This page was published on the old Macroption blog on 26 July 2013 and may be outdated.
Performance by Day of Week
In the table below you can find average daily performance of S&P500 on individual weekdays over different periods in history. The end of data is always Thursday 25 July 2013.
Since start of 2013 the differences between individual days of week are relatively large, which is of course due to the small sample size (26-30, depending on the weekday). Tuesdays have been the best (you probably remember the attention financial media paid to the "Tuesdays up" streak). Fridays have been close second. The weakest day of week YTD has been Monday, which is in line with long-term historical statistics.
The differences between weekdays become negligible when using longer periods of history. Day of week alone is of course not sufficient for making trading decisions.
Volatility by Day of Week
I have also done a similar analysis of daily S&P500 volatility, measured as true range (the greatest of high less low, high less previous close, and previous close less low – see detailed explanation here).
Differences between individual weekdays are quite small even when using the shortest period (YTD). The day which stands out is Wednesday with slightly wider true range (blame the Fed).
One thing that stands out across all time periods is the smaller true range on Fridays. Another day with relatively smaller range was Monday when using the data of the last 10 or 20 years, although that is not the case for the other periods. The relatively narrower range on Mondays and Fridays is in line with my own experience with testing various trading strategies, based mostly on the trend following idea. Mondays and Fridays did consistently show weaker performance than the other weekdays on vast majority of strategies and data periods.
The most striking observation from the table above is the much smaller average true range YTD compared to all the other periods (about 2/3) – but if you are watching the markets regularly, you probably don't need the table to know that. The difference is way larger that the differences between individual days of week. The point is: the big picture (call it long-term trend, economic cycle, or whatever) matters much more than day of week.