Any thoughts? What am I missing?
Regarding BMX, I suspect that you’re missing start/end point effects. Perhaps look at the difference of rolling one year returns to see if there’s a sustained advantage.
CBOE has a ton of option strategies for which they publish backtests. Some are fairly fancy like collars, condors etc. None of the ones that I’ve analyzed are particularly attractive.
Given that their strategies are based on heavily traded indices, perhaps this isn’t surprising - an advantage would quickly be arbitraged out.