Bollinger Bands
BB · module 24 of 33 · volatility
A moving average with bands placed a number of standard deviations above and below it.
- SMA 20
- Upper 2σ
- Lower 2σ
- ▲/▼ where the rule triggers
Bands widen with volatility. In ranges, closes back inside a band are the reversion cue; in trends price walks the band.
How to apply it
- 1. Add it. Start with 20, 2.0 on 5m (Scalping).
- 2. Read the market first. Price can walk the upper band for a long time — a band touch is not a sell signal.
- 3. Adjust for ranges. Closes back inside the band from beyond it, at a confirmed boundary, are the reversion use.
- 4. Manage the trade. Place the stop at 1.5 × ATR and the target at 3.5 × ATR, and only act when the chart above shows the trigger.
How it is calculated
Middle = SMA(20); Upper/Lower = Middle ± (k × standard deviation of closes over 20), k usually 2.
Inputs
- · Period (20)
- · Deviations (2.0)
Reading it correctly
In a trend
Price can walk the upper band for a long time — a band touch is not a sell signal.
In a range
Closes back inside the band from beyond it, at a confirmed boundary, are the reversion use.
Settings by style
| Style | Suggested values |
|---|---|
| Scalping | 20, 2.0 on 5m |
| Intraday | 20, 2.0 on 15m |
| Swing | 20, 2.0 on 4h |
These are conventional starting points, not optimised values. Test any change on your own pair and timeframe before relying on it.
Worked example
SMA(20) = 1.0880 with a standard deviation of 0.0015: upper band 1.0910, lower band 1.0850. Price at 1.0915 is beyond the upper band, meaning it is more than two standard deviations above its 20-candle mean — unusual, but by construction that happens roughly 5% of the time.
When it failed
In a strong USD/JPY uptrend, price closed above the upper band on 14 of 20 candles. Every fade lost, because in a trend the standard deviation itself rises and price rides the band.
Limitations
- Bands describe recent dispersion, not support and resistance.
- A touch is statistically expected, not exceptional.
- Assumes an approximately normal distribution, which returns do not follow at the extremes.
Works well alongside
- · ADX / DMI
- · RSI
- · Bollinger BandWidth
Redundant with
- · Keltner Channels
- · Standard deviation
- · Linear regression channel
Stacking these produces agreement by construction, which is not confirmation.
Practice exercise
Record 20 closes beyond the upper band on 15m EUR/USD. Split by ADX above or below 20 and count how many produced a 1R reversion.
Check your understanding
1. Bollinger Band width is determined by:
- · ATR
- ✓ Standard deviation of closes over the lookback
- · Volume
- · Aroon
The bands are placed k standard deviations from the moving average, so they widen as dispersion rises.
2. 'Walking the band' describes:
- · A range condition
- ✓ Price repeatedly closing at or beyond one band in a strong trend
- · A volume spike
- · A gap
In trends, price can hug one band for many candles, which is why band touches are not reversal signals.