A British "Taktfahrplan" - Feasible or Fantasy?

[zwk500]

Factually, very possible. It wouldn't be Swiss-style though, they would never let a "Colwich Jn"-situation arise in the first place.

If it wasn't Colwich, it'd be Stockport or Manchester Piccadilly in any event, at least until HS2 Phase 2 would have been fully complete.

 

[Krokodil]

Equally, we could get clockface integrated services working across London, the south East, Midlands, south Wales, urban north-west, Yorkshire, the north-east and the Central Belt, and just leave off the very rural areas in other parts of the country that have extremely low population density.

The rural areas are the easiest to integrate as they're less congested. The challenges are our biggest cities - London, Birmingham, Manchester and Glasgow. None of them have a centralised railway network as they all have at least a second significant interchange. How do you integrate services calling at Birmingham New Street with Moor Street/Curzon Street.

Provincial interchanges (Chester, Shrewsbury and Hereford for example) are already well-equipped for node status with enough platforms and lines for each service pair, but New Street, Piccadilly and London terminals are just trying to squeeze as much throughput as possible.

I suspect your natural starting point would be Euston, being the most important intercity station in the country and on a very congested section of line, on which the current timetable is extremely efficient and hard to change.

No, Birmingham New Street is the primary hub. Start with the XC services and build the timetable out from there.

There is also the issue that Euston to Birmingham/Manchester services are 3tph so will never fit neatly.

 

[Bletchleyite]

There is also the issue that Euston to Birmingham/Manchester services are 3tph so will never fit neatly.

Euston-Brum is now half hourly plus an hourly semifast, not a 20 minute pattern.

 

[zwk500]

The rural areas are the easiest to integrate as they're less congested.

Not when they're single line. Unless you are going to accept long dwells in some services then you will need to be very careful about how you time them or be prepared to move loops to fit the slots available out of the big cities.

 

[Krokodil]

Euston-Brum is now half hourly plus an hourly semifast, not a 20 minute pattern.

It'll become a 20 minute pattern after HS2 opens though. It may be simpler to treat that one as a TUAG [Turn Up And Go] flow and ignore it for the purposes of takt, just as metro services are.

Not when they're single line. Unless you are going to accept long dwells in some services then you will need to be very careful about how you time them or be prepared to move loops to fit the slots available out of the big cities.

The Cambrian will be fine then. Working with rural services is a lot easier than trying to get an entire half-hour's worth of trains to arrive at Manchester Piccadilly or Birmingham New Street in a 10 minute slot, dwell, and then leave in another 10 minute slot.

 

[Austriantrain]

If it wasn't Colwich, it'd be Stockport or Manchester Piccadilly in any event, at least until HS2 Phase 2 would have been fully complete.

In Switzerland, it would probably still be Birmingham in the end , because they would have done the entire station situation differently (and much more expensively). They might not even made an "integrated Takt hub" out of Birminham (it will become less and less relevant for Zurich e.g. with increasing 15-minute-frequencies), but they most probably would have centralised all major traffic flows at a single station.

But again, Switzerland does not lend itself as an example for the UK, or at least for the most important parts of England.

 

[The Planner]

It'll become a 20 minute pattern after HS2 opens though. It may be simpler to treat that one as a TUAG [Turn Up And Go] flow and ignore it for the purposes of takt, just as metro services are.

It might be on HS2, it will be half hourly on the existing network, which makes it a lot easier to deal with.

 

[NCT]

With the current limit of HS2, surely Colwich Jn becomes the Swiss-Style centre?

I think you are conflating two types of timetabling considerations.

The first type is defining a desired set of concentrated arrival and departure times at nodes to provide quick interchanges. Those are the times the passengers see and care about.
The second type is working out a physically possible and robust sequence of trains through constrained junctions. These are times that are absolutely crucial the industry needs to get right, but passengers don't care a hoot about.

Ideally, you start at locations with the first type of drivers, and then look at the second type of constraints to see if there are options where the first type of considerations can still be satisfied.

== Doublepost prevention - post automatically merged: 7 May 2026 ==

... trying to get an entire half-hour's worth of trains to arrive at Manchester Piccadilly or Birmingham New Street in a 10 minute slot, dwell, and then leave in another 10 minute slot.

Is this ever possible, or even desirable? Take New Street (think post HS2 for simplicity, and a 'reasonably ambitious' timetable), all the major directions are 4tph -
- Coventry (4 fast services per hour, 2 direct to Milton Keynes and 2 to Northampton; and 4 stoppers, 2 to International 2 to Coventry)
- Derby (2tph Sheffield and 2tph Nottingham)
- Stafford (2 Manchester and 2 Liverpool)
- Cheltenham (2 Bristol and 2 Cardiff)
Then you have metro corridors of Cross City (restored to 6tph) and Walsall (4tph).

In my view the 'turn-up-and-go' advantages of distributing those 4tph services far outweigh the slight disadvantage that not all services will connect perfectly to all other services - i.e. if from any one train you have a good connection to Manchester your connection to Liverpool will be 15 minutes longer.

The distributed 4tph model is essentially the Dutch model where every major urban corridor is treated like a metro one. This is still a kind of 'takt', just a Dutch takt instead of a Swiss takt - it removes the toughest constraints (that stations must have enough platforms for overlapping dwells, and node-to-node journey times must be multiples of 15 minutes) - but still requires things like symmetry and regularity.

 

[Krokodil]

Is this ever possible

That's the point that I was making. It's much easier to turn provincial interchanges (like Chester, Shrewsbury and Hereford) into takt nodes than to start trying to achieve it in our major cities.

or even desirable?

Certainly making the Voyager routes part of the takt is essential. If you have takt then Bristol and Oxford both have a half-hourly Manchester service and half-hourly Newcastle service - one direct, one with a change, but between them providing journey opportunities every half hour.

Then you have metro corridors of Cross City (restored to 6tph) and Walsall (4tph).

TUAG metro services would be excluded from the takt.

 

[D6130]

Forty-five years ago I was involved in a small way in the planning of Jim Summers' "Tartan Taktfahrplan" for the Scottish Region. Based on the Edinburgh-Glasgow high-speed line, it extended Northwards in a series of triangles to Inverness and Aberdeen via Perth and Kirkcaldy....with 1 train every two hours on the Glasgow - Aberdeen, Edinburgh - Aberdeen and Aberdeen - Inverness axes and 1 train every four hours on the Glasgow - Inverness and Edinburgh - Inverness routes. This gave 1 train per hour Glasgow - Perth (with an additional Glasgow - Dundee semi-fast every two hours); and Dundee - Aberdeen (which then - as now - incorporated the small number of Anglo-Scottish trains to and from Aberdeen.

The rural lines North and West of Glasgow and Inverness were planned on a 1 train every six hours basis....although the West Highland lines had a two hour/four hour takt to accommodate both the Oban and Fort William services. Crossings were synchronised with a remarkably-little increase in dwell times.

Of course the Glasgow suburban services - like the Southern Region - had had their own Taktfahrplan for many years and this was easily incorporated into the main timetable. The only services that had to be tweaked slightly were the Glasgow -Stranraers, which were timed to connect with the Northern Ireland ferries....but these were incorporated within the Glasgow -Ayr- Ardrossan - Largs takt with a minimum of disruption.

 

[NCT]

TUAG metro services would be excluded from the takt.

TUAG metro services need to part of the takt, as they are made of lower-frequency (often half-hourly) building blocks.

If Birmingham - Rugeley Trent Valley is 2tph, then they should connect with (hopefully 2tph) Trent Valley - Euston services. Ditto at Lichfield Trent Valley for the Cross City.

While (post HS2) Milton Keynes - Euston will likely have a 6tph 'metro' fast service to Euston (2 Birmingham, 2 Northampton, 2 Trent Valley), if at all possible at Rugby, the Trent Valley should connect into Birmingham - Northampton services.

 

[Bletchleyite]

TUAG metro services need to part of the takt, as they are made of lower-frequency (often half-hourly) building blocks.

They can't completely do though. If your base Takt is half hourly and your 4tph is made up of two to A and two to B offset by 15 minutes, you can't connect with both. You have to choose one.

You could group the A and B services together so they're only 5 minutes apart, but if you do that you've then made that local service less useful. And when the majority of Merseyrail users are going to Liverpool (for example) that is harmful, not helpful. See also most Marston Vale journeys being entirely local with very few connecting onto the WCML, which meant that breaking what was already a Takt style connection so the trains were better timed for the school day was actually beneficial? It was a beautiful piece of (Dutch style) Takt, but it didn't serve the majority of passengers and there was if I recall a local campaign (ish) to change it.

 

[NCT]

They can't completely do though. If your base Takt is half hourly and your 4tph is made up of two to A and two to B offset by 15 minutes

This is still a takt structure, part of your active timetable design considerations.

You can have each 2tph building block making good Swiss-style connections at their extremities and combined into a 15-minute frequency pattern in the core. Granted you won't be lucky everywhere, but it you don't consider them in your design philosophy you miss these opportunities.

 

[The Planner]

Is this ever possible, or even desirable? Take New Street (think post HS2 for simplicity, and a 'reasonably ambitious' timetable), all the major directions are 4tph -
- Coventry (4 fast services per hour, 2 direct to Milton Keynes and 2 to Northampton; and 4 stoppers, 2 to International 2 to Coventry)

Good luck with that one.

 

[NCT]

The issue with picking up any single point is that a lot of the relationships are equally as constraining both ways. E.g. if you build the timetable around Manchester oxford Road, you are dictating the slots for trains at Water Street Jn, which dictates the slots into Victoria. Conversely, if you picked Victoria, you're dictating the slots at Water Street Jn, which dictates the slots at Oxford Road. One is not really any more obviously better than the other.

Timetabling is a multi-dimensional geometric puzzle. You look at what works across a complex geography like Central Manchester as a whole, not any one particular location in isolation. Part of the consideration includes looking at necessary timetable simplification to reduce / remove unsolvable multi-dimensional dependencies. London Bridge was one such example of simplification - having Cannon Street, Thameslink and Charing Cross lines work independently. Some sort of similar simplification will have to happen in Manchester.

If it wasn't Colwich, it'd be Stockport or Manchester Piccadilly in any event, at least until HS2 Phase 2 would have been fully complete.

Colwich isn't actually that bad because there are only 2tph to/from Hixon in each direction.

When a network has a mixture of Swiss (low frequencies where well timed interchanges are valuable) and Dutch (good interchanges by virtue of having frequency) characteristics, you start in the Swiss parts. Dutch timetabling starts in Zwolle.

On the west side of England you'd start on the Chester - Shrewsbury - Hereford corridor. Shrewsbury - Manchester becomes the first train on the graph into Manchester Piccadilly. Shrewsbury - Birmingham becomes the first train on the graph at New Street.

You then plan the remainder of the trains around these Shrewsbury trains, and then you hope that Crewe and Stoke direction trains will form parallel moves at Cheadle Hulme, and that Hazel Grove and Mid Cheshire Line trains will form parallel moves at Edgeley Junctions while leaving space for Northenden - Denton freight services, and then Alsager line - Stafford services compliantly cross Kidsgrove and Stone while not crashing into each other on the single line. As luck would have it (and there IS a huge element of luck), all of this is possible, and Hope Valley fast paths happen to be in the right positions at Grantham, Peterborough and Ely for an ECML and Anglia takt structure.

== Doublepost prevention - post automatically merged: 7 May 2026 ==

You'd pick the single national takt point on the basis of it being the most constrained place. If crossrail was largely independent from the GWML I'd agree with you, but it's not, because building the 5th and 6th tracks from Paddington to Heathrow wasn't ever likely to be viable.

However I could see arguments for Several other places including Blackfriars (flat junction for Thameslink), Manchester Oxford Road (due to the range of traffic feeding Castlefield), Coventry (as it's got limited platforms and the single line to Leamington), Birmingham New Street (due to the crossing moves between XC) and so on.

The locations you listed are the 'make the timetable work' locations (second type of considerations) not 'arrival / departure time defining' locations (first type of considerations).

Crossrail is largely independent. It has near sole use of GWML Relief Lines (only interaction is freight) which is operationally independent from the Main Lines, and Airport Junction is grade separated. There's very little external constraint to Crossrail's timetabling. Crossrail's unreliability is overwhelmingly down to infrastructure defects and broken down trains, not the timetable.

At Birmingham New Street you start the timetable from Shrewsbury, then Cheadle Hulme, then the Trent Valley (Rugeley Trent Valley and Lichfield Trent Valley determine the first metro trains on the graph through Birmingham). The way to make New Street work is to have it operationally as simple as possible - keep the Stour (Soho - Coventry) and Gloucester (University - Water Orton) sides separate, operate each side as outer lines for through services and inner lines for terminating, and don't be too precious with how services pair up across New Street. This would give you a lot of flexibility to ensure service outer ends form nice Swiss-style interchanges while the lines close to Birmingham provide well distributed metro service frequencies.

For Thameslink, yes at Blackfriar South Junction you want to ensure there are 16 London Bridge paths per hour on parallel moves and 8 Elephant paths per hour on parallel moves for the timetable to work. However, this in itself neither constrains nor informs, for example, where the Peterborough or Cambridge ideal arrival and departure times are. A sequence of logic might be as follows:
- Can the Thameslink Peterborough times be optimised for a Peterborough Swiss-style node simultaneously with King's Cross - Ely/Lynn service optimised for the Ely node (noting Peterborough - Ely travel time)
- Can the Thameslink Cambridge (semi-fast) service coexist with the King's Cross - Ely/Lynn while providing a 15-minute interval service to Stevenage with the Thameslink Peterborough
- Can these paths be part of the Blackfriars South Junction's 16tph London Bridge parallel moves?
As luck would have it, there is a sequence of 'yes' - and it's not just luck, there are options for achieving these aims. I won't go into the technical details - you'll have to trust me when I say this is based on much more than a hunch.

Separately, while Haywards Heath - Three Bridges - Victoria and Basingstoke - Woking - Waterloo are you intensive metro-like corridors, your starting locations actually want to be places like Salisbury, Havant, Barnham and Lewes. Salisbury has all the ingredients for being a Swiss-style full takt node; Southampton - Brighton wants to connect into Portsmouth - Waterloo at Havant (choice of fast or semi-fast, with trade-offs), Portsmouth - Victoria wants to connect into Southampton - Brighton at Barnham, Victoria - Ore wants to connect into Brighton - Seaford at Lewes.

 

[The Ham]

I think that there's two factors in play.

Can we get lines to +1tph where they are tph.

Where that's not possible can we get 1tph services so they cluster with interchange services at key junctions.

I think that for the first one we need to accept that if there's 1tph as a through service that the other service(s) doesn't have to be but rather goes to 2 or 3 key interchange points but not the main town/city without a change.

The thing to be conscious of with the second option, is that we run the risk of the trains being slowed down, that's not too bad if you do it once or twice, but after that we could be adding 10 minutes to the journey time.

 

[Route115?]

The current network - and indeed HS2 - is very London centric, apart from some good regional networks for example in Glasgow & Merseyside. The problem is that many journeys not to/from the capital are not very fast. I wonder when GBR rail HQ at Derby opens how many staff will travel in to meetings rom the regions by car because it is so much quicker. (Unless someone does the sensible thing and hold them in London.) This is what a Taktfahrplan plans needs to change, but the potential rewards are huge.

The current timetable is based around maximise the utilisation of resouces. If you run an hourly service between two stations fifty minutes apart there will almost certainly be two sets of stock rostered when a taktfahrplan might require three with a total of 80 minutes per hour sitting at the terminii in order to fit in with other services. We would probably baulk at the cost in this country.

Whilst the Swiss railways typically have a 7% (4 minutes in the hour) recovery allowance to allow for recovery from delays - more on single track sections - we would still need a step change in reliability to make a taktfahrplan work in this country. Could we manage it? It might also mean slower journeys on the main routes - we could have a detailed discussion on recovery times.

I have travelled down to Cornwall after a holiday in Switzerland and do despair when the connections to branch lines are so poor. Sadly, I doubt that will change in a while.

 

[zwk500]

I wonder when GBR rail HQ at Derby opens how many staff will travel in to meetings rom the regions by car because it is so much quicker.

The answer will be very close to zero, because the meetings will almost certainly be over MS Teams.