I think this is overblown. A well maintained geared bicycle drivetrain is 97.5-98.5% efficient except when cross-chained. At ~250W, this would be a ~5W difference between a wheel-off trainer and a pedal/crank/spider-based PM. No one should be satisfied with >2% difference between two power meters regardless of the position of the device.
Here is a ride of mine from hours ago. Except for one 40-second divergence in a 1h30m ride, my KICKR BIKE SHIFT and Vector 3 pedals were basically within 1W of each other.
https://analyze.dcrainmaker.com/#/publi ... 4bb7eceacf
