Better metrics would include: - Looking at MSAs or CSAs rather than cities proper, since cities like SF are tiny and you get a distorted view by ignoring peopl

Better metrics would include: - Looking at MSAs or CSAs rather than cities proper, since cities like SF are tiny and you get a distorted view by ignoring people that work or live outside the city - Including non-commute usage data. A good transit system has people who use it even for leisure. - Equitably. Systems disproportionately used by rich or poor people are bad. - Development patterns. If real estate is cheaper near transit, that suggests the system is bad. If you pay more to be near transit, that suggests the system is valuable. - Addressable market. A sprawling metropolitan area with lots of freeways is basically never going to have a high mass transit mode share, but a system can still be ok for the people that do use it. A simpler and better metric here would just be ridership per capita per CSA, although even that misses a lot of important points.