October is the bad boy of the stock market. The Panic of 1907, the Crash of 1929, Black Monday in 1987.

It's notable for another reason, too. The performance of Standard & Poor's 500-stock index from July 31 to Oct. 31 has a curious way of predicting the winner of the presidential election. 

As with every prediction, take it with a giant grain of salt. But the pattern is solid, as shown in the chart above by Sam Stovall, equity strategist for S&P Global Market Intelligence.

Should the S&P 500 record a positive return from July 31 to October 31, it signals the re-election of the party in power, while a decline suggests replacement. The S&P ended September slightly below its July close, so the election results are at the mercy of the market's October performance.

Two times the pattern didn't hold were in 1968 and 1980, when influential third-party candidates were in race, including George Wallace, who took about 14 percent of the popular vote in '68. 

The pattern also failed in 1956, which Stovall says could be attributed to geopolitical events putting the markets on edge. That was the year of the Suez Crisis and the Hungarian Uprising, he noted.  

Stovall describes the stock market right now as vulnerable to outside shocks because it is "top-heavy," with a price-to-earning ratio of 25.3 for the trailing 12 months. The last, and only other, time he's seen a higher market P/E was during the tech bubble, when it was almost 32. 

For those who need a "virtual Valium" for stock market jitters, Stovall pointed out that in November and December of election years the stock market is typically up, regardless of the winning party.

If the incumbent's party is re-elected, the market goes up 1.7 percent, on average, and rises in price 70 percent of the time. If the incumbent party loses, the market rises 2.3 percent and increases in price 75 percent of the time.

Nice statistic. But perhaps not enough to comfort the losers.