Key results
- Total return
- -98.6%
- Buy & hold: +137.5%
- Annual return (CAGR)
- -34.9%
- Buy & hold: +9%
- Max drawdown
- -98.6%
- Buy & hold: -42.1%
- Win rate
- 59.17%
- Profit factor
- 0.64
- Trades
- 8,943
- Longest losing streak
- 8 trades
- Period
- Oct 4, 2016 to Oct 1, 2026
- 10 years
- Market
- DAX/EUR
- 15-minute
Equity curve vs buy & hold
Returns year by year
| Year | Return | Trades |
|---|---|---|
| 2016 | -6% | 140 |
| 2017 | -33.4% | 590 |
| 2018 | -28.2% | 898 |
| 2019 | -32.9% | 808 |
| 2020 | -43.6% | 998 |
| 2021 | -33.4% | 854 |
| 2022 | -31.2% | 954 |
| 2023 | -35% | 868 |
| 2024 | -37.5% | 947 |
| 2025 | -38.3% | 1043 |
| 2026 | -29.7% | 843 |
Calendar years, compounded from monthly results. The first and last years can be partial.
VWAP, the volume-weighted average price, is a reference level for many intraday traders: it shows where the bulk of the session's volume traded. Add bands at two standard deviations and you get a popular mean reversion setup, described in guides such as this VWAP mean reversion article: when price stretches beyond a band, it is overextended relative to the session's fair value, so you fade the move and aim for a return toward VWAP. We ran that idea on the DAX for ten years, in both directions.
The rules we tested
- Market: DAX (DAX/EUR), 15-minute candles.
- Period: 4 October 2016 to 1 October 2026, ten years.
- Indicator: session-anchored VWAP with bands at 2 standard deviations.
- Long entry: the 15-minute close crosses below the lower band.
- Short entry: the 15-minute close crosses above the upper band.
- Exit: a fixed take-profit of 30 index points and a fixed stop-loss of 50 points. There is no separate exit rule; 2,404 trades closed on a strategy signal rather than the stop or target.
- Direction: long and short.
- Costs: 0.02% commission per side, full position size on each trade.
What the backtest shows
The result is close to a total loss: -98.6% over ten years, with a maximum drawdown of 98.6%. Buy and hold on the DAX returned +137% over the same window, with a 42% maximum drawdown.
The strategy traded constantly: 8,943 trades, close to 900 a year. It won 59% of them, which sounds good, but the profit factor was 0.64: the strategy won back only about 64 cents for every dollar it lost. Of all trades, 3,681 reached the target, 2,858 hit the stop and 2,404 closed on a signal. The average win was about 0.13% of the account and the average loss about 0.31%, more than twice as large. Losses did not come in long runs: the longest losing streak was only eight trades. The damage came from that lopsided payoff, repeated thousands of times.
Every calendar year finished negative, and every full year lost between 28% and 44%. 2020 was the worst at -44%, over 998 trades. Even the least bad full year, 2018, lost 28%.
Frequency multiplies every cost
Each round trip pays 0.04% of the position in commission. Across 8,943 trades, that adds up to about 358% of position value paid in fees over the decade. On a 30-point target, every fee eats a real share of each win, and the strategy repeats that hundreds of times a year.
Why it doesn't work
Two things work against the strategy, and they compound.
The first is the payoff. The target (30 points) is smaller than the stop (50 points). If every trade ended at one of the two, the break-even win rate before costs would be 62.5%. The strategy won 59%. It was already losing before paying a single commission.
The second is frequency. A 2-sigma band on a 15-minute chart is crossed very often, especially when the session trends. When the DAX trends, price does not return to VWAP; it keeps walking along the band, and the fade gets stopped out. With hundreds of signals a year, the small negative edge per trade, plus costs on every round trip, compounded into an almost complete loss.
None of this means VWAP is useless as a reference. It means this specific fade, with these exits, at this frequency, did not have an edge on the DAX.
What you could test next
- Flip the payoff: a target at least as large as the stop, or a target set at the VWAP itself.
- Use wider bands (2.5 or 3 standard deviations) so only true extremes trigger a trade.
- Add a trend filter and only fade against the band when the higher-timeframe trend is flat.
- Limit entries to one per session to cut the trade count and the commission drag.
Methodology and assumptions
- Starting capital
- 10,000
- Position size
- 100% of equity
- Commission
- 0.02% per side
- Data window
- Oct 4, 2016 to Oct 1, 2026
- Run date
- October 2, 2026
- No repainting: every signal is computed on closed candles only, so the backtest never acts on a price a trader could not have seen yet.
- Past performance does not predict future results. A backtest is a historical simulation, not a forecast.
Frequently asked questions
What is the VWAP band mean reversion strategy?
VWAP is the volume-weighted average price of the session. Bands are drawn a number of standard deviations above and below it. The mean reversion idea is that a close beyond the outer band is overextended, so you fade it and target a move back toward the VWAP.
Why did a 59% win rate lose money?
Because the target (30 points) was smaller than the stop (50 points). If every trade ended at one or the other, you would need to win about 62.5% of the time just to break even before costs. At 59%, and with commissions on 8,943 trades, the strategy lost steadily.
How much did commissions cost in this VWAP backtest?
Each round trip costs 0.04% of the position. Over 8,943 trades that adds up to about 358% of position value paid in fees across the decade, a heavy load for a strategy targeting 30 points per trade.
Does VWAP mean reversion work on other timeframes?
This test only covers the 15-minute DAX with fixed 30 and 50 point exits. Fewer, larger trades, or a target set at the VWAP itself, are the obvious variants to measure before drawing a broader conclusion.
Reproduce it, then change it
Every number on this page comes from the Backtrex engine. Rebuild the strategy in the app, then change the market, the timeframe or a parameter and see whether the result holds.
Ready to backtest your strategies?
Create your free account and take the guided tour: a real 10-year backtest in 3 minutes, no coding, no credit card.
Create your account in 30 seconds and run your first backtest. No credit card.
Other backtests
- Strategy
- +14.4%
- Buy & hold
- +136.7%
- Strategy
- +16.5%
- Buy & hold
- +136.4%
- Strategy
- +13.6%
- Buy & hold
- +136.4%
- Strategy
- +26.9%
- Buy & hold
- +528.3%