[RC5] Stats suggestion

Dan Oetting dan_oetting at uswest.net
Fri Jan 5 13:22:50 EST 2001


on 1/5/01 6:36 AM, bwilson at fers.com wrote:

>Wow, great idea.  That would solve all of my issues, and would only
>require one additional field per participant/team.  Thanks also to Flower
>for suggesting the same thing, though yours was a lot easier to
>understand.
>
>It sounds like there would be a somewhat complex "catch-up" process to get
>all the averages into place.
>
>Do you have a formula handy for calculating k1 and k2 to obtain a given
>half-life?
>
>>todays_avg = k1 * yesterdays_avg + k2 * todays_count;

If todays count equals yesterdays average the average should not change:

     average = k1 * average + k2 * average
or
     k2 = 1-k1

The decayed rate after N days assuming new processing has stopped is:

     todays_average/N_days_ago_agerage  =  k1^N

so we can compute k1 for any half life as:

     k1 = 0.5^(1/N)

-- Dan O.
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list