>>/74381/
yeah they are crap and that's common knowledge in the powermeter space. It's technically possible that it could be correct since the spread is quite large, some over read and some under read