This is all well and good, but how does this methodology eliminate the unknown variable which is battery condition?
If advancedpower showed current consumption - you would be on to something - but it doesn't. Since it shows percentage it is using an assumption that battery capacity is a known - which it isn't for anyone who is having these sorts of issues.
I am aware that the original poster has dismissed battery condition as a cause - and he has a decently good reason for doing so - but that is not the same as proof. Stranger behavior has been seen from a faulty / dying LiIon.