steve wrote:
On Fri, 08 Apr 2005 15:02:43 +0000, Richard J. wrote:

If you have trains A, B, C, D etc. running to a timetabled 2-minute
frequency in the peak, and train B gets cancelled, you'll have a
4-minute gap in the service. That gap will get longer and longer
because at least twice the normal number of passengers will be
joining train C at every station, causing longer dwell times. So
not only does train C get continually delayed and overcrowded, but
trains D, E, F, etc. also get delayed as they catch up with the
slower train C.

By holding train A for one minute, producing two 3-minute gaps, you
spread the missing train B's passengers across two trains. This
means that there is a better chance of limiting the delays to
trains C, D, E, F etc. at the cost of one minute's delay to train
A.

Of course it makes the lights on the control panel look evenly
spread out. However, you failed to explain the logic of holding a
train *full* of people.

If people in a "full" train can be persuaded to move down the car a
little to allow more passengers to board, then there is a net benefit to
passengers on that line of holding the train. If you mean literally
full, i.e. crush-loaded with absolutely no room for anyone else and
people waiting on the platform for the next train, then I agree that
ideally the train should depart and be held at the next station if the
crush-loading has eased. But you would need a more flexible signalling
system than currently exists, where AFAIK only certain signals can be
held on red in this way and the rest are automatic. There is still
probably a net benefit in holding the full train, though.
--
Richard J.
(to e-mail me, swap uk and yon in address)