It is very easy to come up with theoretical numbers and look nice. We already know know the subway infrastructure works. Forget what exists currently, and try t

It is very easy to come up with theoretical numbers and look nice. We already know know the subway infrastructure works. Forget what exists currently, and try to calculate the theoretical throughput of the subway system. I bet you'll come up with much more that what it is currently. On first look at your calculations the flaws I find are these: 1. 125mph is a very, very optimistic number in my opinion. In reality I think anywhere near 100mph is what I would expect. 2. You can't fit 50(!) people in a 20ft x 6.5ft 'sled' for hours. If something goes wrong, they'll trapped in a tunnel with no escape to surface for hours!!! 3. The numbers might make financial sense if you cram them with theoretical people at peak hours. But the thing is, most of the day (and night) those sled will mostly be empty. So you can only count the sleds to be used close to their limit for a couple hours a day morning and night. 4. You can't just daisy chain 33k sleds along the track. As I've said before, one malfunction would possibly wreck havoc in the system. Which can actually lead to people's death. Think of them as you would an airplane, but you can't land in an emergency like you can in an airplane 5. This is close to what @so33 said. The last couple miles and the first couple miles will be very congested because people slow down, and getting off them would take time and many stations. The only way to remove bottlenecks would be to make lots of places to easily disembark, which would increase the costs.