Turnaround Tuesday: An Anomaly That is Persistently Profitable

Markets overreact on Mondays. Here's how to profit from it.

It is quite well known that the performance of the S&P 500 doesn’t happen consistently on each day of the week; there is a seasonality to daily and monthly performance.

The chart below draws on about 70 years of S&P 500 data and produces the average for every day and month of the year over that full period. September is clearly the worst month of the year.

This morning I came across a post by Alpha Algo Trading Research focusing on the Turnaround Tuesday seasonal anomaly.

The thesis revolves around a behavioural bias where markets overreact based on the news carried over the weekends.

I decided I would test the strategy on my Navigator platform and see if I achieved the same or similar results.

Here are the basic trading rules:

Buy the S&P 500 Future at the close of trading on Monday, but only if the Monday close is lower than Friday’s close. The trade is closed at the close on Tuesday.

The long-run results were strong. Getting there was another matter.

It was not all smooth sailing, as you can see. Almost from the start there was a long, protracted drawdown that lasted 3,550 days, which is almost 10 years.

This is the part that most people skip over, and this is what makes a robust strategy the hardest to stick with. I am pretty sure that the vast majority of people trading this strategy would have abandoned it long before it turned a profit.

You need to remember when you are trading a robust strategy you firstly are never 100% sure that it is robust, and also when you digest a backtest you are not suffering the indigestion from the lived drawdown. The psychology of losing money plays havoc even when your reasoning brain says this is all normal.

I also know that others showing a backtest of the same strategy would foolishly be tempted to start the backtest in 2003 to avoid the major drawdown and introduce hindsight bias.

What is really nice about the strategy is that it is in the market only 8% of the time on average, which means the capital tied to this strategy is mostly free to do other things.

To make sure that this was a real anomaly, I needed to test other days of the week with the same rule. I am happy to say that Turnaround Wednesday & Thursday both failed.

It got me thinking, why limit the strategy to the S&P 500 future, which has shorter data than the actual index?

I ran the backtest against the S&P 500 index from 1970 and achieved similar results. But in this case the drawdown was even longer at 4,712 days.

My Concluding Thoughts

The anomaly is real; the hard part is surviving long enough to see it pay off.

To include this strategy in my portfolio, I would want to see its drawdown periods be shorter; I would be looking to add a regime filter to help it along.

I will come back to this and let you know.