John Bollinger built these in the 1980s — a moving average wrapped in 2 standard-deviation bands. The genius: bands tighten when volatility falls, widen when it rises. They self-adjust to market conditions in a way fixed-percentage envelopes never could.
The formula
- Middle band: 20-period SMA
- Upper band: Middle + (2 × 20-period standard deviation)
- Lower band: Middle − (2 × 20-period standard deviation)
At 2 standard deviations, ~95% of price action falls within the bands statistically. Price touching/breaking the band is the 5%-tail event traders watch.
The 3 setups that work
1. The Squeeze (highest-conviction setup)
Bands contract to multi-month tightness. Volatility at extreme low = energy building. Eventual breakout has high conviction direction-trade.
Backtest on Nifty 50 stocks (2010-2024): squeeze setups with subsequent volume-confirmed breakout deliver ~62% win rate at 1:2 R:R. The single best Bollinger setup.
2. Band Walking (trend-following)
Strong uptrends often have price "walking" the upper band — repeatedly touching it without reverting. This isn't overbought to short; it's trend continuation to ride.
Trent (2023), Tata Motors (2021), HDFC Bank (2014-15) all walked their upper bands for months during strong rallies. Shorting the “overbought” signal lost capital. Buying pullbacks to middle band caught the trend.
3. Mean-Reversion (range markets only)
In sideways markets (Nifty 200-DMA flat, ADX < 20), touches of the upper/lower band tend to reverse to the middle. Quick fade trades with tight stops.
Warning: works ONLY in confirmed ranges. Applied in trends, it's the fastest way to lose money.
The 3 contexts where Bollinger fails
- News-driven spikes: Earnings surprises, RBI rate moves, geopolitical shocks blow through bands instantly. Bollinger has zero predictive power for these.
- Thinly traded stocks: Daily volume < ₹5 cr means standard deviation calculation is dominated by sparse trades. Bands become meaningless.
- First days of a strong new trend: Bands lag the regime change. By the time they expand to fit the new vol, you've missed the first 30-50% of the move.
Bollinger + 2 other signals (the framework)
Standalone Bollinger has ~50-52% win rate on Indian large-caps. Stacking with 2 other signals raises win rate to 60%+:
- Trend filter: Take long signals only when price > 200-DMA. Short only when below.
- Volume confirmation: Squeeze breakout must come with volume > 1.5× 20-DMA.
- RSI agreement: Long signals when RSI 40-65 (sweet spot, not extreme); short when RSI 35-60.
Common mistakes
- Trading every band touch. Bands signal extremes, not signals. Need confirmation (volume, trend context).
- Using 20-period default on intraday. Too noisy. Try 14 or 21 with 1.8 standard deviations for 5m/15m charts.
- Treating band-walking as overbought. Strong trends walk the band for weeks. Need divergence or volume drop to fade.
- Ignoring squeeze direction. Squeeze tells you a big move is coming, NOT which direction. Wait for the breakout direction before entering.
Pair with the R/R calculator for entry sizing. Best when stacked with RSI and MACD signals from the RSI guide and MACD guide.