Fun With Numbers: Daily Trading Patterns For The S&P 500

Back in 2012 and 2013, I wrote pieces describing some interesting daily trading patterns in the S&P 500 (SPY). The effects looked big enough to matter. In particular, Mondays and Tuesdays showed outsized performance differences from other days in the week. The performance divergences were very large in the first 5 months of 2013 as shown below.

Average Daily Price Change of the S&P 500 Based On Day of Week and Year (Jan 3, 2007 - May 15, 2013)

Average Daily Price Change of the S&P 500 Based On Day of Week and Year (Jan 3, 2007 – May 15, 2013)

Thanks to questions from a reader, I decided to update this analysis. However, this time around, I started with the statistics of difference: calculating whether observed differences are very unlikely random or chance occurrences – statistical significance. I used t-tests for paired comparisons and ANOVA (the analysis of variance) for group comparisons (single factor). All price data came from Yahoo Finance for the S&P 500. (A link to the Google sheet is available upon request).

This more rigorous lens was very instructive. While the year 2013 through May 15th did indeed produce statistically significant differences in performance across the weekdays that confirmed visual inspection (ANOVA p-value of 0.02) , the year ended with a kind of reversion to the mean. By the end of 2013, there was NO statistically significant difference in performance across the weekdays (p-value of 0.12). The chart below shows how overall performance compressed. For example, Tuesday went from an average 0.50% daily performance to 0.19% and Friday became the day with the largest average daily performance.

The spread of average performance for the S&P 500 (SPY) in 2013 was not unusually wide.

The spread of average performance for the S&P 500 (SPY) in 2013 was not unusually wide.

This compression could have easily occurred from the equalizing force of smart traders and algorithms noticing the anomalous performance divergences. Whatever the reason, these dynamics highlighted a danger in acting on such patterns: the window of opportunity is short. The bigger opportunity may be in fading anomalous patterns back to more historic norms (this approach forms the basis of my strategy for trading market extremes). Of course, the parallel risk in 2013 lay in timing the end of the anomaly.

1 2
View single page >> |

Disclosure: Long SSO, long SPY calls.

Follow Dr. Duru’s commentary on financial markets via  more

How did you like this article? Let us know so we can better customize your reading experience. Users' ratings are only visible to themselves.

Comments

Leave a comment to automatically be entered into our contest to win a free Echo Show.