How can we make changing trains more bearable for the travelling public?

[jfollows]

It's clearly not simple, and a line has to be drawn somewhere in how long a train can be held for a connection - some branch line trains have such short turn-round times that delaying a train in one direction delays the next train back and so on.
In my personal experience:

1. When I lived in Fareham in the 1980s I never had a problem with the branch line connections at Eastleigh in either direction, trains were held for a short while to make connections.
2. More recently I have seen shocking examples of connections not held, such as the Barrow train being sent out of Carlisle in sight of the arriving main line train which was supposed to connect to it.

In the second case, nowadays it gets a shrug and a "you can claim delay repay" answer. It's not acceptable to me and, yes, I know that holding trains for connections causes other problems, but it's the sort of behaviour which causes people to abandon travelling by train entirely.

So I think there is scope for a more "sensible" approach than the Balkanised "not our fault, some other company's fault" approach coupled with penalties for late operation caused by holding late connections which aren't in anyone's interest.

Whatever happens now, GB Rail or not, it's time the railways understood that leisure travel and a lack of a captive audience is going to matter more and more, and to be fair I think a lot of people do. The process needs to be more "bearable" as the thread says, people can understand why connections aren't held in certain circumstances but the bean-counting which underlies the operation of today's railways makes silly decisions which aren't understandable in the eye of the travelling passenger.

 

[Bletchleyite]

Fundamentally what is needed is to stop penalising TOCs for trains that are delayed as a result of awaiting a delayed connection. If that is done, then TOCs will generally make sensible operational decisions most of the time. TOCs are also not penalised for delays resulting from providing access to a disabled person, so there's scope to exclude certain reasons.

It could be done better still with "big data", but that would be a start.

 

[Snow1964]

Might be controversial, but perhaps sensible capacity limits need to be introduced on some sections, rather than trying to maximise line usage, to the point where there is no slack for contingencies

It is much better to have each train built around half hourly intervals, with some at hourly or quarter hourly, than a mad mix of random times each hour.

3 trains an hour sounds great, but is hopeless on long distance services, where connections to services based on other intervals exist. So nothing really connects well. Also far too many sections have squeezed an extra train in so that the gaps are nowadays uneven

Maybe it needs some more dynamic loops, or a non platformed reversible fast track laid between others so that there is a way to untangle a delay. Far too often trains get stuck for 15-30 minutes (or 10-20 miles) because there is no way to overtake a stopper. Even to the point a train with recovery time cannot be looped, as no loop exists.

One other thing, perhaps seat reservation software needs programming so it offers seats in the coaches that will stop nearest a connection footbridge (and location of stairways is different at most stations). The Railway then needs to print things on itineraries to show it is trying to help connections. No one wants to be traipsing down a whole platform causing platform congestion, or facing a long walk. (Of course last few seats issued would have to take what is left), but a minimise walking policy when allocating seats would help connections.

 

[Bletchleyite]

Might be controversial, but perhaps sensible capacity limits need to be introduced on some sections, rather than trying to maximise line usage, to the point where there is no slack for contingencies

It is much better to have each train built around half hourly intervals, with some at hourly or quarter hourly, than a mad mix of random times each hour.

Yay, a Taktfahrplan! Yes, I totally agree. Spare paths are a virtue (resilience), while for good connections you need to choose between intervals of 120/60/30/15/5, OR 120/60/20/10/5, OR 120/60/30/10/5 - you cannot mix all three without having some badly aligned connections. I think 120/60/30/15/5 is probably the one to go for, as only tiny numbers of places justify 6tph. Though if you ask the Swiss you can sort of mix them if you merge the 3tph into 2 before meeting at a node with 2tph of connections by adding stops into one to slow it down and terminating one short.

 

[miklcct]

3 trains an hour sounds great, but is hopeless on long distance services, where connections to services based on other intervals exist. So nothing really connects well. Also far too many sections have squeezed an extra train in so that the gaps are nowadays uneven

If 2 trains an hour aren't enough, but there isn't the capacity to operate 4 trains an hour, than the only way to go is 3 trains an hour.

Yay, a Taktfahrplan! Yes, I totally agree. Spare paths are a virtue (resilience), while for good connections you need to choose between intervals of 120/60/30/15/5, OR 120/60/20/10/5, OR 120/60/30/10/5 - you cannot mix all three without having some badly aligned connections. I think 120/60/30/15/5 is probably the one to go for, as only tiny numbers of places justify 6tph.

I will go for 120/60/30/20/15/12/10/turn-up-and-go, so under most pairs of combinations at least half of the changes are good changes.

 

[zwk500]

St Pancras is out of the direct line between Hendon and East Midlands Parkway and involves approximately 25 km of extra distance compared to going direct.

Nevertheless it's still possible. On a pedantic note, Kettering is out of the direct line between Hendon and EMP as well (admittedly only 10km off the direct line)

Similarly, going from Harpenden still requires 3 changes on the direct line as well (Luton, Kettering, Leicester).

But how often are people travelling from Hendon or Harpenden to East Midlands Parkway?

It isn't even possible in both directions in the case of Lymington branch line!

For example, the last few trains from Lymington to Waterloo is a 30-minute wait at Brockenhurst on paper, but I'm not sure if in practice passengers can usually make the 0-minute unofficial connection there as it's not a cross-platform interchange.

This is something that should be focussed on for tidying up. Branch lines should have sensible connections with London trains, and receive investment to make it possible if it currently isn't through line speed improvements or Rolling stock upgrades.

 

[Bletchleyite]

This is something that should be focussed on for tidying up. Branch lines should have sensible connections with London trains, and receive investment to make it possible if it currently isn't through line speed improvements or Rolling stock upgrades.

The Lymington example looks like it probably did connect at one stage, and the mainline train was tweaked by a few minutes and so it now misses. Probably fairly easy to fix. This happens far too often - sometimes it seems people don't think.

On the other hand, there are good examples - the connections between the Marston Vale and both London and Birmingham in the December timetable look straight out of Switzerland - Marston Vale arrives, then a full cycle of arrivals/departures to all destinations, fast and slow, then it departs again. Wonder how long before a timetable tweak mucks it back up?

 

[zwk500]

Yay, a Taktfahrplan! Yes, I totally agree. Spare paths are a virtue (resilience), while for good connections you need to choose between intervals of 120/60/30/15/5, OR 120/60/20/10/5, OR 120/60/30/10/5 - you cannot mix all three without having some badly aligned connections. I think 120/60/30/15/5 is probably the one to go for, as only tiny numbers of places justify 6tph. Though if you ask the Swiss you can sort of mix them if you merge the 3tph into 2 before meeting at a node with 2tph of connections by adding stops into one to slow it down and terminating one short.

Also strong agree on the 30/15 option. Tbh if you're getting down to 10 minute headways you really should be segregating the traffic out as a proper metro. In my experience 3tph tends to not really work as usually it has a 1 fast 2 semi-fast pattern, meaning most stations get a 20/40 minute service and the main station has a 2 out of 3 chance of a slower than headline journey time. If train capacity is a problem, lengthen the train.

== Doublepost prevention - post automatically merged: 8 Nov 2022 ==

The Lymington example looks like it probably did connect at one stage, and the mainline train was tweaked by a few minutes and so it now misses. Probably fairly easy to fix.

If that's the case, it really should be caught by the planning process. I know MK are busy at the moment but this kind of thing just needs a 'key connections' checklist on a post-it note.

 

[Ken H]

Yay, a Taktfahrplan! Yes, I totally agree. Spare paths are a virtue (resilience), while for good connections you need to choose between intervals of 120/60/30/15/5, OR 120/60/20/10/5, OR 120/60/30/10/5 - you cannot mix all three without having some badly aligned connections. I think 120/60/30/15/5 is probably the one to go for, as only tiny numbers of places justify 6tph. Though if you ask the Swiss you can sort of mix them if you merge the 3tph into 2 before meeting at a node with 2tph of connections by adding stops into one to slow it down and terminating one short.

Put in a few flying junctions. So you dont have delays going 1 way causing disruption the other was when there is a turnout.
One wonders how much Norton Bridge has made the west coast more resilient?

 

[zwk500]

Put in a few flying junctions. So you dont have delays going 1 way causing disruption the other was when there is a turnout.
One wonders how much Norton Bridge has made the west coast more resilient?

It'll help, no doubt, but most of the junctions you'd need to rebuild are in built-up areas. How do we pay to relocate everybody out of affected areas?

 

[The exile]

For business and commuting journeys, every second matters. Especially the former one, every second spent on the trip is wasted productivity for the employer, and for the latter, every second saved on the commute is a second of sleep gained.

If the railways want businessmen and commuters back, they need to run a service where the reliability can be timed to seconds, like Japan.

Have to disagree there- what matters even more is predictability. One of the reasons that the networks that used to be famed for their punctuality were so was the amount of “slack” in their timetables. It seriously doesn’t make the slightest bit of difference if the train takes 32 minutes or 34 - what matters is whether connections on to and off it (and that includes to/from buses) are reliable. If the swanky new timetable advertising speeded up journeys does so at the expense of reliability then everyone loses.

== Doublepost prevention - post automatically merged: 8 Nov 2022 ==

It’s not just about train frequencies and connection times - it’s also about clear signage and information at junction stations. This means, for example, clear working monitors - preferably in pairs so the next departures are always visible and you don’t need to wait for it to scroll through pages 2-4 to find the platform for you train which leaves in 5 minutes. It means human oversight which can react to the reality of the situation and is authorised to do something about it. It means automatic acceptance of tickets by all alternatives (and that includes buses) on an “accept now, sort out afterwards” basis. It needs delay penalties based on delays to passengers not delays to trains. It also requires acceptance that a half hourly service on a line that takes 29 minutes end to end can’t be done with two trains!

 

[miklcct]

It also requires acceptance that a half hourly service on a line that takes 29 minutes end to end can’t be done with two trains!

In such case I'd adjust the headway to be 40 minutes such that it can be done with 2 trains.

 

[zwk500]

In such case I'd adjust the headway to be 40 minutes such that it can be done with 2 trains.

which doesn't help if the connection is every 30 minutes. Correct solution in that case is to reduce the end-to-end time to 25 minutes. Be that closing smaller stations, electrification or linespeed improvements.

 

[SussexSeagull]

When I commuted from Durrington to London Bridge 20 odd years ago in the slam door era there was a man who got on the train before me who we nicknamed the doorman because he used to get up and slam the door between carriages shut if anyone left it open.

Anyway when the guard came through checking tickets he would invariably ask something like 'Why was the train 4 minutes late leaving Burgess Hill?' Now at the time I thought this strange as most people were delighted if the train got to London Bridge 15 minutes late. However looking back I think he had a point. The default for a train service should be to be on time. Now if there is a genuine problem then fair enough but that train arrived on time - and I mean at 0832 or whenever it was due - to London a handful of times a year.

I do question why if a train is constantly not on time - and I remember a Brighton to London service a few years ago that made the news after not being on time for an entire calendar year - the timetable isn't looked at and a moe realistic one is issued. That way people can be more confident of reaching somewhere in good time or actually meeting a connection.

 

[Bletchleyite]

which doesn't help if the connection is every 30 minutes. Correct solution in that case is to reduce the end-to-end time to 25 minutes. Be that closing smaller stations, electrification or linespeed improvements.

That or you look for another route nearby it could potentially interwork with. For instance, interworking Ormskirk-Preston with Preston-Colne and Preston-Blackpool South, while it has other downsides, makes more efficient use of units. Otherwise you couldn't do hourly on Ormskirk-Preston (running time about half an hour) without two units with excessive, wasteful layovers. While that unit has now been added back in as the circuit couldn't *quite* make do with one unit fewer, it now effectively provides an extra 10 minutes of resilience to all three routes rather than 30 minutes to one.

Or if you have a branch line that is much quieter at one end (doesn't work for Ormskirk-Preston or the Marston Vale as neither is a true branch line, but rather they are both quiet through routes with demand at both ends), but does for instance with the Conwy Valley) you shorten it and replace the end bit with a bus. If that was truncated to Betws one unit could run two-hourly clockface.

Or if you still can't do that, you either drop to 0.5tph or accept the near-waste of a unit, as the Takt is more important.

 

[miklcct]

which doesn't help if the connection is every 30 minutes. Correct solution in that case is to reduce the end-to-end time to 25 minutes. Be that closing smaller stations, electrification or linespeed improvements.

Then I'll make the headway 45 minutes such that half of the trains are good connections.

 

[bib]

Think I saw somewhere that the swiss railways app has station maps which show you the platforms and where you need to go to make your connection. Might be more useful for bigger stations and infrequent travellers, eg stopping people getting lost at new street

 

[Trainbike46]

Then I'll make the headway 45 minutes such that half of the trains are good connections.

If the mainline runs every thirty minutes, the branch will have to run either every 30 minutes or every 60 minutes, to connect properly. If the length of the branch is so, that you can't create an efficient diagram for the branch by itself, the best solutions seem to be:
- interwork with other branches nearby, to make more efficient diagrams
- Speed up the train (linespeed improvements, station closure, electrification, more suitable rolling stock, for example), so that it does fit into efficient diagrams

you can't just reduce the frequency by a little, such as moving to every 40/45 minutes, because then there will never be a connection. It's halfhourly/hourly/every 2 hours, or not having connections, which is a terrible idea for a branch line

 

[Bletchleyite]

Think I saw somewhere that the swiss railways app has station maps which show you the platforms and where you need to go to make your connection. Might be more useful for bigger stations and infrequent travellers, eg stopping people getting lost at new street

You only need to do one thing not to get lost at New St - go to the B end. It just needs massive signs.

 

[leytongabriel]

Getting architect and railway engineers to understand that people don't like long walks to connecting trains in new stations might help. They are a misery for people with luggage, mobility issues, kids and prams etc and add several minutes to journey times which can nullify the effects of 'improved' services. A trend for long traipse seemed to start with the Jubillee line extension and there have been discussions on here about when to use the Central or Elizabeth line (given that changing at Liverpool St is really no fun). the tunnel round from the Victoria line to Kings Cross mainline via the northern councourse, or even sticking main Brighton and Gatwick services in a far-flung corner of Victoria. In other words, bringing passengers a bit futher forward in the equation, giving them a bit more priority. Of course putting a shoppig centre in front of the platforms will always be more important to some.

 

[Dr Hoo]

Getting architect and railway engineers to understand that people don't like long walks to connecting trains in new stations might help. They are a misery for people with luggage, mobility issues, kids and prams etc and add several minutes to journey times which can nullify the effects of 'improved' services. A trend for long traipse seemed to start with the Jubillee line extension and there have been discussions on here about when to use the Central or Elizabeth line (given that changing at Liverpool St is really no fun). the tunnel round from the Victoria line to Kings Cross mainline via the northern councourse, or even sticking main Brighton and Gatwick services in a far-flung corner of Victoria. In other words, bringing passengers a bit futher forward in the equation, giving them a bit more priority. Of course putting a shoppig centre in front of the platforms will always be more important to some.

Recognising that many station and interchange designs are heavily influenced by original 19th Century land footprints, listed buildings and so on, I know what you mean but modern levels of demand, masses of wheeled luggage, etc. mean that interchange 'pathways' just have to be bigger. Then there are factors like having secondary emergency routes and disabled access.

So, at somewhere like Birmingham New Street the fact that passengers entering or exiting the station (often regulars or locals) can easily use the 'A' end, takes the pressure off the 'B' end, making it easier for interchange passengers for example.

Having experienced stations like the 'old' New Street, London Bridge, Reading, Victoria Underground and so on there's no way that I would want to go back to the originals even if things were more compact back then. Un-modernised and un-expanded relics like Lincoln and Exeter St Davids may look quaint but can't handle crowds expeditiously and safely.

 

[61653 HTAFC]

Every Train should call at every station.

The next train to depart platform 10 is for Norwich.

Calling at Stratford, Maryland, Forest Gate, Manor Park, Ilford, Severn Kings, Chadwell Heath, Romford, Gidea Park, Harold Wood, Brentford, Shenfield, Ingatestone, Chelmsford, Hatfield Perveral, Witham, Kelvedon, Marks Tey, Colchester, Mannigtree, Ipswich, Needham Market, Stowmarket, Diss and Norwich.

Only about 17 extra stops...

London to Glasgow an extra 30 stops..

Imagine the sleeper calling at every station!

I'm assuming this post is a bit tongue-in-cheek, but a few weeks ago on a Sunday the amended timetable on North TPE involved the usual Leeds to Huddersfield stopper being merged with the Huddersfield to Manchester stopper. There were definitely some positives about this along with a few negatives, and being one of only two (at best) trains per hour along the route it wasn't the disaster that was the 2018 debacle. The positives: a six-car formation, using P1 at Huddersfield (so less of a walk to the exit), and a more visible staff presence. The negatives: some or most hours this was the only service running at all along the route, so even a six-car set was woefully inadequate. This also meant lots of longer-distance passengers with large amounts of luggage which was blocking aisles and doors.

 

[Purple Train]

Every Train should call at every station.

The next train to depart platform 10 is for Norwich.

Calling at Stratford, Maryland, Forest Gate, Manor Park, Ilford, Severn Kings, Chadwell Heath, Romford, Gidea Park, Harold Wood, Brentford, Shenfield, Ingatestone, Chelmsford, Hatfield Perveral, Witham, Kelvedon, Marks Tey, Colchester, Mannigtree, Ipswich, Needham Market, Stowmarket, Diss and Norwich.

Only about 17 extra stops...

London to Glasgow an extra 30 stops..

Imagine the sleeper calling at every station!

That's a bit of a strange routing...