xHL did come to pass yesterday. The cQuatz looks like c1, with a small negative return probable from CO.
The quatz logic has “worked” since about the beginning of 2020. The cQuatz looks “better”. The analysis seems to show that equities, in general, follow SPY CO but then sort of do their own thing OC. Note that c0 and c1 returns and win rates are similar – CO is negative for both, almost by definition. c2 and c3 CO is positive for the same reason.
That suggests that binary decisions on buying CO have been working well if OC is negative and vice versa. The oQuatz is less clear.
The technical issue is exploiting these examples of high dimensional clustering before they inevitably start working differently.
I’m debating with myself whether to take the rest of the day off.