Dr Prius and the Gen4 Prime (bad news)

All gauges have a bias. The bias needs to be computed and taken into account to eval a trust percentage.
The only way that can be done is over time using base readings and comparing them with current readings.
Or on the devels side by reading the specs doing the math and writing the program, than comparing the apps gauge data stream to the specs for each model the app supports..

And that trust percentage is going to be different across different supported models.

Beyond that there are more ways to eval and compare reading in one app to readings in another app for consistency between apps. Than you can check and compare an apps data output to data seen on the OEM gauges over time. Consistency is the key, because a gauge isn't necessarily going to be calibrated exactly to each supported models data stream, unless the user does the calibration and double checks for inconsistencies in the gauge(s) full range, if the full range is even known, exactly..