdict in three lines

• Does the anchored VWAP strategy work? Not in our test: buying pullbacks to the weekly anchored VWAP after a confirmation candle (and selling the mirror) made -0.112 R a trade on EUR/USD from 2004 to 2026 taking profit before the next barrier (828 trades), and -0.087 R trailing the stop (8,541 trades), after a cost of 1 pip.
• Before costs it was barely positive (+0.015 R and +0.041 R a trade). A stop behind a 30-minute swing point is small, a median of 9 pips, so the cost took about 0.13 R from every trade.
• Buried by costs: the trailing version beat the same trades in random weeks, the barrier version did not, and both lost in every period, most of all in 2020 to 2026.

01Trader Dale's anchored VWAP rules

The rules come from Trader Dale's webinar “How to Trade With Anchored VWAP | Ultimate Guide” on YouTube (August 2025). A VWAP is the volume-weighted average price since a starting point; an anchored VWAP starts it at a point you choose instead of the session open.

• Anchor: at the start of the day, the week or the year, at a prominent swing point, or at the end of a heavy-volume rotation.
• Bias: when whole candles are above the VWAP, buyers are in control and you look for buys; below it, sells.
• Entry: on a pullback to the VWAP. Without other levels to back it up, wait for a confirmation candle: for a buy, a bullish candle that closes back above the VWAP.
• Stop: behind the nearest swing point; the next one if the nearest is too close; a fixed, volatility-based stop if the swing is too far.
• Exit: take profit just before the next barrier (a swing point, a heavy-volume zone, another VWAP), or in a trend trail the stop below each new swing point. Close before the weekend.

His favourite timeframe for intraday trading is 30 minutes, and EUR/USD is among his favourite markets, so that is what we tested. The webinar also cites testimony that large institutions execute their orders with VWAP “program trades”. Those are execution algorithms that spread a big order over the day to get the average price; that is not the same claim as prices turning at the VWAP, which is what this test measures. The webinar promotes a paid course and indicator pack; we have no connection with it.

02How we made the rules mechanical

• The anchor: the start of each trading week (the first minute after the weekend). The swing-point and heavy-volume anchors are drawn by eye in the webinar, so we did not test them. The VWAP: the average of open, high, low and close of each minute, as the webinar computes it, weighted by the tick volume (the number of price changes in the minute), accumulated from the anchor.
• Bias: bullish after a 30-minute bar whose low is above the VWAP, bearish after one whose high is below it; reset each week.
• Pullback and confirmation: with a bullish bias, a bar whose low reaches the VWAP; then, on that bar or one of the next three, a bullish bar closing above the VWAP. Entry at the next minute's open, not in the last two hours of the week.
• Stop: exactly at the low of the nearest 30-minute swing low below the entry (a bar lower than the two on each side), if at least half an ATR(14) away; otherwise the next one down; a fixed stop of 1.5 ATR if the swing is more than 3 ATR away or there is none in two weeks.
• AV-BAR: target 1 pip before the nearest barrier above: a known swing high or last week's final VWAP. No trade if that is less than 1R away. AV-TRAIL: no target; the stop moves up to each new swing low. Both close at the end of the week; one trade at a time; sells are the mirror image.

Of 19,502 pullbacks to the weekly VWAP, 12,320 were confirmed in time. AV-TRAIL traded 8,541 of them (the others came while a trade was open). AV-BAR skipped 10,963 because the next barrier was less than 1R away: with the stop behind one swing and the target before the next, the target was usually the closer of the two.

03How we tested it

EUR/USD, the 1-minute bid history of an IC Markets MetaTrader 4 demo server with its tick volume, January 2004 to 28 September 2026, built into 30-minute bars. Costs: 1 pip a round trip, as in our other EUR/USD tests, with 1.5 and 2.5 pips and two financing scenarios shown too. Results in R: the result of a trade divided by its risk, entry to stop. Periods: 2004 to 2012, 2013 to 2019, and a holdout from 2020 already used by our earlier EUR/USD tests; first-period bar t ≥ 2.5, because pullback rules have been tried several times on this pair. Control: every trade moved to another week, entered at the same time into the week, with the same stop and target distances and the same exit rule, 500 times. Engine check: 46 checks, among them hand-worked VWAP values, each rule on hand-made weeks, and an independent re-implementation that matched every one of the 12,320 signals and every trade; 15 deliberate bugs put into the engine were all caught (av_selftest_result.txt).

04Results: does the anchored VWAP strategy work?

Cumulative net R by month. * Holdout already used before this test.

Anchored VWAP strategy win rate: how the trades ended

AV-BAR won 40.9% of its trades with a median target of 1.36 R; AV-TRAIL won 20.6%, waiting for the few long runs, and its ten best trades made +308 R. Without them it lost 1,054 R. Longs and shorts both lost (AV-TRAIL: -0.082 R and -0.093 R a trade).

05What a swing stop does to the cost

A 30-minute swing point near the VWAP is rarely far from the entry: half the trades risked 9 pips or less, so 1 pip cost 0.13 R a trade. That is several times what either version made before costs. Financing A charges 2% a year of the price per day held; B uses today's IC Markets EUR/USD swap. Most trades closed within hours and none was held over a weekend, so financing changes little here.

Which stop rule was used

In most trades the nearest swing point was less than half an ATR from the entry, too close by the webinar's own logic, and the stop went to the next swing. The fixed stops were rare (77 of 8,541 trailing trades); their wider stops paid less cost in R, but they are too few to build on.

06VWAP pullback or random week?

The trailing version did better than the same trades moved to other weeks (the 99.6th percentile): the pullback and confirmation picked better moments than the same time in a random week. But the random weeks themselves lost -0.170 R a trade, so beating them still left a loss. The barrier version sat in the middle of its random sets (the 67.8th percentile).

07By period and by year

* The holdout years were already used by our earlier EUR/USD tests. Ticks a minute: the mean tick volume of the data, the weight of the VWAP.

AV-TRAIL lost in 15 of the 23 years and AV-BAR in 17. The trailing version made money in every year from 2008 to 2012 and lost overall before and after (-308 R in 2004 to 2007, -565 R from 2013); its holdout loss (-396 R, t -3.07) is the largest of the three periods.

08Does the thin tick volume matter?

On spot forex there is no exchange volume; MetaTrader and most charting tools weight the VWAP by tick volume, the number of price changes in each minute. This history has about 6 ticks a minute in 2007 and 30 to 100 from 2009 on. Because the VWAP only compares the minutes of the same week with each other, the level does not matter, but very thin weeks give a rougher average. The test answers it directly: AV-TRAIL lost -0.155 R a trade in 2004 to 2008 and -0.068 R from 2009; AV-BAR -0.061 R and -0.131 R. Both versions lost in both eras, so the thin early volume does not explain the result. A market with real exchange volume could still behave differently; we tested VWAP strategies with real volume on ES futures, where they were buried too: VWAP strategy backtest on ES.

09The gates

The data-gap check removed 3 and 12 trades open across a known gap in the data; no verdict changed. Costs, in the protocol's sense: positive before costs and negative after. For AV-BAR the amount before costs (+0.015 R a trade) is so small, and its trades so ordinary against random weeks, that cheaper execution would not have made it a proven edge either.

TWhat this shows, and what to take into your own tests

What the results show

• Trader Dale's weekly anchored VWAP pullback, made mechanical on EUR/USD 30-minute bars, lost money from 2004 to 2026 in both versions: -0.112 R a trade taking profit before the next barrier, -0.087 R trailing the stop.
• Before costs both were slightly positive (+0.015 R and +0.041 R a trade). The stops behind a 30-minute swing were small, a median of 9 pips, so a cost of 1 pip took about 0.13 R from every trade.
• The trailing version picked better moments than the same trades moved to other weeks (the 99.6th percentile), but the margin was a few hundredths of R, and it lost most in 2020 to 2026.

What remains unknown

• The anchors the webinar draws by eye: a prominent swing point, the end of a heavy-volume rotation, and the combinations with volume profile and price action. We could not code them without inventing rules, so we did not test them.
• The daily and yearly anchors, other timeframes, and markets with real exchange volume. On EUR/USD only the broker's tick count is available.
• Whether staying out around major news, which the webinar advises and we could not do without a calendar, changes the result.

Why it was buried

Both versions made a little before costs and lost after them in all three periods; neither passed the first-period test or the holdout. The trailing version beat random weeks, the barrier version did not, and taking out the ten best trades left both deeply negative. In this reading the weekly VWAP pullback did not leave enough per trade to pay for 1 pip on a 9-pip stop.

What you can apply to your own tests

• Divide your cost by your stop before anything else: on a 9-pip stop, 1 pip of spread and commission is more than a tenth of the risk on every trade.
• If a rule says "target before the next barrier" and "stop behind the nearest swing", count how often the target is less than 1R away: here it was most of the time, and those trades were skipped.
• Compare against the same trades at another time: the trailing exit did beat random weeks, which is the kind of hint worth a separate, frozen test rather than a tweak.

A note on the numbers: drawdown figures are measured on closed trades unless a table says it includes open losses. Open losses inside a trade are not counted in a closed-trade drawdown, so real account risk is higher, especially for systems with no stop. Backtests are not live trading.

FAQCommon questions

Does the anchored VWAP strategy work?

Not in our backtest on EUR/USD from 2004 to 2026: 8,541 trades at -0.087 R each with a trailing stop and 828 at -0.112 R with a target before the next barrier, after a 1-pip cost. Both lost in every period, most in 2020 to 2026.

What is an anchored VWAP?

The volume-weighted average price measured from a starting point you choose (the start of a day, week or year, or a swing point) instead of the session open. On spot forex there is no exchange volume, so charting platforms weight it by tick volume, the number of price changes.

How do you trade the anchored VWAP?

As taught in the webinar we tested: when whole candles sit above the VWAP, buy a pullback to it once a bullish candle closes back above it, with the stop behind the nearest swing point and the target just before the next barrier, or a trailing stop below each new swing; sells are the mirror. We anchored at the start of each week on 30-minute bars.

What was the anchored VWAP strategy win rate?

40.9% with a target before the next barrier (a median of about 1.4R) and 20.6% with a trailing stop, which relies on a few long runs. Neither covered the cost.

How was the anchored VWAP strategy tested?

On EUR/USD 1-minute data with tick volume from 2004 to 2026, built into 30-minute bars, with the anchor, bias, pullback, confirmation, stop and exits written down before any result, and rules written and dated before the data was run, costs included on every trade, results split into development, validation and holdout periods, a random-entry control as the benchmark, and every variant reported. Read the protocol.

10Check our arithmetic

Code provenance: our own code, written from the rules as taught in the webinar and the readings in the specification. The webinar is cited for its rules only.

11Data notes

Source. A broker's demo history of spot EUR/USD (bid prices) with its tick volume, on the New York clock; gaps longer than 30 minutes inside the week are listed in data_holes_eur.json and never filled. A week starts after any gap of more than 24 hours, so a holiday can split one. No rule depends on the time of day, so the clock offset our audit found in some weeks of 2004 to 2014 changes no signal. Fills: entries at a minute's open; stops and targets at their level, or at a minute's open if it opened beyond it; a flat 1 pip a round trip, which understates the wider spreads around news and the 17:00 rollover.

12Corrections

None since publication.

+More tests and guides

• Vwap strategy backtest
• Order block strategy
• Stochastic rsi macd strategy
• Trading costs
• Free prop firm challenge calculator
• What all the backtests found so far
• The backtesting protocol
• How to backtest a trading strategy without fooling yourself
• Backtesting glossary