I am not using exponential smoothing. I experimented with it early
on but decided against using it. I just didn't have as much success
with it. Being able to catch a player in a hot streak didn't seem
to be as profitable as being able to identify a winning player (or
team) with a solid long term rating but whom may actually have
hit a temporary cold streak from which I think it is best to not
devalue their much better longer term rating. Been awhile since
I worked through this stuff so correct me if you see any glaring
errors in my thoughts on this subject.
By the way, great question!!
