Bollinger Bands provide a statistical way to compare a stock price to the recent historical prices for that stock. The Bollinger Bands formula typically looks at the 20 trading days preceding the time in question, and it gives you a range of reasonable prices for that period. The center of that range is the average price during that historical period. The size of the range is based on the volatility of the stock for that time period. If the price goes outside of this range, that is a statistically significant event. For the exact formula see http://www.bollingerbands.com.
Our software allows you to filter based on Bollinger Bands. You can create an envelope based on the standard Bollinger Bands to see only stocks which are trading within their normal trading range. Or you can set a minimum level of three standard deviations, to see only stocks which are trading far above expectations. These filters are user configurable, so the possibilities are endless.
Bollinger Bands differ from many of our other types of volatility analysis because they are based on a trading range. Bollinger analysis assumes that a price will typically move around within a range; if a stock suddenly jumps up, it is more likely to go back down than further up. Most of our analysis is based on the efficient market theorem. The efficient market theorem allows you to look at historical prices to estimate volatility, but does not allow you to predict a specific direction for the market based on historical prices.
Most of our volatility analysis is based on recent intraday data. We also work with Bright Bands, which look at volatility over a year. Bollinger Bands fill in the middle ground by looking at the closing prices for the last 20 trading days.