[matt4077]

Read the actual paper. They seem to have used two methods to make the alternative ("counterfactual") scenario more realistic than just "remove Uber cars":

- They compare increase in congestion and Uber/Lyft activity for short road segments, and show that these two changes correlate

- They compare reality to a model that does not include ride sharing, but does model increased population etc. This (generally well-regarded model) shows lower increases in congestion in a world without ride sharing.

[hedora]

Increases in congestion at popular drop off / pick up sites are not necessarily bad. It could be that dropping people off is the new bottleneck because the rest of the road network is less congested, and now there are more trips (i.e., the road is providing more value because it is more efficient).

A study that measured door-to-door time, and the number of passenger-miles per hour would be much more useful.

I’m not sure why traffic engineers refuse to use metrics that are directly related to the quality of the experience for the people using the roads, or the economic value delivered.