[RC5] Stats suggestion
Pedro J. Cabrera
pjcabrera at mindspring.com
Sat Jan 6 15:42:42 EST 2001
> Going back to the problem about what half-life to choose for the fixed
> exp.avrg.: How about starting with a two-day half-life for one week, then
> setting it on 7 days for a month and finally setting it on 30 days (or
> whatever value d.net will choose in the end) forever? This will help you
> get an accurate value quickly and will not discourage people because it
> goes up so slowly.
> Only problem with this is it's difficult to do with new participants...
Not really. We need to add an extra field to all participants' records,
with their exp. avg. begin date. For those that are not new, this field
will have the date the exp. avg. stats began. For those that are new, it
will have their join date.
The process that computes the exp. avg. will look at this field to decide
which half-life to apply (two-day half-life for a week, 7-day for a month,
etc., just like Anne Heizmann suggested.) This way, there is no code to
treat as 'special cases,' because there are no special cases.
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5