If you have been involved with almost any form of technical trading you may have heard and probably used Bollinger Bands. This technical indicator named after its creator - John Bollinger- became one of the most popular tools used by traders during the 20th century. Today I am going to dedicate this post to a description and analysis of this indicator and how it can be used to create successful mechanical trading strategies. I am going to talk about the calculation of the indicator, what it tells us about price action and how we can analyze it to find different types of market inefficiencies. As always the objective of this post will be to teach you how to understand this indicator better and how to design effective systems that tie its logical relationship with price in a rational manner.
So how is the Bollinger Bands indicator calculated ? This is one of the simplest technical indicators available to modern traders. It is formed by two parallel lines plotted around an X period moving average. These lines are distanced from the MA by a certain number of standard deviations. The value of the standard deviation is calculated at the same number of periods as the X period MA. To sum it up the indicator is formed by the following line :
upper band = X period MA + Y times the standard deviation
lower band = X period MA - Y times the standard deviation
central line = X period MA
But what does it tell us ? The standard deviation is simply a measure of how much price moved away from its average. A high standard deviation implies that price moved further away from the central average while a low standard deviation implies that price remained more rangebound and close to its average. When this value is low, volatility is low, when it is high, volatility is high. The plotting of the traditional Bollinger Band indicator with the moving average and standard deviation calculated on close prices over 20 periods is shown below. The upper and lower bands are placed 2 standard deviation measures away from the central line.
-
-
The problem now seems to be to find an exploitable ineffiency using the above indicator. We must therefore find a behavior related to the standard deviation that will produce a forecast of future market movements with a positive mathematical expectancy over a long period of time. However when you attempt to do this you will find that most of your attempts will be frustrated by the fact that volatility related behavior - which is what the standard deviation of price tells us - does not appear to have a relationship with any particular inefficiency. In particular, attempting to capture volatility breakouts using Bollinger Bands is very difficult in forex trading and such a strategy does not hold a positive mathematical expectancy on most currency pairs.
Contrary to popular belief, there is also no clear statistical tendency when evaluating price "touching" the Bollinger Bands. If you attempt to create a strategy to profit from bounces from one side of the band to another you will find that profitable periods will exist but unprofitable periods in which the market will not bounce, but follow a particular band will happen. The same happens if you attempt to do the opposite. None of these approaches seems to give you a positive mathematical expectancy and in the end they do not lend themselves to the creation of mechanical strategies.
Is it not possible then to create a profitable system using this indicator ? Of course not ! Certainly there are ways in which this indicator might be exploited to give entries with positive mathematical expectancies. For example, statistically we might expect price levels outside the bollinger bands to be quite rare and in fact, significant price moves outside the bands might prove to be signals that price is moving decisively in that given direction. We could therefore use these signals to exploit both a short term retracement and a longer term trending movement, expecting price to return within the bands but to continue to move in that direction. Overall entry rules based on this approach have a positive mathematical expectancy meaning that they do provide us with a way to create Bollinger Band-based long term profitable systems.
-
