Reversion to the Mean

Reversion to the mean is really all about something called 'sample size'.  Put simply, sample size is a description of the portion of a set of data being considered in an analysis.  That analysis is often used to predict the nature of the larger data set or even to project into the future.  The larger the sample size, the closer the set of date approaches to the mean (average). Confused?  Consider this.

During an election, polling organizations hire people to stand outside the voting locations and ask departing voters how they voted.  These 'exit polls' are then used between the voting and final tally to estimate what the final vote will be.  As more voters answer, the exit poll is likely to become more accurate.  In other words, the exit poll trends more correct as the sample size grows.

An investor's sample size isn't exit polls or even the number of funds she's invested in. It's time. Just like each voter's choice oscillated between two candidates, the investors daily returns will bounce between big losses, small gains, small losses and big gains.  This variation in return is known as volatility, and different investments display different levels of volatility.

As more folks are queried and the exit poll grows, the exit poll will begin to reflect the likely election outcome. Similarly, with each passing year, the losses and gains offset one another to approach the average (or mean) return. For the investor, given enough time the series of ups and downs will begin to average out to the expected (or historical) market return.  The longer that investor's time horizon, the better able she is to accept volatility because time will smooth out the variations in return.

This also means that when the sample begins to depart from average (i.e. the market is on a “hot streak”), it is increasingly likely that it will correct back to average.  Meaning a bull market with high returns will likely moderate over time with flat or negative returns and a bear market with low or negative returns will likely be followed by better times.  Some recent examples include the dot com price crash after the 1990s technology stock surge or the strong recovery in the stock market in the years following the big losses associated with the global economic crisis of 2008-2009.

Deep dive here: http://johncbogle.com/speeches/JCB_Morningstar_6-02.pdf