# Mean reversion

> Mean reversion is a strategy family that treats a large move away from a reference level — a moving average, a session VWAP, a prior close — as more likely to be partly retraced than continued. It assumes the reference is stable and the deviation is noise. The assumption is what fails: when a market re-rates, the reference moves to the price instead of the price returning to the reference, and every entry taken on the way down is a loss.

Canonical HTML page: https://qanterion.com/glossary/mean-reversion
Updated: 2026-08-18
Also known as: reversion to the mean, counter-trend trading, fading the move

## How it is calculated

```
A mean-reversion rule needs three numbers, not one:
  · Reference — the level being reverted to (moving average, VWAP, prior close)
  · Deviation — how far price must travel from it before entry
  · Invalidation — the move that says the reference itself has shifted

Expectancy = (Win% × Avg win) − (Loss% × Avg loss)
A 31% win rate breaks even once the average win clears ~2.2× the average loss.
```

## Worked example

At a 31% win rate, a rule averaging +1.0% per win and -0.4% per loss returns (0.31 x 1.0) - (0.69 x 0.4) = +0.04% per trade before costs.

## A low win rate is the normal shape, not a defect

Short-horizon mean reversion exits small and often, and takes its losses when a deviation keeps deviating. QANTERION publishes a 31% typical win rate on its precious-metals strategy for exactly this reason. Expectancy, not win rate, decides whether a rule pays: at 31% the average win has to clear roughly 2.2 times the average loss to break even, and that ratio is the number worth checking before the win rate.

## Trends are the failure mode

Mean reversion and momentum are not competing opinions about one market — they are bets on opposite regimes. A rule that fades every extension earns while a range holds and loses continuously once a trend establishes. This is why the regime, not the entry signal, is the thing to monitor, and why a hard drawdown cap matters more here than it does in a trend-following rule.

## How it gets misread

A high win rate is routinely read as evidence that a mean-reversion rule works. It often means the opposite: wins are being closed early and losses left to run, so the strategy is carrying an unrealised tail it has not paid for yet. The figure worth trusting is the maximum drawdown printed next to the win rate, never the win rate on its own.

## Calculate it

- Risk of Ruin Calculator: https://qanterion.com/tools/risk-of-ruin-calculator
- Position Size Calculator: https://qanterion.com/tools/position-size-calculator
- Kelly Criterion Calculator: https://qanterion.com/tools/kelly-criterion-calculator

## See also

- Momentum trading: https://qanterion.com/glossary/momentum
- Maximum drawdown: https://qanterion.com/glossary/max-drawdown
- Expectancy: https://qanterion.com/glossary/expectancy

## Sources

- [Mean Reversion in Stock Prices: Evidence and Implications](https://www.nber.org/papers/w2343) — National Bureau of Economic Research (Poterba & Summers), Working Paper 2343

## Other languages

- 简体中文: https://qanterion.com/zh/glossary/mean-reversion
- 繁體中文（台灣）: https://qanterion.com/zh-TW/glossary/mean-reversion
- 繁體中文（香港）: https://qanterion.com/zh-HK/glossary/mean-reversion
- 日本語: https://qanterion.com/ja/glossary/mean-reversion
- 한국어: https://qanterion.com/ko/glossary/mean-reversion
- Português (Brasil): https://qanterion.com/pt-BR/glossary/mean-reversion
- Türkçe: https://qanterion.com/tr/glossary/mean-reversion
- Bahasa Indonesia: https://qanterion.com/id/glossary/mean-reversion
- العربية: https://qanterion.com/ar/glossary/mean-reversion

## Risk notice

Definitions are educational. Nothing here is investment advice, and no metric described on this page predicts future results.