Intraday VWAP Reversion
A stock's volume-weighted average price acts as a running center of gravity for the day, and price often pulls back toward it after stretching too far away on short bursts of order flow.
Prerequisites: TWAP, VWAP & POV
The volume-weighted average price (VWAP) — every trade's price weighted by its size, running from the open through the current moment — is used across the industry as a fairness benchmark for execution. It also behaves, informally, like a magnet for price during the day: because so many large orders are explicitly instructed to trade at or near VWAP, price that strays far from it attracts execution flow trying to get back to it, and that flow itself pulls price back.
VWAP is both a benchmark and, because so much institutional order flow is instructed to track it, a self-reinforcing anchor — price that drifts unusually far above or below the running VWAP tends to attract offsetting flow that pulls it back toward that line over the course of the day.
Why VWAP pulls price back
A large fund executing a VWAP algorithm (see TWAP, VWAP & POV) slices its order across the day in proportion to expected volume, but most such algorithms also include logic that trades more aggressively when price is favorable relative to VWAP and less aggressively when it isn't. If a stock spikes well above its running VWAP on a burst of buying, a fund with a large sell order benchmarked to VWAP becomes relatively more eager to sell into that spike — its own performance versus benchmark improves by doing so — and that incremental selling nudges price back down toward the VWAP line. The reverse holds for dips below VWAP.
Worked example
A stock's running VWAP sits at $34.00 at midday. A short burst of buying pushes the last-traded price to $34.60, about 1.8% above VWAP — an unusually wide gap for a stock that typically trades within 0.5% of its running VWAP intraday. A trader running a VWAP-reversion strategy sells short at $34.60, expecting institutional sell flow benchmarked to VWAP to lean into the spike. Over the next hour, price drifts back to $34.15, and the trader covers, capturing most of the overshoot as the gap to VWAP narrows back toward its typical range.
What this means in practice
The strategy works best on liquid names with heavy institutional participation, where a large fraction of the day's volume really is executed by VWAP-style algorithms — the more of the volume is benchmark-driven, the stronger the magnet effect. It works far less reliably on names dominated by retail flow or momentum-driven participants who have no reason to trade back toward VWAP.
A stretch away from VWAP driven by real news — an intraday earnings guidance update, a rating change, a large corporate announcement — is not the same as a stretch driven by a transient order-flow imbalance, and the two look identical on a price chart in the moment. Fading a news-driven move because it deviated from VWAP is a common way to lose money running this strategy; checking for a specific catalyst before fading a wide VWAP gap is essential.
Related concepts
Practice in interviews
Further reading
- Johnson, Algorithmic Trading and DMA (ch. 3, execution benchmarks)