We are writing this piece to explain the context of our drawdown. As a systematic investor, you are making a probabilistic bet, a gamble. Not a blind gamble like pulling the lever on a slot machine. A calculated gamble like increasing your bet at the blackjack table when the count is at 25[1]. While a good bet and one that I would make every chance I got, there is no guarantee it will pay off each time as each bet is discrete, independent and subject to the chance inherent to any gamble. Even the best card counters only estimate that they increase their raw probability of winning from 50% when betting randomly to 52% when counting.
To achieve attractive risk-adjusted returns in the long haul (a Sharpe ratio of ~1.2-1.5), we think we need to be right ~54% of the time: i.e., making money 54 days in the average 100-day window. That is the metric of our backtest[2]. So, we would expect to lose money on 46 out of 100 days. The standard deviation of that 46 is six days, meaning a two-sigma bad 100-day window is 58 down days. In our recent drawdown, it so happens that the trough of our DD was the 58th down day of the previous 100- day rolling window.
[1] This metaphor is a useful one, but there are incongruencies that we will address in a second. Also, for the non-counters, a +25 is very high and is an indication that the odds are in the players hands.
[2] Winning on 54 of 100 days might not sound very impressive, but winning on 54% of days on a gross basis has yielded winning on 64% of months and 90% of years in our track record.