What Is Mean Reversion Trading and When Does It Work?

A range holds when buyers and sellers are roughly balanced over a period, with price repeatedly finding support at a lower boundary and resistance at an upper one without breaking decisively through either. This tends to happen when there is no strong new information driving the market, such as during quiet periods between major data releases or when a prior trend has exhausted itself and participants are waiting for the next catalyst.

Confirming a range rather than assuming one usually involves checking that price has tested both boundaries at least twice without a decisive close beyond either, and that momentum indicators are oscillating rather than trending in one direction. A range that has only been tested once on each side is far less reliable than one that has been tested and held multiple times.

The core of a mean reversion entry is quantifying how far price has moved from its average, since a small deviation is not tradeable and an unusually large one is the actual signal. Bollinger Bands, which plot standard deviation bands around a moving average, are a common tool for this — a touch of the outer band represents a statistically unusual distance from the recent average, given the market's own recent volatility.

VWAP standard deviation bands serve a similar purpose intraday, showing how far price has stretched from the session's volume-weighted average relative to that session's own volatility. In both cases, the measurement is relative to current conditions rather than a fixed number of points, which matters because what counts as an unusual stretch changes as volatility changes.

In a trending market, price does not oscillate around a stable average — the average itself is moving, and price can remain stretched away from any fixed reference for an extended period while the trend continues. A mean reversion trader who fades a stretched reading in a strong trend is effectively betting against the dominant flow, and can be stopped out repeatedly as the market makes new extremes rather than reverting.

The clearest evidence of this failure is a string of consecutive losing trades that all look the same: a stretched reading, an entry against it, and a continuation past the entry that stops the position out before any reversion occurs. When that pattern appears, it typically means the regime has shifted from ranging to trending, and the honest response is to stop applying the strategy until the range reasserts itself rather than increasing size to recover the losses.

Because the strategy is periodically wrong in a specific, correlated way — it loses repeatedly once a trend starts — risk controls need to account for clustered losses rather than assuming they will be independent and random. A hard rule to stop trading the strategy after two or three consecutive losses, pending a reassessment of whether the range still holds, limits the damage from a regime change.

Stops should sit beyond the level that would confirm the range has actually broken, not merely at a fixed distance from entry, since the entire premise of the trade depends on that boundary holding. If the boundary breaks, the premise is gone regardless of how far into loss the position currently sits.