Mean reversion is the opposite bet to a breakout: instead of following a stretch, it fades it. It tends to win more often than it loses, with small gains, and its danger is the rare trade where the stretch keeps going. This guide explains how the bands frame mean reversion, complete rules, worked examples on Nifty futures and a stock, the filters that matter most, the arithmetic of a high win rate, option versions, costs, and how to build, test and run the strategy in QuantPower.
At a glance
| Bollinger Band mean reversion | |
|---|---|
| Style | Mean reversion |
| Instruments | Index futures, liquid stocks |
| Timeframe | 5-minute to daily candles |
| Indicator | 20-period moving average with bands at 2 standard deviations |
| Entry | A close below the lower band, then a close back inside |
| Target | The middle band |
| Stop | Beyond the recent low, or a multiple of ATR |
| Market | Ranges; avoid strong trends and news days |
| Win rate | Usually above half, with small average gains |
A short history of mean reversion trading
The idea that extremes tend to be followed by more ordinary values is older than markets. In 1886 the statistician Francis Galton described "regression towards mediocrity" in the heights of parents and children: unusually tall parents tend to have children closer to the average.4 Applied to prices, the idea says that a stock that has moved unusually far from its recent average is more likely to move back than to keep going, at least over short horizons.
Academic studies gave the idea support in the 1980s and 1990s. Werner De Bondt and Richard Thaler found in 1985 that US stocks with the worst returns over several years tended to do better over the following years than the previous winners, and Narasimhan Jegadeesh found in 1990 that stock returns over a month tended to reverse the next month.56 Traders built short-term reversal and pairs strategies on the same observation.
John Bollinger's bands, set in the early 1980s at standard deviations from a 20-period average, gave the idea a visible boundary: a close outside the band marked a stretch.1 Bollinger himself warned that a close outside a band is not a signal by itself, and that in a trend prices can "walk the band" for a long time. The rules below are built around that warning. In India, Bollinger Bands and the Average Directional Index became standard on retail charting platforms, and range trading on Nifty and Bank Nifty futures with band-based rules became common. Arth.live charts, inside QuantPower, include Bollinger Bands, %B, Bandwidth and ADX.
How the bands frame mean reversion
The middle band is a 20-period moving average. The upper and lower bands sit 2 standard deviations away. If prices were normally distributed around the average, about 95% of closes would fall inside the bands. Real prices are not normally distributed, but the bands still show when a move is stretched relative to recent history.
A mean reversion trader reads a close below the lower band as an overreaction and expects the price to return toward the middle band. The strategy works when two things are true: the market has a stable average to return to, and the stretch came from a temporary imbalance rather than new information.
%b as a stretch gauge
%b puts the stretch in one number: 0 is the lower band, 0.5 the middle and 1 the upper band. A close at −0.1 is below the lower band by a tenth of the band's width. Some mean reversion rules use %b directly: buy when %b crosses back above 0 after a close below it, and exit when it reaches 0.5. The rule is the same as the band version, written in one indicator.
The rules
- Range filter. Trade only when the market is not trending strongly, for example when ADX is below 20, or when the middle band is flat.
- Setup. A candle closes below the lower band.
- Entry trigger. The next candle closes back inside the band. This avoids buying while the price is still falling.
- Entry. Buy at the next price.
- Target. Exit at the middle band.
- Stop. Place the stop below the setup candle's low, or at a multiple of ATR below entry.
- Time stop. Exit if the target is not reached within a set number of candles.
The mirror rules apply for selling after a close above the upper band.
Settings for mean reversion
The standard 20-period average with 2-standard-deviation bands is a reasonable start. Two changes are common for mean reversion:
- Wider bands, such as 2.5 standard deviations. A close outside them is rarer and more stretched, so the fade has more room; the cost is fewer setups.
- A shorter average, such as 10 periods, for faster intraday trading. The middle band then sits closer to the price, the targets are smaller and the setups more frequent.
Whatever setting you test, keep it fixed across instruments on the same timeframe, and prefer a rule whose results change little when the setting moves a step either way.
A worked example on Nifty futures
This is a hypothetical example, not a result.
- On a 15-minute chart of Nifty futures, the middle band is flat at 25,050 and the lower band is at 24,980. ADX is 16, showing no strong trend.
- A candle closes at 24,968, below the lower band.
- The next candle closes at 24,990, back inside.
- The order buys one lot of 65 at the next price, 24,992. The stop is below the setup candle's low of 24,955, a risk of 37 × 65 = ₹2,405.
- The target is the middle band. The price reaches 25,048 two hours later, where the trade exits for a gain of 56 × 65 = ₹3,640 before costs of about ₹968 for the round trip at the rates from 1 April 2026.2
After costs, the trade made about ₹2,670, a little over its risk. The costs took about a quarter of the gross gain. That share is typical of intraday mean reversion on index futures, and it is why the strategy needs a high win rate to pay.
A worked example on a stock
Again hypothetical. A large, liquid stock has traded between ₹960 and ₹1,060 for three months. On the daily chart the middle band is flat at ₹1,010 and the lower band at ₹972; ADX is 14.
- Setup. After a weak market day, the stock closes at ₹966, below the lower band, on no company news.
- Trigger. The next day it closes at ₹981, back inside the band.
- Entry. 100 shares are bought for delivery at the next open, ₹983.
- Stop. Below the setup low of ₹958, at ₹955, a risk of ₹28 per share, ₹2,800.
- Target. The middle band, which has stayed near ₹1,010. The stock reaches it five sessions later, and the position is sold at ₹1,012: a gain of ₹2,900 before delivery charges of about ₹270 for the round trip.
Delivery trades carry STT on both the buy and the sell, and the overnight gap is part of the risk: a bad result or a downgrade can open the stock well below the stop.
A short-side example
The rules work in reverse after a close above the upper band. Hypothetically, on a 15-minute chart of Nifty futures in a flat session, the middle band is at 25,120 and the upper band at 25,190; ADX is 15.
- Setup. A candle closes at 25,204, above the upper band, after a quick rally with no news.
- Trigger. The next candle closes at 25,182, back inside.
- Entry. One lot is sold at the next price, 25,180.
- Stop. Above the setup candle's high of 25,215, at 25,218: a risk of 38 points, ₹2,470.
- Target. The middle band, now at 25,122. It fills an hour later: a gain of 58 points, ₹3,770, before costs.
Short trades in index futures carry the same risks as long ones, with one extra: news that lifts the market tends to arrive as gaps at the open, and a short held overnight can open far above its stop.
Confirming the stretch with RSI
Many traders require a second sign that a move is stretched before fading it. The Relative Strength Index is the usual choice: a close below the lower band with a 14-period RSI under 30, or above the upper band with RSI over 70. The two indicators measure different things, the band distance in volatility terms and RSI in momentum terms, so requiring both removes some setups where only one is stretched. As always, test the combined rule against the plain one; a second filter is worth keeping only if it removes more losing trades than winning ones.
The filters that matter most
Avoid strong trends
The biggest danger is a trend day. When the market falls steadily, the price can "walk the band", closing below the lower band again and again. Buying each close below the band in a downtrend repeatedly catches a falling market. A range filter keeps the strategy out of these days. Three common filters:
- ADX below 20, which marks a weak trend.
- A flat middle band, for example a change of less than 0.2% over the last ten candles.
- Price inside a longer-term range, for example between the highest and lowest close of the last 60 sessions.
Avoid news-driven moves
A gap or a sharp move on news, such as results or policy announcements, can change the fair price. Mean reversion assumes the move was an overreaction; news moves often are not. Many traders skip the strategy on result days for a stock, on RBI policy days and on Budget day for the index.
Require a turn
Waiting for the close back inside the band costs some of the move but avoids many losing trades. The first close outside the band is often not the last.
Watch Bandwidth
A close outside the bands after a long squeeze is more likely to be the start of a breakout than a stretch to fade. If Bandwidth is near a multi-month low, the setup belongs to the breakout strategy, not this one.
The arithmetic of a high win rate
Mean reversion strategies often win more often than they lose, which feels comfortable and can hide a weak edge. Suppose a rule wins 65% of the time with an average gain of 1 times its risk, and loses 35% of the time with an average loss of 1.3 times its risk, because some stops slip and some trends run through them. The expected result per trade is 0.65 × 1.0 − 0.35 × 1.3 = 0.195 times the risk, before costs.
That is a thin edge. If costs are a quarter of a typical gain, as in the Nifty example, the edge nearly disappears. Two habits protect it: keeping losses close to their planned size, which means never moving a stop, and avoiding the rare trend day, which is where the losses larger than 1.3 times the risk come from. A backtest that reports only the win rate hides all of this; read the average loss next to it. A mean reversion rule's worst trades decide whether it works.
Scaling out of the trade
The middle band is a modest target. Some traders take half the position off at the middle band and hold the rest for the opposite band, with the stop on the remaining half moved to the entry price. The first half secures the usual small gain; the second has a chance at a full swing across the range at no further risk to the entry. The cost is a smaller average gain on the trades that stop at the middle band and turn back. For index futures with a lot that cannot be split, the same idea works with two lots, or with an option position whose size can be divided.
When ranges tend to form
Ranges are more common in some conditions than others. Quiet periods before scheduled events, when the market waits for news, often trade in narrow ranges; so do sessions after a large move, as the market digests it. Index levels with heavy open interest at nearby strikes can also hold prices in a band as expiry approaches. None of this is reliable enough to trade on alone, but it explains why the strategy's results come in streaks: several good weeks while the market ranges, then losses when a trend starts.
Bollinger mean reversion vs breakout
The same bands support the opposite trade. After a squeeze, a close outside the band can start a breakout instead of a reversion. Many traders use Bandwidth to decide: low Bandwidth favours breakouts, normal Bandwidth in a flat market favours reversion. See the bollinger band breakout strategy.
| Mean reversion | Breakout | |
|---|---|---|
| Market | A range with steady bands | After a squeeze |
| Entry | Fade the stretch after a turn | Follow the break |
| Target | The middle band | Open; trail the stop |
| Win rate | Usually higher | Usually lower |
| Main danger | A trend that keeps going | Breaks that fail |
Timeframes
- 5-minute and 15-minute. Many setups a day, small targets, and costs that take a large share of each gain. The opening 15 minutes are best skipped, because the gap distorts the bands.
- Hourly. Fewer setups, larger targets; trades often last a session or two.
- Daily. A few setups a month per instrument, targets of several per cent, and overnight risk. Suits liquid stocks in long ranges.
A rule rarely works equally well on every timeframe. Test on the one you will trade.
Time stops
A reversion that does not come quickly often does not come at all. If the price has not reached the middle band within a set number of candles after entry, the stretch was probably the start of something else. A time stop closes the trade at that point, usually at a small loss or a small gain. On 15-minute candles, eight to twelve candles, two to three hours, is a common limit; on daily candles, five to ten sessions. Intraday trades also close by 3:15 pm whatever the bands say. The time stop is the least exciting rule in the set and often the one that saves the most money.
Bands and VWAP together
On intraday charts, the day's volume-weighted average price is another level prices tend to return to. A close back inside the lower Bollinger Band while the price is also below VWAP gives two reasons to expect a move up, and VWAP becomes a natural first target if it sits below the middle band. On a day when the price stays above VWAP and keeps tagging the upper band, the session is trending, and fading it is a fight with the day's buyers.
Trading the reversion with options
A mean reversion view can be traded with options in two common ways:
- A long call after a close back inside the lower band. The loss is limited to the premium, but time decay works against a slow return to the middle band.
- A bull put spread, which sells a put below the recent low and buys a lower put. It earns a credit if the index simply stays above the short put, which suits a view that the stretch is over even if the full return to the middle band does not come.
Option legs in QuantPower can carry their stop in underlying points, so an option trade can still be closed when the index breaks the setup low.
Mean reversion between two prices
The same idea can be applied to the difference between two related prices rather than to one price alone. If two stocks in the same sector usually move together, the spread between them, or the ratio of their prices, tends to revert to its average. Bollinger Bands drawn on that ratio mark when it is stretched: buy the cheap stock and sell the dear one when the ratio closes outside a band, and close both when it returns to the middle. This is pairs trading, a field of its own with its own statistics; a cointegration test checks whether the two prices really share a long-run relationship before any money is put on it.
Position sizing and risk
Size each trade by the stop distance, using the position size calculator, and cap the number of trades per day. Mean reversion rules can fire several times on a volatile day, each time against the same move. A cap of two or three trades a day, or a rule to stop after two losses, limits the damage on the day the range breaks.
What it costs to trade
A round trip in one Nifty futures lot costs about ₹968 for the intraday example above, most of it STT on the sale. Delivery trades in stocks pay STT on both sides. Slippage matters too: a reversal entry often comes in a fast market, and the fill can be a point or two worse than the trigger. On small intraday targets, costs and slippage together can take a third of the gross gain.
Building the strategy in QuantPower
QuantPower runs rules on your own broker account, through the two static IPs included with every account.
- Set up the chart. Add Bollinger Bands, %B and ADX to the Arth chart in the app.
- Test the rule. QuantPower's rule backtest reads up to 2,000 past candles of one instrument from your broker, from 5-minute to daily. Write the entry as "close crosses above the lower Bollinger Band" with "ADX below 20", set the stop in per cent, and add the exit rule "close crosses above the middle band". Slippage and charges are deducted from every trade. See backtesting.
- Create the signal. In Algos → Signal algos, set the same conditions on closed candles. A signal can combine up to 20 conditions in ALL and ANY groups.
- One-touch or auto. In one-touch mode the order is ready when the signal fires and you place it with one touch; in auto mode, once armed, the signal places it itself, with every risk check. A signal that places its own order costs no tokens.
- Set the stop and target. Place them as a two-leg OCO GTT order: the stop below the setup low and the target near the middle band. The first to trigger cancels the other.
- Forward test. Point the signal at the Forward test account and run it on live prices with test money. See forward test.
Account safety above the rules
Smart Exit sets a profit target, a stop-loss and a trailing step for the whole account; on a hit it stops the algos and squares off. The daily loss limit refuses new orders once today's loss reaches it, which is a natural cap for a mean reversion rule on a trend day. The kill switch stops a segment until its next session.
Backtesting tips
- Include trend periods. A test over a calm year only will flatter the strategy.
- Look at the largest losses, not just the win rate. They come from trend days and gaps.
- Count costs and slippage on every trade; small targets make them decisive.
- Test the filter on its own. Compare the rule with and without ADX before adding anything else.
- Check how many trades the rule takes. Too few, and the result is luck; too many on one day, and the cap matters more than the entry.
Reviewing the trades
Keep a record of each trade: the setup candle's distance outside the band, the ADX at entry, the time of day, the result in multiples of the risk, and whether the exit was the target, the stop or the time stop. After forty or fifty trades, look at the losers together. If most came after 2 pm, or with ADX above 18, or on days with news, the filter can be tightened. If they are spread evenly, the rule's edge is probably as thin as its results say.
Who the strategy suits
Bollinger mean reversion suits a disciplined trader who is comfortable taking many small gains and accepting an occasional larger loss without moving a stop. It suits intraday traders who can watch the session or automate it, and swing traders in liquid stocks that trade in long ranges. It does not suit anyone who averages down on a losing fade, which is how a range trade turns into a trend loss.
Common mistakes
- Buying the first close outside the band without waiting for the turn.
- Trading through a trend because the setups keep appearing.
- Moving the stop when the price nears it, hoping for the return.
- Ignoring news. A stretch on results is often a new price, not an overreaction.
- Treating a high win rate as safety. The losses decide the result.
- Skipping the time stop. A fade that drags on for hours is usually a trend in the making.
- Raising size after a winning streak. Mean reversion wins come in streaks while the market ranges; the next trend arrives when the size is largest.
- Fading illiquid stocks. A thin order book can push a price outside the bands for reasons that have nothing to do with a stretch, and the exit fills poorly.
Where Bollinger mean reversion is going
Short-term mean reversion has become more crowded as more traders use the same bands on the same instruments, and the easy edges on the most watched charts have narrowed. The rules that last tend to be stricter about when not to trade: trend filters, news filters and caps on trades per day. Scanners that check many stocks for a stretch at once, and signal algos that act only on closed candles, make those filters easier to apply consistently. Under the retail algo framework in force since 1 April 2026, such rules run through brokers as registered algos.3 The idea underneath, that an unusual stretch tends to ease, is as old as statistics; what keeps it working is knowing when the stretch is not unusual at all.
Start with 500 free tokens
Connect your broker, build a strategy and test it on paper before a single live order goes out.
Frequently asked questions
Is touching the lower Bollinger Band a buy signal?
Not on its own. In a downtrend the price can stay at or below the lower band for days. Combine it with a range filter and a confirmed turn.
Which timeframe suits Bollinger mean reversion?
Intraday traders often use 5-minute or 15-minute candles on index futures; swing traders use daily candles on liquid stocks. The rules are the same, but test them on the timeframe you trade.
Why does Bollinger mean reversion fail in trends?
Because the middle band moves with the trend. In a strong downtrend the price keeps closing below the lower band, and each entry is a bet against the trend.
How many trades a day should a mean reversion rule take?
There is no right number, but a cap protects the account on trend days, when the setups keep appearing. Many traders cap it at two or three trades a day, or stop after two losses.
Related guides
- RSI mean reversion strategy: momentum extremes instead of volatility bands
- Opening range breakout strategy: the intraday breakout alternative
- Bollinger Bands: the indicator, %B and Bandwidth, step by step
References
- John Bollinger, *Bollinger on Bollinger Bands* (McGraw-Hill, 2001)
- Union Budget 2026-27: securities transaction tax on futures and options raised from 1 April 2026 (indiabudget.gov.in)
- SEBI, circular on safer participation of retail investors in algorithmic trading, 4 February 2025, in force for every broker from 1 April 2026 (sebi.gov.in)
- Francis Galton, "Regression towards Mediocrity in Hereditary Stature", *Journal of the Anthropological Institute* 15 (1886)
- Werner F. M. De Bondt and Richard Thaler, "Does the Stock Market Overreact?", *Journal of Finance* 40(3) (1985)
- Narasimhan Jegadeesh, "Evidence of Predictable Behavior of Security Returns", *Journal of Finance* 45(3) (1990)