No better today, 1T48 spent 9 minutes at Cambridge, to the irritation of those of us waiting for it on a very cold platform at Cambridge North. (1T50 took 8 minutes despite already being very late, 1T52 was cancelled between Kings Cross and Cambridge - are these the only 3 evening services that split?).
For comparison, when the class 317s split at Letchworth in the afternoon peak, the front set would be on the move as little as two minutes after arrival. Has anyone observed what goes on at Cambridge to make the process so slow?Tonight 1T48 took 9 minutes to split at Cambridge, 1T50 took 7 minutes, 1T52 took 11. It's actually quite impressive that GN have found yet another way to be rubbish at running trains, but they seem to have managed it with their apparent inability to split 379s in a reasonable time.
Good point. Does the "splitting process" commence as soon as relevant pax have alighted? Or are incoming drivers not a part of the "splitting process" bcause they are booked off on arrival and there is potentially a delaying factor until subsequent drivers are present?For comparison, when the class 317s split at Letchworth in the afternoon peak, the front set would be on the move as little as two minutes after arrival. Has anyone observed what goes on at Cambridge to make the process so slow?
In the case of Letchworth, the driver from King's Cross remained in the front cab. Another driver got in the middle, shut the doors, split the rear unit off and gave a toot on the horn to let the front driver know that the split was complete. Then the doors on the leading unit would be released again for a few seconds and then the train departed. I assume the process of checking the rear unit was empty happened after the front had departed. There were always announcements about the split en route from Hitchin.Good point. Does the "splitting process" commence as soon as relevant pax have alighted? Or are incoming drivers not a part of the "splitting process" bcause they are booked off on arrival and there is potentially a delaying factor until subsequent drivers are present?
That's because the second driver got on at Hitchin, herded the through passengers into the front unit (and there were not very many), and closed the interior gangway doors before the train had arrived at Letchworth. I wouldn't recommend trying to do that at Cambridge.For comparison, when the class 317s split at Letchworth in the afternoon peak, the front set would be on the move as little as two minutes after arrival.
Trains arriving at Cambridge for splitting have lots of passengers, and a disproportionate number of them don't know where they are or what they are doing. That includes passengers who are in the rear 4 cars wanting to go beyond Cambridge but not realising they need to move. If the train is splitting in platform 4 at Cambridge then this results in two conflicting flows of passengers going in opposite directions between the front and rear parts of the train. At Cambridge "as soon as relevant pax have alighted" is not trivial.Does the "splitting process" commence as soon as relevant pax have alighted?
I haven't watched any splits since the class 379s were introduced, but I did watch class 387 splits, where GN used 3 drivers, one in each cab at the location of the split, plus another at the front. I don't know to what extent splitting class 379s is different from splitting class 387s, but hopefully there is a learning curve and drivers are moving along it.It's actually quite impressive that GN have found yet another way to be rubbish at running trains, but they seem to have managed it with their apparent inability to split 379s in a reasonable time.
Interesting. I never observed that because I was always in the front unit. It sounds a sensible plan. Perhaps someone can board at Royston to usher people forward? Or have staff at King's Cross to do the same. Failing that, the dwell time at Cambridge needs to be changed to reflect reality.That's because the second driver got on at Hitchin, herded the through passengers into the front unit (and there were not very many), and closed the interior gangway doors before the train had arrived at Letchworth. I wouldn't recommend trying to do that at Cambridge.
Presumably there is no scope to reduce or retime until later in the evening the current level of splitting activity in order to improve reliability?Interesting. I never observed that because I was always in the front unit. It sounds a sensible plan. Perhaps someone can board at Royston to usher people forward? Or have staff at King's Cross to do the same. Failing that, the dwell time at Cambridge needs to be changed to reflect reality.
A 12 car train from Kings Cross has to split because 8 cars is the maximum north of Cambridge.Presumably there is no scope to reduce or retime until later in the evening the current level of splitting activity in order to improve reliability?
I haven't watched any splits since the class 379s were introduced, but I did watch class 387 splits, where GN used 3 drivers, one in each cab at the location of the split, plus another at the front. I don't know to what extent splitting class 379s is different from splitting class 387s, but hopefully there is a learning curve and drivers are moving along it.
I'd never noticed that - I can see the need for the two drivers facing forwards, but why was the 'facing backwards' driver needed? Did the same happen with 365s back in the day - I do think I'd seen quite a lot of those attaching/detaching but don't recall it needing 3 drivers.
Today's splitting scores : 9, 7, 6 - so 1T52 improved over yesterday, but not the others.
Either way, punctuality at Ely was dismal as ever - 1T48 5 late (due to the time taken to split), 1T50 17 late (mostly due to being caught behind a late-running stopper, same as yesterday), 1T52 31 late (started very late from KGX). So keeping the record that none of these three key trains have got to Ely on time once since the timetable change.
Just my two cents - the supposed '6tph' on the GN inners to Gordon Hill/Hertford at peak is not really 6tph because there is far too often a cancelled train or two, causing overcrowding. If one thing breaks in the NCL, the whole next hour of Hertford & Welwyn trains is done for. And the most striking part is that with the current timetable, there's always at least 6 717s sitting in the depot through all the disruption. It's quite frankly abysmal that GTR still cannot sort it out. I don't even know if they'll be able to reliably run the 4tph off-peak when it returns in Dec 2026.
GN has well and truly been forgotten when it comes to the TSGN franchise, and I don't think that comes as a surprise. The very fact that two distinct networks were paired into one franchise still seems ridiculous to me.
I am certainly not pleased with the fact that services out of KGX that worked reliably for decades were replaced with considerably less reliable TL services. It honestly still seems like a complete afterthought, being done only as a 'why not' kind of thing. Compared to the MML the infrastructure simply cannot cope as well, and it doesn't really work with the fact that there's still some contained regional and a lot of contained commuter services. The southern MML's regional and commuter services are almost exclusively TL so it just works more fluidly. I understand it technically makes sense, with Cambridge to Gatwick being a large flow for example, but I think everyone would much rather have an actually reliable service even if that means slight inconvenience in terms of changing trains.It was always a terrible idea to introduce Thameslink to the awkward southern end of the ECML
Even this has gone worse than expected. Ridiculous numbers of 5 cars on Leeds trains to name something off the top of my head. KGX is completely at capacity which is going to cause huge headaches in the future. We are going to need a new terminus (or at least major works at KGX) in the not too distant future because Thameslink really will not benefit from even more trains through the core, but the ECML will. Effectively we've reached a point where Thameslink's idea of through trains is no longer helpful and we actually need more contained services.Clearly the new ECML timetable has been deemed more important
A great example of the fact that mixing different services like this requires a level of precision we currently cannot achieve on our railway unfortunately.It was amended to run fast from Welwyn GC to London, but has just been held in Welwyn platform 2 for five minutes to allow 1T61 and 9S65 to pass. It will now likely get tangled up with 9S65 at Finsbury Park, meaning the train’s next working will not be able to depart on time.

Friday's numbers:
- 1T48 takes 11 (!) minutes to split at Cambridge, 8 late at Ely.
- 1T50 takes 9 minutes to split, 16 late at Ely.
- 1T52 takes 12 minutes to split, 17 late at Ely.
So, 0 out of 30 on-time for these at Ely since the timetable change - and a whole 5 out of those 30 arrived less than 5 minutes late. Actually the record is worse than that, as looking at RecentTrainTimes, 1T54 and 1T56 also have the same punctuality issues (though they don't split at Cambridge).
On-time arrivals for evening peak services at Ely since the timetable change:
View attachment 195593
(chart showing no on-time arrivals for 5 consecutive evening peak trains).
It must have been quite difficult to manage to degrade the service further from last year, but this does look like this year will be even worse than the dismal results seen with the previous timetable.
The common demoninator is the split taking much longer (almost double the time) than scheduled.
That being said, this is the sort of issue that can be ironed out after a few weeks/months with proper attention paid to it by performance teams, so it is not guaranteed to be awful forever.
The thing with the splits, one would think that this should have already been been able to be improved upon, but it’s the same with the emptying out at Letchworth, GTR just don’t seem to have any capability of focussing on these sorts of issues. There is no evidence of *any* focus on these sorts of details which make the difference between the service working and not working.
Sure, but I would wait for a month or so of the new timetable before it has been categorically been deemed a failure.
Surely it has to get better from here?Friday's numbers:
- 1T48 takes 11 (!) minutes to split at Cambridge, 8 late at Ely.
- 1T50 takes 9 minutes to split, 16 late at Ely.
- 1T52 takes 12 minutes to split, 17 late at Ely.
So, 0 out of 30 on-time for these at Ely since the timetable change - and a whole 5 out of those 30 arrived less than 5 minutes late. Actually the record is worse than that, as looking at RecentTrainTimes, 1T54 and 1T56 also have the same punctuality issues (though they don't split at Cambridge).
On-time arrivals for evening peak services at Ely since the timetable change:
View attachment 195593
(chart showing no on-time arrivals for 5 consecutive evening peak trains).
It must have been quite difficult to manage to degrade the service further from last year, but this does look like this year will be even worse than the dismal results seen with the previous timetable.
That should start on 5 January, not 14 December. The intervening period was 11 days of trying to run the service with the wrong track and signalling, followed by an 11 day closure for the new track to be joined up and the signalling installed.Sure, but I would wait for a month or so of the new timetable before it has been categorically been deemed a failure.
I hope that you are right here.The common demoninator is the split taking much longer (almost double the time) than scheduled.
That being said, this is the sort of issue that can be ironed out after a few weeks/months with proper attention paid to it by performance teams, so it is not guaranteed to be awful forever.
There’s just one consolation, which is one can only imagine how much worse still all this would be had Cambridge-Maidstone happened.
That should start on 5 January, not 14 December. The intervening period was 11 days of trying to run the service with the wrong track and signalling, followed by an 11 day closure for the new track to be joined up and the signalling installed.
I hope that you are right here.
Whereas I know that you are right here, and I started doing my imagining more than 10 years ago.
That should start on 5 January, not 14 December. The intervening period was 11 days of trying to run the service with the wrong track and signalling, followed by an 11 day closure for the new track to be joined up and the signalling installed.
I hope that you are right here.