I like your thread, and your analysis, but I'm not sure what this is showing. Why the reset? Doesn't this just show how quickly 10% drawdowns are reached?

SMCJB 

Why the reset? Doesn't this just show how quickly 10% drawdowns are reached?

The reset is my attempt at making the visualization of drawdown streaks possible even if we are not at all time highs (this can be seen in the 2000-2012 when the dd from all-time high count was useless).

In the dd calculations, if we make a "new" high then the value is 0 for those observations. So the days are calculated since the last 10% drawdown OR from the local next peak.

Edit:

It was also an attempt to replicate this:


I am not sure how GS calculates it but Bespoke calculates drawdowns from all time high (same as my lower panel and as can be seen it is of no use from 2000-2012):

Still long gld and tlt and added a long sh (inverse spy). I am potentially looking to add to sh due to low cross-asset vol, low cross-asset corr, spy sector cross correlation, seasonality and potential triple top in credit.