Firstly, thanks to bowroaduk for the copy of the report - speaking as the
man on the platform, it's great when we get to hear the story behind the
messes we get in; being kept in the dark really adds insult to injury.

On Tue, 7 Jun 2005, Boltar wrote:

tunnel and at 10.16 BT Police were advised. Formal Incident Management
(Na100) was declared with DSM Sparrow appointed Silver Control. A
special service was introduced west of White City and east of Holborn
to all destinations. With technical staff assessing the damage to the
points, two of the three trains 3 and 42 were authorised to work back
to Lancaster Gate, Approximately 1000 passengers were detrained to the
platform by 10.45.

Why did it take 30 mins to decide to move the trains back? Why not do
this immediately?

Because it was thought that it would be possible to keep the trains going
forward. If it had been possible to resolve the problem quickly - as on
the Victoria line the other day - then it would be hugely preferable to do
that, resulting in nothing but a five minute delay, rather than to reverse
and detrain, delaying 1500 people by however long it took them to get into
the station, get off the train, then catch another, plus another several
thousand people in the trains behind them by however long it took to do
all that.

As it happened, it wasn't possible to resolve the problem quickly, it took
until about 1016 to determine this, and people suffered a longer delay
than if they'd been detrained immediately. However, it couldn't have been
known at 0926 that this would happen. The managers took a gamble, and
lost, but it was a sensible gamble.

Also, where do you get this 30 minute figure from? The interval from 1016
to 1045? That wasn't decision-making, that was the actual doing of it -
1016 is, from the sound of it, when the managers decided that they weren't
going to be able to fix the problem quickly, and ordered the reversal of
the trains; 1045 is the time by which this was completed. I don't think
that's an outrageous length of time to get 1000 people off two trains
under these conditions.

How long does it take to examine a set of points? We're talking 1000
people here stuck in a sweatbox.

The points failed at 0926; the first train was authorised to have a go at
the points at 1001. That means it took 35 minutes to diagnose the problem,
decide what to do, work a train in and out of the siding, determine that
it hadn't been fixed, switch off the power, get someone to the points, fix
the points, get him back, switch on the power and determine that
everything seemed okay to move. I don't think that's an unreasonable
amount of time.

One thing that i do wonder about, though: immediately after the failure at
0926, we're told "Initially, three eastbound trains were stalled in
section between Lancaster Gate and Marble Arch"; we find out later on that
the first is number 33, the second is number 5, and that there's one more
train behind that (we're not told the name, but from what comes later, i'd
guess it was number 3). Now, after the first train (number 33) has got
over the points and they've broken again, and the second train (number 5)
has approached the points, we're told "This train and a further two trains
were then queued in the tunnel" (the train i'm thinking is number 3, plus
what i assume is number 42). So, there are now two trains behind number 5
- implying that between in the time between the failure and number 5's
attempt at the points, a fourth train entered the section.

Why on earth was a train allowed to enter a section of tunnel leading to
points which were known to be broken, with three trains ahead of it? I
appreciate that the station supervisor at Marble Arch thought he'd solved
the problem, but until the points were in a known good state, they should
have been treated as broken. That train could have been emptied much
faster if it had been waiting at the platform, time would have been saved,
and unnecessary suffering prevented.

Also, i'm not sure what was happening with trains further back than
Lancaster Gate, but it seems to me that it would have made sense to bring
the last train in the queue back to the station as soon as the problem was
detected. If it was resolved, it could have started going forward again,
and since the time to reach Marble Arch would have been dominated by the
time taken for the queue to empty, no time would have been lost. If it had
not, it would have made detraining that little bit quicker, plus it would
have given passengers the opportunity to leave the train instead of
waiting.

Anyway, i'm sure LU will be investigating fully.

tom

--
It's the 21st century, man - we rue _minutes_. -- Benjamin Rosenbaum