Mean reversion trading systems operate based on the belief that while prices trend at times, they tend to get over extended above or below the trend and then 'snap back' to some mean value. These trading systems aim to enter on an extreme and hold until this reversion to the mean occurs before exiting with a profit.
Mean reversion trading strategies can be applied to individual instruments that oscillate around some mean level, but are most often applied to pairs of instruments that move relative to each other in a predictable way – this is called a pairs trade.
As discussed here one of the most powerful ways to improve profitability in a trading program is to diversify your trading systems and take advantage of different profit drivers.
Say for example you already have systems that are long and short in the stock market. How can you add diversity if you are already trading in both directions?
One solution may be to introduce a spread trading strategy which allows you to profit from the relationship between two stocks or indices. Note that it does not then matter if that pair of stocks goes up or down.
Some pairs of stocks can tend to move in unison. If a statistical relationship between the two stocks is observed, the relationship between them should tend to oscillate around the mean level. When this relationship moves to an extreme, a position could be taken with the expectation that the relationship would return to 'mean' levels. Generally this would involve buying one instrument and shorting the other. This sort of position will profit if the relationship returns to the expected level.
This style of trading is generally nimble 'in and out', relying on some precision of entries and exits to capture a move back to the mean. It also requires a higher degree of statistical understanding and modelling than simpler strategies such as trend trading and swing trading.
This style of trading strategy can also be described as statistical arbitrage. This simply means there is some statistical relationship between the instruments that the trader has measured historically. When the relationship moves away from the statistical relationship there is an increased probability it will revert back to the mean level.
There is a significant assumption made in pairs trading that the statistical relationship between the pair of instruments will hold constant over time. This may be true for a while, but ultimately many, if not all of these relationships will break down.
If the spread between the instruments widens too far then it makes sense to cut the trade because something may have changed in the relationship between your two instruments. As always, risk management, conservative position sizing and careful monitoring will be critical to maintaining survival and profiting over the long term.
Because certain pair relationships can change quickly and dramatically, conservative position sizing is important to protect your capital in case you get caught on the wrong side of a historical pairs relationship that is breaking down.
If you can design a mean reversion system that has a high enough expectancy to be tradeable this can provide good diversification from your trend following or swing trading systems. This is because the underlying profit driver of the strategy is how the two instruments move relative to each other and not the direction of the two instruments as it is in trend following and swing trading.
As an example of how this diversification benefit works, lets say you have a long side trend following system and have several long positions in the stock market. Your spread trading system gives you a signal to go long one stock and short another stock.
Immediately afterwards the broader market undergoes a correction and your long positions all drop by 10%. Your spread position is not directly impacted as you are long one stock and short another (approximating market neutral).
During the correction the spread may revert to the statistically expected levels – giving you a profit even though your long trend following system is having a drawdown. Of course the spread position could also move against you at the same time, but the point is the reversion to the mean in the spread position is not correlated to the broader market direction. This improves the performance of your overall portfolio of trading systems.
Because the relationships between instruments changes over time and can be subject to external shocks, this strategy requires constant attention, diversification across multiple spread positions and conservative position sizing to control your exposure if there is an external shock.
Mean reversion is a trading strategy that is worth investigating if it suits your personality and skillset because it can add significant diversification benefits to your trading program.
As a result of the higher level of complexity involved, specific software (or programming skills coupled with strong knowledge of statistics) is required to test mean reversion strategies.
Experienced programmers who are into analysing complex data sets and relationships will have their own favoured tools to develop their models. Another solution is to acquire specialised pairs trading software to perform the analysis for you.
Regardless of your software solution self-education is critical to understand the intricacies of mean reversion / statistical arbitrage / pairs trading to add diversity to your trading program. There are many great resources to learn about mean reversion trading strategies.
Several outstanding books on the subject are available. In particular Ernie Chan’s book covers these concepts particularly well. I also suggest you read 'The Complete Guide to Spread Trading' which covers many different types of spread trading strategies – including strategies other than mean reversion.
I absolutely believe in the value of pairs trading and mean reversion trading systems, especially to diversify an existing trading program which has several other trading systems. Because of the higher level of complexity, however, if you are a new trader and have not previously developed your own trading systems then I strongly suggest you look at the trend trading and swing trading strategies first.
Use one of these simpler approaches to learn the basics of systems trading, trading system development and risk control. Once you have done this then come back to pairs trading and mean reversion to diversify your portfolio of trading systems.
If you do want to develop your own mean reversion trading system then I recommend you consider Van Tharp’s home study course ‘How to develop a winning trading system that fits you’.
This course will help you design a trading system that matches your beliefs, objectives and lifestyle preferences. If you do not match your system to these things your chance of success is probably pretty low, and you will also be in for a very stressful ride in the markets. On the other hand, if you do match your trading system to your beliefs, objectives and lifestyle preferences you have a greater chance of following the system, generating profits and avoiding stress in your trading.
Return to top of Mean Reversion