How to Set Up Bollinger Bands
Bollinger Bands have two parameters: the period of the average and the number of standard deviations. Instead of advice like 'use 20 and 2', we measured what each setting actually does to signal frequency.
How often price closes outside the band
- Period 10, deviation 1.5 — 25.9% of bars
- Period 10, deviation 2 — 7.9% of bars
- Period 10, deviation 2.5 — 1.6% of bars
- Period 20, deviation 1.5 — 28.7% of bars
- Period 20, deviation 2 — 10.8% of bars
- Period 20, deviation 2.5 — 3.2% of bars
- Period 50, deviation 1.5 — 30.5% of bars
- Period 50, deviation 2 — 11.9% of bars
- Period 50, deviation 2.5 — 3.4% of bars
Measured over 518 daily bars for each of 12 currency pairs; medians across pairs are shown. The figures describe how often events occur, not profitability — those are different questions. The data is open and the calculation reproduces from a script in the repository.
Bollinger Bands settings: what deviation changes
Deviation sets the frequency of events, and the dependence is steep. At two deviations roughly a tenth of bars close outside the band, at 2.5 only a few percent, at 1.5 almost a third. The difference between 2 and 2.5 changes the number of signals more than threefold — turning the strategy into an entirely different one in terms of trade frequency and costs.
What the period changes
Far less than is commonly assumed. Moving from a period of 10 to 50 at the same deviation raises the share of band exits only slightly, by a few percentage points. The period affects sensitivity rather than frequency: a shorter average adapts faster to changed volatility, so the bands contract sooner. Choose the period from your trade horizon, not from a desire for more or fewer signals.
The theoretical 5% and the real 10%
Textbooks say about 5 percent of observations should fall outside two sigma. Our measurement produced roughly double that — around 10 to 11 percent at a period of 20. This is not a calculation error but a known property of market data: the distribution of returns has heavier tails than the normal one. The practical consequence matters — an exit beyond the band is not a rare event, and building logic on 'this is an anomaly, therefore it will revert' does not hold.
Choosing for your task
Start from the trade frequency you need. For rare entries at genuinely extreme deviations, 2.5 works — but signals will be a handful per year and there is nowhere to gather statistics. For regular work the standard two deviations are sensible. A setting of 1.5 produces too many entries to be covered by costs and serves better as an indicator of market state than as a trigger.
This material is educational and is not individual investment advice. Trading forex carries the risk of losing capital.
Frequently asked questions
Which deviation should I use for Bollinger Bands?
It depends on the trade frequency you need. At two deviations about 10% of bars close outside the band, at 2.5 about 3%, at 1.5 almost a third. The setting changes signal count severalfold.
Does the period affect signal frequency?
Weakly. Moving from 10 to 50 raises the share of band exits by only a few percentage points. The period determines how fast the bands adapt to volatility, not how many signals appear.
Why do 10% of bars exit two sigma rather than 5%?
Because the distribution of market returns has heavier tails than the normal distribution. An exit beyond the band is not a rare event and cannot be treated as an anomaly.
New backtest studies — to your inbox
When a new reproducible study comes out or the data is updated, we send a short email. No spam, no "signals", no selling your address.
By leaving your email you agree to receive occasional emails from the portal. Unsubscribe in one click from any of them.