Thread

So I'm backtesting a flaky algo and suddenly realize that it works unreasonably well in some places. Craps the bed on most tradfi but kicks ass with many crypto This is by far the best test result so far but .... (2 yr hrly)