He was excluded from the beginning, as his numbers really mess up the average.
| [reply] |
ahhh ... *_X are the corrected numbers .... got it.
You might be interested in Median and other techniques to eliminate outliers.
You could even cluster different user types ( "voters" vs "posters" )
Personally I'd like to see a XP per response level stat ( Re^n: ), but accessing the DB is outside of your possibilities.
| [reply] [d/l] [select] |
Sure, using median might help, as would grouping. Any suggestions as to the criteria? :)
I'd like to see a XP per response level stat ( Re^n: ), but accessing the DB is outside of your possibilities.
Yeah, if only i had db access. It'd be fun to write and run all those queries.
| [reply] |