There is a counterintuitive idea in road systems where a shortcut slows down the whole system, as any one car has no incentive to go a longer route, but if the shortcut is closed most peoples travel time will be decreased.
The same is probably true of transit systems, but is probably harder to observe, predict and model. The slowdowns are probably in transfers being slowed down by congested hallways ect.
Also when do you have too many lines? When you don't have the money to run them at good frequencies. Could you close a line and increase the frequency on other lines?
This is the logic of many small city bus systems, where there are too many bus routes, but non of them are frequent enough to be useful. Spending the same money on fewer routes would bring better service to the whole system and possibly make it self-sufficient.
