Investment lore has a lot of sayings meant to pass long wisdom from one generation to the next, usually after some calendar or valuation pattern becomes obvious, but may not prove to be everlasting when viewed with the benefit of hindsight a few years later.

One popular calendar based prognostication is upon us now. Some call this the "Sell in May and go away indicator." Others call it the "Halloween indicator." The theory behind this market pattern is that over time, the bulk of the stock market's return is delivered from the end of October through the end of April, the most favorable six months of the year to invest.

The chart below, prepared by Plexus Asset Management, shows the long-term monthly average returns for the S&P 500 over the 60 years ending April 2009. It becomes evident that the 7.9 percent return in the favorable November through April period is over three times that of the 2.5 percent returned on average in the May through October half of the year.

One could easily conclude that selling on May Day and buying back into the market on Halloween would deliver the bulk of the returns and help avoid risk during half the year, leading to greater risk-adjusted returns. However, it's not quite that easy. As I've highlighted in previous articles, historical returns averaging data over many years may reveal tendencies, but they are rarely rock solid in their future repetition.

For example, entering the market the on October 31, 2008, would have caused a loss of about 9 percent to May 1, 2009. The supposedly favorable period did not materialize in 2008–2009 due to the panic surrounding the unfolding financial crisis. Selling on May 1 of 2009 would have caused you to miss out on large returns achieved between that date and Oct. 31, when the S&P 500 rose over 17 percent as the recovery began. Obviously, events like this shake the confidence of someone buying into this sort of mechanical methods.

Some further research has also revealed that buried in the very long-term average data, there are patterns which exist during shorter periods of time – the time in which we as investors are actually investing and reacting. Mark Hulbert, of the Hulbert Financial Digest, recently pointed out that since 2002, the return of the S&P 500 was 6.5 percent annualized on a buy and hold basis, while the timing efforts of the Halloween Indicator would have produced a lower return of 6.3 percent, albeit with lower volatility due to spending half the year out of the market.

I believe part of this skewing of the results is due to the tendency of the stock market to perform especially poor in September and October in years where there is already a negative bias unfolding. In more normal, less chaotic years, May through October half of the year may well provide better results. Unfortunately, it is hard to make that assessment in the heat of battle in any one particular year, but obviously easy to assess in hindsight.

In summary, I'm not a fan of a fixed calendar pattern system. I am in favors of rules based investment plans however. When using a system of some sort I suggest you follow the rules and not be swayed by the feel of the market. Rules based investing works, but not perfectly. It should be approached with the expectation that there will be periods of time where your rules may not work. Altering the plan, however, is likely to be a bigger mistake as you then subject yourself to emotional decision making or guessing. Also, we must acknowledge the tax consequences of a system of managing investments that creates more frequent trading activity in non-qualified accounts.

Tom Breiter is president of Breiter Capital Management, Inc., an Anna Maria based investment advisor. He can be reached at 778-1900. Some of the investment concepts highlighted in this column may carry the risk of loss of principal, and investors should determine appropriateness for their personal situation before investing. Visit www.breitercapital.com.