(08-25-2012, 10:33 AM)Poor Yorek Wrote: I suppose one might improve the analysis (were one so inclined) by normalizing each game rating by a given player to that player's "global" average. Of course, this would mean that any game/scenario's rating would be fluid as members continue to log plays (and their average changes) and would incorporate any "drift", that is, whether a given member grades more easily or harshly as more plays are logged (which would be an interesting study in and of itself).
If you want to try to put together a more meaningful algorithm than just a straight average, I can give you a data dump of scenario, rating, and userIDs to play with in Excel.
EDIT - the offer is for anyone interested, not just PY.
Fluidity of ratings is meaningless to PG-HQ. Submitting a new play & rating drops the corresponding scenario, units, maps, nations,
everything basically into the queue which is checked in 30 minutes or less. Anything in the queue gets a full rating recalc.
