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 greed, a damaging type of investor psychology, tends to entice investors into oversizing their positions as the impressive 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.
From the start of VXX on January 30, 2009 to December 31, 2015, the worst drawdown periods (greater than -15%) are :
8/31/15 - 12/28/15 -21.8% = As of 1/8/16, the date the model went to cash, the model has rapidly recovered nearly all of this loss due to the highly profitable sell signal on 12/30/15. The date of this writing is 1/13/16.
10/24/14 - 2/26/15 -19.9% = Back to original high by 4/1/15. Stated another way, it took just over a month to recover all losses buying XIV on SELL signals and buying UVXY on BUY signals had one started investing on the 10/24/14 signal, a worst case scenario, since most members would have already been profiting from prior signals.
3/1/13 - 6/26/13 -16.3% = Back to original high by 7/31/14.
<<There were no drawdowns exceeding -15% in 2011 and 2012.>>
2/3/10 - 5/11/10 -22.7% = Back to original high by 5/24/10.
These results do not include partial or full profit taking should gains spike in a few days or less as they can, thus our suggestion of taking quick partial profits while you have them.
12-month rolling period gains are typically well above 50% and often exceed 100%, especially in 2011 which was a highly volatile year. 2016 is shaping up to be a volatile year given that it had the worst start to the year since 1932, a year when the Dow Jones Industrials fell nearly -55% peak-to-trough.
And again, 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.