Here are some observations to prove that the current signaling on the 4/5 prevents more than about 28 trains per hour. Observations were taken under totally normal conditions between 5:06 PM and 5:31 PM at the downtown platform at Union Square. This is a good place because (a) the moving platforms add 15 seconds of delay* and make it a potential bottleneck and (b) you can see the next train coming as soon as the previous train has started moving out. There were no major delays, and no excessive crowding.
Some major observations:
- Even the most skillful T/Os can’t bring an express train to a stop in the station in less than 80 seconds after the previous train has started to move. Out of 10 trains, 7 were in a good position to make it into the station quickly, being in the general vicinity of 23rd St when the previous train started to move (3 were half a minute or more further north). Five of the 7 T/Os achieved times of 80 to 87 seconds. Two of the 7 achieved much worse times of 102 and 107 seconds.
- Time spent stopped in the station wasn’t strongly related to platform and train crowding. It was more closely related to arbitrary events.** I don’t see how the average stopped time of 45.7 seconds could have improved by more than a few seconds if crowding had been less.
The observations:
A. Time between trains: length of time since last train started moving till this train comes to a stop (in seconds).
B. Length of time this train is stopped.
C. Comments.
A, B, C
80, 45, no local to feed the express***
85, 58, local came at worst possible moment
102, 36, local to feed the express
107, 29, no local
120, 46, no local, big BMT crowd; express well back of previous train
134, 53, very full local, express well back
87, 33, local
83, 67, no local; C/R was asked a question; door holders
147, 60, local, express well back
86, 30, local
* The 10 seconds of delay before the train doors can open is included in stopping time. The 5 seconds during which the departing train must move extremely slowly contributes to the time between trains.
** Unfavorable arbitrary events included a local pulling in at the worst possible moment (people streaming across the platform just before the C/R might have considered closing the doors), a door holder accommodating passengers coming down the stairs, and a passenger asking a question just as the C/R was ready to close the doors.
*** The platform was much more crowded if a local had come since the last express than if none had.