Practically every modern receiver works fine under 700 mph(edit: apparently the COCOM limits are approximately 1,200 mph). But the data can be tampered with, and if you’re not clever about how you do it there will be some discrepancies. Presumably a company validating their data would be a relatively trustworthy way to catch the tampering. There are a couple people on Strava (a fitness app) that doctor their GPS data to shave a few seconds of their times and take the top spot on the leaderboard. Sometimes they get caught because of the discrepancies.

This is really about Shelby getting caught with a doctored video and lying about a fake record but the author is too nice to call them out on it.

Any measurement inaccuracy or calibration issue here is going to be <1mph anyway...

It could never explain what was seen. Either someone foolishly read km/h numbers as mph, or the whole thing is faked. Considering the silly animation of the camera view on the video, I suspect the latter...

It is very hard for me to understand what they're referring to with discussion of calibration. They would presumably be using some type of augmented GPS/differential GPS which could require calibrated ground stations depending on accuracy targets, but the error you'd get from using unaugmented GPS or a low-setup augmented GPS like WAAS wouldn't be large enough to account for the errors discussed here. A cheap consumer GPS receiver like the one in your phone shouldn't produce the errors in question over a meaningful distance outside of some really fringe situations.

One of the specific claims is that the GPS equipment doesn't show a speed of zero when the vehicle is not moving - it's hard to understand how this would happen, a poorly set up DGPS scheme would not produce a 20mph rate of change in fixes for a non-moving object unless it was seriously defective.

All in all it feels a lot like the "calibration" here is just a hand-waving excuse for the outcome - the margin here is so large that calibration issues are not going to account for it unless there was outright incompetence.

The vendor involved, DEWE, makes various types of augmented GPS equipment but also inertial measurement products. Calibration would be a much bigger issue with inertial measurement and that could make more sense, but all the reporting on this says GPS was in use.

They probably use some RTK / dGPS setup for increased accuracy, which does need calibration.