[jcranmer]

The actual practice behind metrorail and PRT operation, not to mention the actual operation of the system here, puts some lie to the theory you're working under.

When you design a system so that vehicles have to shunt to and from a station, you strongly limit the capacity of the system by that shunting system. Unbranched metrorail lines can reach 45TPH in normal operation, while branched lines do worse, and systems with heavy reverse branching (splitting from one line to rejoin another, NY's and DC's systems do this something fierce) struggle to get 24TPH. PRTs get better headways because they're smaller vehicles, but they're not really any better than a standard highway lane in the best case (which is beaten out by just about any other rapid transit system in existence).

If you look at the station design here, it's phenomenally bad from a throughput perspective. They appear to be using low sedans (which mean people take longer to get in and out of their seats compared to typical trainsets or buses). Furthermore, the parking spots are angled parking, which requires the cars to back into the travel lane to exit the station, backing up even traffic trying to "express" through the station. There are honest-to-god traffic jams on this system at even moderate capacity events. (Compare this with subway stations, where the station being at crush capacity utilization leads to trains skipping the station until it clears out; that's harder to do in this system.)