Shorter Backtests

July 2022

More and more traders have a suspicion that traditional option trading strategies do not perform as well as they used to.

It is not clear why this is the case. It might be that there is no free lunch after all. If too many people are trying to go for the same edge, the market automatically adjusts. But that is speculation.

Let's see what the data says. For this, we computed statistics only since 2018. Let's compare statistics for the Netzero since 2006 and since 2018. Just looking at the charts is already quite interesting: the charts since 2006 look quite smooth, while the carts since 2016 look wild. It is also interesting that even since 2018, all strategies end up in a plus and outperform the ^RUT.

Let's go through the indicators and see whether can make interesting observations.

  • While the average risk is relatively comparable, the average size nearly halved (from 41.95 to 23.62) for the shorter backtest. This is because the RUT has been raising, and risk capital of $100,000 does not have the buying power that it used to have. Deriving the risk capital from the underlying might be a better idea
  • The average profit per trade and per day is smaller for the shorter backtest, and so is the return on risk and the percentage of winners
  • While drawdowns regarding the initial risk capital are above 200% in both backtests, the max drawdown on a previous peak is much larger for the shorter backtest (-70.90%).
  • Both backtests contain very long drawdowns
  • Sharpe and Sortino are consistently above 1 for the longer backtests, while they are mainly below 1 for the shorter backtests.

To conclude, we can summarise that the runtime of a backtest has a strong impact on performance indicators. These observations are all interesting and might help you to decide whether you would like to trade such a strategy, and if so, with which settings.

Future ideas include to look further into the data itself. Can we see Netzeros being bought in the CBOE data? Did the IV for the short puts decrease over time?

If you have questions or insights to share, please reach out!