I've been noticing something I think is a bit fishy when backtesting an algo that is trading TVIX. There was a 1:10 split in June 2015 that seems to have been handled correctly for all of 2015 and 2014, but somewhere in the 2012-2013 range the split is reversed. For example, first week of 2015 it's trading around 25 (post-split adjusted price), which Q agrees with. First week of 2012, Q has TVIX's price around 30,000 when I believe it should be two orders of magnitude lower, around 300. I wasn't able to find the point where the division/multiplication actually switches (if that's what happened). Is there something I'm missing?