I guess an important piece (conveniently left out of this story) is that it does not work if there are other cars on the road, as these cars quickly invalidate the data coming from the slow cart.
Which makes it all the more infuriating that Google Maps still has huge flaws when it comes to motorbike navigation.
While they could, I doubt they do... do you have evidence of this?
I would guess (unsubstantiated) that it classifies a motorcycle trip as being unexpectedly fast through many traffic jams.
For a long very straight road represented as a single edge, I would assume if you had 100 actual cars, you'd get an average of the manipulated data rate and the real flow rate.
Then do the calculations at the server for speed and such.