Can this have anything to do with auto-zero or temperature correction? I have bought my meter in 2013, so it should be the version that has these features. Has somebody else had similar issues, and should you during trainer rides also stop pedaling at some points to auto zero, or?
I find this happens with all 3 of the P2Maxs I have, regardless of whether it has the temperature correction feature or not. For example, the graph below shows what happens when I use a Type S P2Max in parallel with a Lemond / Wattbox, and when I use an SRM with the same Lemond / Wattbox. The issue is clearly the P2Max's offset drifting because each time the P2Max is zeroed, it returns to the same power relative to the Lemond / Wattbox. If the issue were the characteristics of the Lemond changing over time, then zeroing the P2Max wouldn't cause this shift in the relative power readings.