The obvious way to estimate the difference would be to assume Uber isn't actually lying about the average fare/trip length/speed. The "math he's done" involved

The obvious way to estimate the difference would be to assume Uber isn't actually lying about the average fare/trip length/speed. The "math he's done" involved arbitrarily cutting the average fare in half based on a guess that Uber is lying somehow to juke the stats. My guess would be that they're not and the better numbers they get are merely the result of differences between how people use Uber and how they use cabs.