Read Daxx's post. The random sampling is going to be horribly inaccurate until some time goes by. Actually, if it doesn't record until 1000 cards are played without a refresh, it'll be (even more) skewed to what the more serious players grind with. Stats for what's used in PvP would interest me more.