Originally Posted by
Silvercivic27
but if buyers are going to have remorse a year down the line because the thing they bought isn't as useful anymore, it's not a good thing.
Maybe, but lots of people aren't doing much analysis (or are doing a poor job of the analysis they do). The real issue is whether a particular buyer is one of the people for whom this kind of thing is a deal killer. That's why Ray's review (and Greg Kopecky's, and a few others) are so long and detailed -- not everyone has the same needs or checkbook balance or experience with power meters. I've said that in many ways training is the least demanding use for a power meter; comparing the average power for two power meters (or one power meter and an indoor trainer) is the least demanding test of data quality. What's important isn't how close two power meters read on average; what's important is knowing the conditions under which they differ and by how much. Once you know that you can figure out whether a particular purchase meets your particular needs. Most reviews never get to the "under what conditions and by how much" level. Those kinds of reviews are less helpful than they should be.
Originally Posted by
robbyville
But if your data has always been on the same device wouldn't analysis still be consistent event if off in the actual numbers (ie if high level power has always been off then any increase would be relative), or is there no consistency in the inacurracies?
Depends on the kinds of analysis you're doing. NP (and measures that rely on NP, like TSS) depend on actual power values. NP is already a nonlinear estimator so adding in some extra unknown nonlinearity can be a pain. Likewise, CP/W' estimation (though I'm not a huge fan of CP/W' estimation) also depends on MMP at different durations. However, if you're just starting off with training this sort of problem tends to be less important than after you've been training for a while and you get closer to your real ability limit.