CRTLAB Answers · measured, not opinion
Does price return to VWAP?
Answered with real trades on 11 markets of CRTLAB's own price history, 2017–2026. The rules were fixed before any result was seen, and every one of them is on this page.
CRTLAB is a browser market replay tool: smooth candle-by-candle replay, a synced second chart for SMT, footprint and volume profile. This was measured on the same price data you can replay yourself: Replay it candle by candle.
The answer
No more often than any level the same distance away. After price closed 1 standard deviation from the session VWAP in the New York session, it traded back to that VWAP on 51.9% of 23,934 days — and to a mirror level the same distance on the other side on 53.4%. Fading the 2σ band to VWAP averaged 0.00R ± 0.01 a trade.
The short answer
VWAP was reached first — before the mirror — on 41.1% of those days, and the mirror first on 42.5%. If VWAP pulled price back, the first number would be clearly bigger. It is not.
The live VWAP line, which keeps moving toward price as the session trades, was touched again on 63.1% of days. That is the number most "price always comes back to VWAP" claims are quoting: the line comes to price, as much as price comes to it.
London and New York
| Session | Stretch | Days | Back to VWAP | Mirror level | Live line touched |
|---|---|---|---|---|---|
| London 03:00–08:00 | 1σ | 24,045 | 77.3% | 78.1% | 91.8% |
| London 03:00–08:00 | 2σ | 16,177 | 68.6% | 69.1% | 87.6% |
| New York 09:30–15:00 | 1σ | 23,934 | 51.9% | 53.4% | 63.1% |
| New York 09:30–15:00 | 2σ | 14,432 | 32.6% | 34.2% | 43.8% |
Every market (New York, 1σ)
| Market | Since | Days | Back to VWAP | Mirror level | 2σ fade to VWAP |
|---|---|---|---|---|---|
| GBP/USD | 2017 | 2,339 | 46.4% | 45.6% | −0.01R ± 0.03 |
| Crude oil (WTI) | 2017 | 2,431 | 57.9% | 57.3% | −0.04R ± 0.03 |
| Silver (XAG/USD) | 2017 | 2,311 | 48.8% | 48.3% | −0.01R ± 0.04 |
| AUD/USD | 2017 | 2,370 | 42.8% | 42.8% | −0.02R ± 0.03 |
| Nasdaq 100 | 2017 | 2,445 | 66.7% | 68.2% | +0.05R ± 0.03 |
| S&P 500 | 2017 | 2,446 | 68.8% | 70.6% | −0.01R ± 0.03 |
| USD/JPY | 2017 | 2,378 | 41.2% | 43.1% | −0.01R ± 0.03 |
| EUR/USD | 2017 | 2,404 | 45.8% | 48.3% | 0.00R ± 0.03 |
| Bitcoin | 2022 | 1,109 | 54.5% | 57.3% | 0.00R ± 0.05 |
| Gold (XAU/USD) | 2017 | 2,386 | 47.8% | 51.9% | 0.00R ± 0.03 |
| Ethereum | 2022 | 1,100 | 50.0% | 55.5% | +0.01R ± 0.05 |
Does fading the VWAP bands work?
Entering at the close of the first New York candle beyond 2σ, targeting VWAP with a stop half that distance beyond the entry (a 2R target): 14,432 trades, 22.9% reached VWAP, 55.2% were stopped, average 0.00R ± 0.01. The same trade the other way — with the move, to the mirror — averaged +0.02R ± 0.01. Neither is an edge before costs.
What this means for your trading
- VWAP is where the session's volume averaged out — a fair-price reference, not a magnet.
- A band touch alone is not a reason to fade: the move kept going as often as it came back.
- A VWAP setup needs something the band does not have — a sweep, a failed breakout, a news candle — and it has to be tested on its own.
Test it yourself
NAS100 on the 5-minute chart with the session VWAP and its bands. Replay any New York session and see when price came back to it — and when it never did. The chart opens in your browser with nothing to install — 2,000 candles free, no card. Replay VWAP on NAS100.
How we measured it
CRTLAB's own 5-minute price history, 3 January 2017 to 6 November 2026, on 11 markets: Nasdaq 100, S&P 500, Gold (XAU/USD), Silver (XAG/USD), Crude oil (WTI), EUR/USD, GBP/USD, USD/JPY, AUD/USD, Bitcoin, Ethereum (Bitcoin from 2022, Ethereum from 2022). A trading day runs 18:00 to 17:00 New York and carries the date it ends on; Monday to Friday only, and a day counts only if it is complete (half-holidays are left out).
VWAP is the session VWAP as the CRTLAB chart and TradingView draw it: from the 18:00 New York open, source (high + low + close) / 3, weighted by real exchange volume (CME, Nasdaq, Eurex and Coinbase prints), with bands at the volume-weighted standard deviation. Days with volume missing on more than 5% of candles are left out.
The stretch is the first 5-minute candle in the window (London 03:00–08:00, New York 09:30–15:00 New York time) that closes more than 1σ (or 2σ) from VWAP. From the next candle to the 15:55 candle we check whether price trades through the VWAP level at that moment, and through the mirror: the same distance on the other side of the close.
The fade enters at that close toward VWAP, target VWAP, stop half the distance beyond the entry; a candle that reaches both counts as a loss; open trades close at 15:55. No spread or commission. The rules were written before any result was seen.
Frequently asked questions
Does price return to VWAP?
No more often than any level the same distance away. After price closed 1 standard deviation from the session VWAP in the New York session, it traded back to that VWAP on 51.9% of 23,934 days — and to a mirror level the same distance on the other side on 53.4%. Fading the 2σ band to VWAP averaged 0.00R ± 0.01 a trade.
What is VWAP?
The volume-weighted average price: the average price of every trade in the session, weighted by how much traded there. It starts again every session (18:00 New York for futures and forex, 00:00 UTC for crypto).
Do VWAP bands work for mean reversion?
Not on their own. Fading the 2σ band back to VWAP averaged 0.00R ± 0.01 a trade over 14,432 trades in the New York session.
The average says what usually happens. Replay tells you if your version beats it.
Replay real market history candle by candle with the future hidden, take the setup by your own rules, and every trade is scored for you — on 38 markets and every timeframe.
Replay VWAP on NAS100