Bollinger Bands, developed by and named for noted technical analyst John Bollinger, employ a concept frequently used in the technical analysis of securities – standard deviation. Standard deviation is, essentially, a measure of how far the price of a security diverges from its mean average. In short, it is a measure of volatility. Bollinger bands provide a sort of range trendline where the range expands or contracts in conjunction with increased or decreased volatility. They do this by measuring how far closing prices are away from a 20-period moving average.
Bollinger found that by plotting the bands at two standard deviations, both above and below the moving average, roughly 90% of all closing prices should fall within the range of the bands. (Three standard deviations should contain approximately 99% of closing prices, but two standard deviations is the standard setting for Bollinger bands.) So, Bollinger bands are constructed by an upper band line that is two standard deviations above the 20-period simple moving average, a midline that is the 20-period average, and a lower band line that is two standard deviations below the midline.
A key element of Bollinger bands is that they are not focused on the absolute price of a security – such as $50 or $55 – but on where the price is relative to the Bollinger bands. In other words, is the price closer or further away from the outside bands. It is because of the expansion and contraction of the bands with volatility. The absolute price of a security might go higher, but at the same time be lower (further away) relative to the upper Bollinger band because the bands have expanded in line with increased volatility.
The fact is illustrated in the chart below – the first sign of impending trend change is the long red (down) candlestick, roughly in the middle of the chart. Although the absolute price is a new low, the price is higher relative to the lower Bollinger band, as it is contained with the band – compared to the previous low that went below (outside) the band.