Home › Guides › Strategies

Mean reversion strategy: betting on the snap back

Stretch a rubber band and it snaps back. Mean reversion trades bet price does the same after moving far from its average. Sometimes it does. Sometimes the band is a trend, and it keeps stretching.

Beginner guide · 8-minute read · Updated 2026-10-21

How mean reversion works

"Mean" is a math word for average. Mean reversion is the idea that price tends to drift back toward its average after moving far away from it.

A mean reversion trade:

• Picks an average, like a 20-candle Moving average or VWAP.
• Waits for price to stretch unusually far from it.
• Bets on a move back toward the average.

It's the opposite mindset from trend following. Trend followers buy strength. Mean reversion traders fade it.

How "stretched" gets measured

You need a fixed rule for "far". Common ones:

• Bollinger Bands. A close outside the 2-standard-deviation band. See Bollinger Bands explained.
• RSI extremes. RSI below 30 or above 70, or a stricter 20/80. See RSI explained.
• Distance in ATR. Price more than 2 ATR away from the average.
• Distance from VWAP on an intraday chart.

Pick one. Using all four at once mostly measures the same thing four times.

The market it needs

Mean reversion works best in calm, sideways conditions.

• No strong trend on the bigger chart.
• ADX low, often under about 20.
• Moves that stretch and come back, again and again, without news behind them.

Some markets and timeframes spend more time like this than others. Big, busy markets on short intraday charts often chop back and forth. A single stock after major news often doesn't.

Where it fails

Here's the honest part, and it's a big one.

A stretched move can keep stretching. In a strong trend, price can stay far from its average for a long time. Every "it's too far, it must come back" trade loses.

That's the classic failure:

• Price drops below the lower band. You buy.
• It drops further. The band moves down with it.
• It drops again. "Even more stretched" looks even more tempting.

Traders call this catching a falling knife.

Mean reversion also struggles:

• After news. An earnings surprise or a rate decision can reset where the "fair" average is.
• When it works small and fails big. Many small wins back to the average, then one trend day that wipes them out. That pattern is common.

The stop problem

Where does the stop go on a trade that's betting against the move?

There's no perfect answer. Options traders use:

• Past the last turning point. If price makes a new low beyond it, you're out.
• A fixed distance in ATR beyond your entry.
• A time stop. If price hasn't moved back toward the average within a set number of candles, exit anyway.

What you must never do: move the stop further away because the stretch got bigger. That's how one trade becomes the whole month's losses.

The rules a beginner would write down

1. Market check: "Daily chart is not trending hard. ADX under about 20 on my trading chart."
2. Entry: "Price closes outside the lower Bollinger Band, then a candle body closes back inside it. I buy that close." (Upside down for shorts.)
3. Stop: "Just past the low of the stretch. It never moves further away."
4. Target: "The middle band (the 20-candle average). If that's less than 1.5 times my risk away, I skip it."
5. Size and limit: "About 1% risk, sized with the position size calculator. Two losses in a day and I stop. That's the falling-knife guard."

A worked example in units

A stock's 20-candle average is 60.00. Its lower band is 57.00.

• A 1-hour candle closes at 56.50, outside the band. You wait.
• The next body closes back inside at 57.20. Entry: 57.20.
• The stretch low was 56.20. Stop just below: 56.00. Risk: 1.20 a share.
• Target: the average, which has drifted to 59.50. Reward: 2.30, about 1.9 times the risk. That fits.
• Risk 1% of 10,000 units = 100. Size: 100 ÷ 1.20 = 83 shares.

If the stretch keeps going, about 100 units are lost, more with a gap. If price returns to 59.50, about 190. A math example, not a promise.

Notice the "wait for the close back inside" step. It doesn't stop the falling knife. It just avoids buying while it's still clearly falling.

How to practice it on replay

1. Open a 1-hour chart in TradingView bar replay. Add Bollinger Bands (20, 2) and ADX (14).
2. Step forward. Before each trade, write the market check result: trending or calm.
3. Take only setups that pass every rule. Write entry, stop, target and size first.
4. Step forward until the stop, the target or your time stop is hit.
5. Log it in your trading journal. Add a column: "Was the daily chart trending?"

After 30 trades, sort by that column. The split between calm days and trend days usually says more than anything else.

Questions