Frequently Asked Questions
The model's worst drawdown since backtesting began on February 2, 2009, is -20.6%. The results table reflects a larger drawdown because the rebuy rule was implemented in September 2016 which prevents the model from getting sidelined as the market goes higher. It also prevents an automatic rebuy after a VIX spike buy signal.
The fail-safes of the VIX model were instituted during the beta phase. They are self-learning thus are dependent on data in real-time. That said, I was able to go back over all prior data and adjust the fail-safes accordingly. Even without the fail-safes, the model was able to well outperform but drawdowns were substantially greater in some cases, especially from 8/31/15 to 12/28/15.
Up to now, I have refrained from posting any backtested results as impressive as they are, since greed, a damaging type of investor psychology, tends to entice investors into oversizing their positions as the backtested results bring false confidence to a strategy that can be highly volatile.
The most important part of the backtests are the drawdowns, ie, risk since the upside gains more than take care of the drawdowns over any rolling 12-month period, and typically over most rolling 4-month periods. In other words, it usually takes less than 4 months to make back the losses even in the worst of cases.
While drawdowns beyond 10% are unusual and typically less than 20%, the model's worst drawdown was originally 26.8% due to the Brexit surprise. Note, this loss was tempered by a +24.15% profit due to a VIX Spike buy signal on 6-10-16 shortly before Brexit. Also note that since the rebuy rule was implemented, this reduced the 26.8% loss from the model's signal that occurred during Brexit down to 12.2% because the rebuy allowed for a better entry price prior to the Brexit gap down.
12-month rolling period gains are typically well above 50% and usually exceed 100% in backtests but please do not take these results as a reason to oversize your positions. These instruments can be exceptionally volatile so make sure your position sizing is aligned with your risk tolerance levels. And always remember that prior results are no guarantees of future performance.