VWAP: RTH vs ETH Session, the Kiss-Back Re-Test, and Whether Mean-Reversion Actually Holds

RTH vs ETH: Which Session Should You Anchor To?

VWAP resets at the start of each session and calculates a running volume-weighted average from that anchor point. The question of which session to use — Regular Trading Hours (RTH, typically 9:30am–4:00pm ET for US equities) versus Extended Trading Hours (ETH, which pulls in pre- and post-market activity) — is really a question about what you're measuring.

If you trade purely during the RTH session, anchoring to RTH gives you a VWAP that reflects only liquid, primary-session participation. ETH anchoring bakes pre-market and after-hours volume into the reference level before the opening bell, which shifts that level based on activity most retail traders can't act on cleanly. Neither setting is wrong — they answer different questions. RTH asks where has the average RTH participant broken even today? ETH includes the extended-session crowd.

The 'my VWAP doesn't match your chart' problem almost always traces back here. TradingView, ThinkOrSwim, and other platforms may default to different session definitions. Before assuming the indicator is broken, check which session your chart is set to display and align it to the session you actually trade.

The Kiss-Back Re-Test: What It Is and What It Isn't

The kiss-back describes a sequence where price moves away from VWAP — say, a sharp opening rally — then pulls back to touch or slightly breach VWAP before resuming in the original direction. The narrative is that VWAP acts as a magnet and, once re-tested, confirms support or resistance. It's visually compelling and straightforward to spot in hindsight.

The structural problem is that 'touch VWAP and go' is a pattern you can draw onto almost any chart after the outcome is known. The question is whether it occurs at a rate that produces real edge over a systematic rule, or whether you're selecting the instances that worked and filtering out the ones where price sliced through VWAP and kept going. That selection bias is the same mechanism that makes nearly every mean-reversion story look better in a highlight reel than in a live account.

VWAP Mean-Reversion and the Win-Rate Trap

Mean-reversion is one of the most seductive setups in technical trading, partly because it generates high win rates that feel like proof of edge. In our database of 660,005 backtests across 903 assets, RSI Mean-Reversion — a conceptually similar 'return to center' setup — showed a median win rate of 71.7% across assets. That sounds impressive. Only 10% of those backtests beat buy-and-hold. That gap between win rate and actual edge is the trap.

The mechanics explain it: mean-reversion strategies capture many small profits as price returns toward the anchor, but they're structurally exposed to the occasions when price doesn't revert — trending moves that can erase a long string of small wins in a single trade. VWAP reversion carries the same structural shape. A price that moves far below VWAP on heavy volume may be doing so because something has changed, not despite it.

VWAP did not emerge as a top performer in our backtests across any asset class. The indicators that consistently ranked at the top for stocks — Fibonacci Pivots, Camarilla Pivots, and the Intraday Momentum Index — share VWAP's general idea of session-anchored reference levels in some cases, but they carry different signal mechanics and showed meaningfully better results.

What the Backtests Found for Stocks Instead

Across our 903-asset database, tested on 1-Hour, 4-Hour, Daily, and Weekly timeframes, the top-performing indicators for stocks were Fibonacci Pivots (top performer for 22 assets), followed by Projection Bands and the Intraday Momentum Index (16 each), and Camarilla Pivots (also 16). These are level-based and momentum-based approaches rather than pure mean-reversion setups waiting for price to snap back to a shifting average.

Fibonacci Pivots and Camarilla Pivots both establish reference levels based on prior-session data — not unlike VWAP in concept — but they define discrete price zones rather than a continuous average that shifts with every tick. A fixed level can be tested cleanly. A moving average like VWAP requires you to know exactly where it sits at the moment you're trading, and that location changes throughout the session.

None of this says VWAP has no value as contextual information. Traders use it to assess whether they're buying above or below the day's average cost basis, and institutions use it as an execution benchmark. But using it as a standalone entry trigger for mean-reversion trades is a different and much stronger claim — and it's one our data doesn't support.

These Are Hypothetical Results, Not Financial Advice

Everything on this page comes from hypothetical backtests run on historical data with realistic transaction cost assumptions. These are not real trades and past backtest performance does not predict future results. Backtests can be influenced by look-ahead bias, data-snooping, and the specific historical period tested. Nothing here is financial advice. You are responsible for your own trading decisions.

Our backtests covered 1-Hour, 4-Hour, Daily, and Weekly timeframes across 903 assets and 382 indicators — 660,005 combinations in total. We do not have results for timeframes shorter than one hour, tick charts, or range bars, and we make no claims about how any indicator behaves at those resolutions.