Unfortunately, Richard, you are completely wrong that using the
Embarcadero diamond is faster.
[...]
Now I will tell you why Embarcadero Diamond doesn't work, even
though, theoretically, you feel it could. First, our trains do not
have cars with end doors. Each car is brought inbound with one
motorman. Back when Muni Metro began, we did have a fallback
system of operators at Embarcadero where a fresh motorman relieved
an inbound train. This was the only time in the history of Muni
Metro where we came close to running a 90-second headway in and
out. This period lasted from 1980 to 1983 or 1984, when the
fallback location was changed to Montgomery--a much inferior
arrangement. Second, we do not have one dedicated route. Instead,
we have five routes. With trains of 3 or 4 cars, trains had
multiple outbound destinations. Combining of trains then, as now,
counted on pinpoint timing of arrivals at portals--which was almost
an impossibility--and the hope that the train were successful.
Anyway, without a fallback system, each motorman has to change
ends. And then there is the "702" situation, which holds things
up. A fallback system would speed things up. But then you have
the problem of assigning a motorman to every car of the train,
instead of just one motorman, which can be done on a heavy rail
system.
While all of the above may be true to varying degrees, they are not
all true all of the time.
During peak service there is _always_ a huge queue of vehicles
entering outbound Embarcadero from MMT or MMX; often four or five
trains are visible to passengers on the platform. These vehicles load
slowly and inefficiently because they must all queue to await the one
single loading location at the west end of the platform. If some
trains could load from the inbound platform -- subject of course to an
outbound path being available ahead of some of the trains boarding
opposite, and subject of course to not causing excessive delay to
inbound service during the crossover operation -- then it would
actually speed up metro operation and throughput to operate using
Embarcadero diamond _in conjuction with_ MMT. Remember that in
close-headway urban rail operation it is dwell time, determined by
vehicle and platform throughput, which almost alone determines
headways. Signalling and speed are secondary issues; which is why
ATCS was such a criminal waste of resources. Overlapped boarding of
multiple trains increases passenger throughput, which increases system
efficiency. Turning back some trains from Embarcadero inbound via the
crossover would achieve that. (It would also slightly increase the
system-wide level of service by decreasing the non-revenue time wasted
in MMT.)
During off-peak service there is usually a free path available for
inbound trains either directly into the outbound platform or for them
to depart from the outbound platform and cross over. And if there
isn't, operation can fall back to MMT. MMT was intended to address
path contention over the diamon at times of very close headways.
Off-peak, headways are not close, and so _always_ using MMT results in
a 4+ minute non-revenue-service penalty to _every_ trip on the system,
which results in a degraded level of service system wide.
Basically, you should only use a slower and less efficient way of
running trains if operational constraints preclude the alternative.
Which means that trains should only be sent into MMT if there is no
way to turn them back using the Embarcadero crossing.
Not only are these "theoretical" scheduling arguments, they are widely
implemented on any number of rail systems. Stub terminals with tail
tracks aren't uncommon, and neither are service patterns far more
demanding than those in Muni's little trolley system.
In a future signup, the M-Oceanview is going to Proof-of-Payment,
with 2-car trains and one operator. Now, changing cabs at Caltrain
is easy. All the doors open, the train is in Cab/Street operation,
and the motorman has anywhere from 5 to 17 minutes to depart.
Changing cabs in a 2-car train in the MMT will not be that easy,
and, in fact, may be impossible. Trains will be in auto mode
entering MMT, which precludes opening a side door even when
changing cabs. Once the motorman opens the door in a location
other than at the designated platform stop, ATCS is lost. The only
solution to this dilemma is for Muni to provide a fallback
arrangement back to Embarcadero to handle this line. Back to
Square One!
I'd fix the software so that opening a door and walking along the MMT
side platforms is possible. Or I'd never have written software which
had such a bug. (Or I'd have software drive the trains into and out of
the turnback and have the operators meet them back at the platform!)
Now, then, Richard. I ask you how *you* would make this situation work.
You have my answers, which are the same as they have ever been.
Muni Metro is being operated inefficiently and wastefully, and this is
only one example of the phenomenon.