Originally Posted by
MountainBear
Ping time can increase for multiple reasons not related to distance. One obvious confounding factor is interference of some type.
They're talking about differences of a few milliseconds over the course of a flight, so the only way I could see interference affecting it would be to add noise to the signal that would make identifying those signals harder and increase the error bounds on the timings; I'm not sure exactly where the delay is measured, but, for example, a noise spike could presumably cause misidentification of the first bit, if that's the point used to determine delay.
So there's no 100% certainty here, but if the delays are consistent with an aircraft moving over time, particularly one moving at a constant speed, and the delays on earlier flights are consistent with position reports on those flights, then we can be pretty sure they're good positions. Bigger issues are probably things like transmitter synchronization drift over the course of a flight, which, again, can be checked by looking at earlier flights of the same aircraft.
Edit: just saw your post above mine about low-level multipath reflections: yeah, there certainly could be oddities like that which would affect the timings. I'm guessing no-one's going to offer a 777 to fly low-level over the ocean to provide calibration data.