Reforming the Service

Essentially the controller or signaller will pick a location where crew reliefs are booked, make a list of what trains are approaching, note down for each train how late it is and whether or not is has a crew relief, and look to see whether there are opportunities to renumber trains in such a way which by the end of it will have made the late running situation better. The key to doing this is making use of the trains which have crew reliefs, as this gives you a fresh driver who you can assume will be ready to make up any train into their own at the time scheduled.

The ideal scenario is to push all the late running onto one train, and then do something else with that train (eg turn it short, or even miss a whole round trip). You’ve essential then “righted” several trains without passengers being aware, and only one or two get hit with being turned short.

So take somewhere like Golders Green. Ideally you’d only touch stuff on the northbound as (1) that’s where the crew desk is so it can be closely monitored, and (2) you don’t change destinations for southbound passengers.

As an example, you have a sequence of trains on a 3-minute headway with a small amount of out-of-order running and a few running slightly late.

001 +6 no relief
002 +6 relief
003 +6 relief
004 +5 relief
006 +3 no relief
005 +7 relief

You could do something like

001 x 002 (departs +3 with fresh driver)
002 x 003 (departs +3 with fresh driver)
003 x 004 (departs +3 with fresh driver)
004 x 005 (departs +2 with fresh driver)
006 - 006 (departs +3, left as is)
005 x 001 (reverse Golders Green, depart in correct order, driver gets off original train and picks up new one when it arrives)

By doing the above sequence of alterations, the late running has been reduced to a point where each train is close enough to being back on time that providing it makes up a minute in the run up to Edgware and back down, they will fit into the correct order at Camden, and are likely to reach their next destinations on time. Only one train has had to be turned short. Southbound passengers should to all intents and purposes see a full service apart from one cancellation from Edgware, and once that slots in everything will head south in its right order.

Notice how I’ve had to be careful with what is done with trains 001 and 006 as these two don’t have a crew relief. I can still do things with them, but only if the booked driver gets there before their train is made up. For train 001 this is fine as the driver is already there before he takes over 005 and makes it 001. 006 is not being touched, so no problem there.

There’s no right and wrong way of doing things, and no doubt it might be possible to find a way of getting everything on time with perhaps a Colindale reverser thrown in, or perhaps leaving one train go north late and then divert via Charing Cross on its next trip, or perhaps reverse at Tooting Broadway, or whatever.

Something like the Northern is a delight as the range of options is literally endless, so you could have a batch of late running and if there’s time for the signaller to spend a few minutes thinking and planning, it can be possible to sort out 10 or so trains in one hit by using a combination of all the various options. If there’s severe late running then it’s more a case of looking at what drivers are ready to take over a train and making up accordingly, but this is more of a firefighting measure.

It does happen on Thameslink, but at the termini. As you say, you can only do things where there is a crew relief planned. On the Great Northern side you could theoretically do something at Finsbury Park on the up side with the Moorgate service - subject to every train being the same length and having crew changes there you could renumber a load of trains and then, for example, turn one at Drayton Park.

The other issue is that renumbering trains means the stock concerned may finish in a different location. Even on LU this can be a problem if a depot are expecting a particular train. Whilst control are entitled to take the view “delivering the timetable comes first”, the trouble is that this can then mean a no ok stock the next day when the train which was due an exam has ended up outstabled unexpectedly and the depot couldn’t do the exam. This is why it’s always a case of doing as little as possible to give the service the opportunity to recover naturally, rather than chasing every last minute.

As ever the question is whether it’s worth taking that ‘hit’ on one service if it avoids a load of reactionary or consequential delays. Control *should* be taking into account a range of factors, not least what the train and driver are scheduled to do on the next trip. It’s like playing snooker - it isn’t about being able to pot the current ball, but looking ahead to set up subsequent shots.

Having said that, I’ve seen some very dubious non-stop runs on the mainline, not least stops removed but the train getting stuck behind another train so ending up getting signal checked in the very platforms being skipped! (Yes Southeastern, looking at you here!)

This happens on Thameslink too.

Classic example now are the xx:16 and xx:46 Luton Rainhams when the incoming service is late into Luton; if it arrives more than a couple of minutes after it's outbound should have gone, the general reaction is non-stop it to St.Pancras. I find one of these every month or so. What then happens is one of two things [1] it stays on the slow Lines, catches up with a St.Albans all station starter by about Mill Hill, and trails that one's tail lights on single yellows all the way to Blackfriars, or, [2] it sits in platform at Luton or Harpenden for several more minutes waiting free path to get on the Fast Lines, then sits again at West Hampstead waiting to get back, often ending up behind the very same StAlbans starter it would have followed as in [1].

Anyway, that was a digression.



I was more aiming at wishing TL could do more of the LU type reforms when the entire service is late. I was not getting at sorting out just one train (which is what skipping stops does), but when the whole pattern in one service group is all late by a similar amount of minutes that approaches that service frequency i.e. every 15 min but they are are 12-14 min late. As a TL commuter, that happens so so often, been there, got more than one tee shirt.

I think that is the thing the main line delay attribution regime is overlooking the case where the delay to consecutive trains is of the same order as the service frequency - which is where reforms make more sense. LU in this sense is high frequency everywhere, even out in the sticks of Buckinghamshire and Essex. So it works.