If you try to maintain sync, one delay on one train makes the whole system late to maintain the sync. This reduces throughput, relative to random sync, so it’s not often used in big systems. Random sync means that your connection takes 0 sec some of the time and the train leaves before your doors open some of the time.
Most (competent) subway systems’ schedules are developed this way so that at rush hour you don’t have people clogging up the platforms waiting for the next train as more people come into the system. It requires a lot of data, knowing where the demand is, how people use the system and where people generally switch trains. Tokyo subways work quite well if you ever have the opportunity to ride them. It’s an example of a system that works very well.
If you look at the clusterfuck that is the New York subway system, it’s a perfect example of a dysfunctional system where the train “schedule” is more for show than anything else – where this doesn’t work.
Latest Answers