Well, the easiest counter-example is those who don't regularly torque-zero their power meters, or inadvertently zero the torque at the wrong value. In those cases the precision is unaffected but the accuracy is off. If you follow different power meter lists you'll see guys who, after an important race, discover their meters were improperly zeroed and can't figure out how to fit that ride's data in among the rest of their data, or who wonder whether their FTP changed. A couple of years ago a guy claimed he had set a new PB FTP due to his way of training. When we examined his data we could tell that he'd inadvertently re-set his torque zero. The precision was, of course, unchanged. Finally, as for the Stages, we actually know that most riders don't have fixed pedaling asymmetry. That's pretty well observed in the literature but if you think that's too rarified you can even observe it in Ray's latest data sets where he rode with four different PMs: his left-right balance changes with pedal force. The Bikeradar reviewers have noticed the same thing from the data they're concurrently collecting from a couple of different PMs. A slightly more complex example is that speed is really the result of the ratio of power to drag. You can do training to maximize the numerator and in that case you, and others, can defensibly say that even an inaccurate power meter can help you do that. However, another use for a power meter is minimizing the denominator. If that's what you're interested in doing (and, admittedly, not everyone is) then you're going to need both precision and accuracy.