Good points again, but I'm actually prone to the other version of recency bias: if something DOESN'T work recently, I convince myself it's flawed and needs to be tweaked, forgetting all the months of data showing it does work over time but can have rough patches. Need to stop doing that and be...