Is a 5 minute minimum connection time too unrealistic even at smaller stations?

[johntea]

I was in Sheffield yesterday and planned to visit Manchester afterwards, totally forgetting the works at Piccadilly rather than catch a service to Stockport or wherever then a rail replacement bus to Piccadilly journey planner suggested I could catch the 22:27 Northern service to Wakefield Kirkgate arriving at 23:12 (Platform 1) then head across to Platform 2 for the 23:18 Transpenine service to Manchester Victoria (usually Piccadilly but they seem to have managed to move that to Victoria during the works)

That would have probably been fine until the 22:18 Sheffield to Leeds express service got delayed until 22:31 and took priority, so the 22:27 departed 7 minutes late and arrived in Wakefield Kirkgate just in time to see the Transpenine service pulling away!

In the end I just went home instead but it did amuse me thinking about it if I had really been stubborn I could have had a valid case for Northern providing me a taxi from Wakefield to Manchester! (not that I bought a ticket in the end for that journey as I knew it was unlikely to happen once we left 7 minutes late - I already had a Wakefield - Sheffield day return ticket)

5 minutes to make a connection seems quite tight even at Wakefield Kirkgate, you're basically relying on things running pretty much bang on schedule

(There is of course the valid point I could have caught the 22:18 express service and been there in good time even accounting for the delay but I was already sat on the 22:27 and didn't want to risk getting off that and then potentially not being on either train!)

 

[Bletchleyite]

I do find 5 minutes too tight except where it's same or cross platform, yes. Given the railway's punctuality or lack thereof, 10 is more realistic in most cases.

It would be good if GBR got an IT system in place that could differentiate.

 

[Crithylum]

Depends on the station (and punctuality of the operators). At Liverpool Street (MCT 15 minutes), I quite reliably (at least 95% of the time) make 5 minute connections from Elizabeth Line to high level without running. I’m really fortunate that I don’t have to deal with advance singles on this route.

I once planned an illegal cross London transfer (LST to PAD), that had the inbound service arrive 8 minutes late to Liverpool Street, and then had my final train cancelled from Paddington to Reading, forcing me to catch a train 2 minutes earlier to Reading. I was able to walk (not speedwalk or run) to the very front of the train, plonk down on a table, and start getting work done before the train started moving.

However, I would describe Liverpool Street as the exception rather than the norm. The likelihood of a train arriving the exact minute is comparable to arriving at all for the worst stations.

I’m not sure what the best solution is. It is probably “too difficult” to factor in punctuality or expected platform, therefore leaving in some pretty extreme edge cases. Anyways, I tend to ignore the MCTs (both ways) unless using advance singles

 

[dk1]

I have always enjoyed the thrill of making unadvertised connections with minutes to spare but the best ones are minus ones you never should make when something is running late.

 

[Mgameing123]

I was in Sheffield yesterday and planned to visit Manchester afterwards, totally forgetting the works at Piccadilly rather than catch a service to Stockport or wherever then a rail replacement bus to Piccadilly journey planner suggested I could catch the 22:27 Northern service to Wakefield Kirkgate arriving at 23:12 (Platform 1) then head across to Platform 2 for the 23:18 Transpenine service to Manchester Victoria (usually Piccadilly but they seem to have managed to move that to Victoria during the works)

That would have probably been fine until the 22:18 Sheffield to Leeds express service got delayed until 22:31 and took priority, so the 22:27 departed 7 minutes late and arrived in Wakefield Kirkgate just in time to see the Transpenine service pulling away!

In the end I just went home instead but it did amuse me thinking about it if I had really been stubborn I could have had a valid case for Northern providing me a taxi from Wakefield to Manchester! (not that I bought a ticket in the end for that journey as I knew it was unlikely to happen once we left 7 minutes late - I already had a Wakefield - Sheffield day return ticket)

5 minutes to make a connection seems quite tight even at Wakefield Kirkgate, you're basically relying on things running pretty much bang on schedule

(There is of course the valid point I could have caught the 22:18 express service and been there in good time even accounting for the delay but I was already sat on the 22:27 and didn't want to risk getting off that and then potentially not being on either train!)

5 minutes is brilliant if you want delay repay XD

 

[yorksrob]

It's not so much about the connection, more about the security of not being stranded when it goes tits up.

 

[The exile]

It's not so much about the connection, more about the security of not being stranded when it goes tits up.

If it’s an advertised connection you should have that security (eventually). Whether you consider it wise to risk the hassle has to be a matter of personal choice - if a railway-issued schedule includes the connection, then the railway must be in the hook for it.

 

[xotGD]

Making a 9 minute connection at Leeds from an inbound LNER to an outbound Airedale service is very easy. What is annoying is when the inbound is late, you miss the connection, and aren't entitled to Delay Repay.

 

[yorksrob]

If it’s an advertised connection you should have that security (eventually). Whether you consider it wise to risk the hassle has to be a matter of personal choice - if a railway-issued schedule includes the connection, then the railway must be in the hook for it.

Unfortunately in many places we don't have the luxury of frequent services late into the night to fall back on, so we do have to consider official connections carefully.

 

[Topological]

Only the railway would have anything to gain from extending the minimum connection time.

For those who would like a longer personal minimum connection time, then surely plan appropriately?

No need to invest in complicated data-driven solutions on platforming etc.

 

[JamesT]

I do find 5 minutes too tight except where it's same or cross platform, yes. Given the railway's punctuality or lack thereof, 10 is more realistic in most cases.

It would be good if GBR got an IT system in place that could differentiate.

Where there's a simple station with only two platforms it's easy to say there would be a consistent connection time. But in a larger more complicated one, although normally you might be booked into adjacent platforms, wouldn't the connection time have to take into account a service being rerouted into a different platform?
If you're making a guarantee that you will be able to make it from one train to another if they're on time, then I think you're stuck with giving the timings of the worst case.

 

[35B]

Only the railway would have anything to gain from extending the minimum connection time.

For those who would like a longer personal minimum connection time, then surely plan appropriately?

No need to invest in complicated data-driven solutions on platforming etc.

The railway also stands to lose - if they get too extended, connections will miss and people won't travel.

If a 5 minute connection is unreasonably tight because it leaves no contingency, then it should be reviewed - but the first priority should be the punctuality.

 

[Bletchleyite]

If a 5 minute connection is unreasonably tight because it leaves no contingency, then it should be reviewed - but the first priority should be the punctuality.

On a railway where a train 5 minutes late isn't even officially delayed, I don't think that's going to be solved. There are some connections that very rarely miss because of operational constraints, e.g. a slow train is almost never let out past Ledburn Jn if a fast is due, and these could have 0 minutes and still be reliable, but most really do need more. Even 10 is tight at somewhere like New St, and 5 at Milton Keynes Central is VERY tight.

AI would seem ideal to do an assessment of the network to work out which specific ones can be lower based on reliability and ease of making the connection, but otherwise I do think it needs to be longer. 5 minutes is very tight for any connection that isn't same- or cross-platform for a non-expert rail user, and can even be tight for an expert. Using MKC as an example again, if you get off a busy train it can easily take 5 minutes to get up the stairs, but if you come in on 1 and need to change to a stopper on 2 one minute would be fine to be honest, they aren't going to send the stopper out in front and make the semifast 10 minutes late if they can hold it a couple of minutes.

 

[Topological]

What does a person gain from minimum connectivity being extended?

As you say, connections will miss.

Currently, people can make their own judgement. Those of us who want the fast connection are offered it, whilst those who want to allow longer can.

 

[Bletchleyite]

What does a person gain from minimum connectivity being extended?

A reliable and low-stress journey.

As you say, connections will miss.

Currently, people can make their own judgement. Those of us who want the fast connection are offered it, whilst those who want to allow longer can.

I think there would be an argument for having planners default to a minimum of ten minutes, but allowing people to knock it down to 5 if they are confident, expert users. However I do think analysis of specific connections would be worthwhile, as as I say there are some very short ones that never miss, and there are some longer ones where the train may not be late but you'll still struggle. Booked platforms being on the same island would be enough, generally, to lower it somewhat - a dataset of which platforms are grouped at all the stations naturally shouldn't be too hard to do, plus a dataset of pairs of platforms which could be a particular issue e.g. if the only way to cross the line is a level crossing which could be closed a few minutes before the connection and so render making it impossible.

I do think WMT <-> Avanti at MKC needs to be 10 (based on experience of making that connection, or rather not making it), but I see no reason why Platform 1 to 2 should be any more than 2 as it'll basically never miss unless the delay is much longer than the present 5.

Bletchley will be similar. 4 to another train on 4 or onto 5 (i.e. semifast Brum onto stopper) could be two minutes, same-platform could even be zero as it'll either miss or it won't regardless of how long you allow as the trains will be in a certain order, but going over to the new high level platforms probably needs to be 7-10 or so, 5 will be tight to the far side for an average user.

 

[35B]

What does a person gain from minimum connectivity being extended?

As you say, connections will miss.

Currently, people can make their own judgement. Those of us who want the fast connection are offered it, whilst those who want to allow longer can.

You write from the perspective of someone with detailed knowledge, and the interest to research a quick change. I also surmise (but stand to correction) that you are able bodied and generally travel relatively light.

The MCT is intended to provide for a normal person to make a connection. That needs to be based on the passenger who does not know the station, and to have allowance for them to be at the wrong end of the train for making a connection. If people are missing connections, or having to run to make them, then that is a sign that the MCT is inadequate

As my local station (Grantham) has a 7 minute MCT, I've sympathy with @Bletchleyite's view of MKC. New St's 12 minutes strikes me as reasonable - but with the caveat that the bottleneck at the staircases is so bad at times that no MCT could adequately represent it.

 

[Bletchleyite]

As my local station (Grantham) has a 7 minute MCT, I've sympathy with @Bletchleyite's view of MKC. New St's 12 minutes strikes me as reasonable - but with the caveat that the bottleneck at the staircases is so bad at times that no MCT could adequately represent it.

New St is a prime example of one where you need differential MCTs. Same island could be 5, if it involves going up top then 15 might be more suitable because of (a) those queues, and (b) how confusing the A end can be at times for infrequent users. Yes, there can be platform alterations, but unless you're just going to AI-analyse the lot and make it dynamic you probably can't legislate for that.

Because it's all about journey planners now, not a number next to a station on the timetable, you don't need to have one for the whole station that's too long for some connections and too short for others.

Manchester Piccadilly is another good example. Main trainshed to main train shed could easily be 5-7 or so, but main train shed to/from 13/14 probably needs to be 15, it might as well be a separate station.

Similarly Liverpool Lime St between platforms on one side could be 5, between halves could be 7 or 8, to/from Low Level would again be 10-12 by the time you've faffed about unless you use the not very well known fare dodgers' lift.

 

[Peter0124]

Its possible to make Platform 1/2 to lower level of Glasgow Central in about 3 or 4 minutes, if its not busy.

 

[Bletchleyite]

Its possible to make Platform 1/2 to lower level of Glasgow Central in about 3 or 4 minutes, if its not busy.

You can't really set MCT based on what an expert user can do if they're lucky, though. It needs to be a time that is always achievable by an average to slow, inexperienced user.

I can do Euston tube to 8-11 up the spiral stairs in under 5 (avoiding escalator and tunnel congestion) but you could never make that the default.

 

[Topological]

Passengers who feel they would like longer can ask for longer. Indeed, journey planners could default to longer.

That is not raising the MINIMUM connecting time, it is just changing the recommended connection time.

You would be right to summarise that I know what I am doing. I would do 5 minutes at Birmingham New Street if the computer would let me. However, connections are rarely priced for that and so unless on a flexible ticket it is a long wait for some journey pairs.

 

[lachlan]

In most cases I would rather chance the connection and enjoy the delay repay if it is missed.

Across London some of the times could be reduced significantly and one could still make the connection. Shortening these would make rail more competitive for long journeys.

Booking websites including this website's own allow extra time to be requested for connections, if needed.

 

[35B]

Passengers who feel they would like longer can ask for longer. Indeed, journey planners could default to longer.

That is not raising the MINIMUM connecting time, it is just changing the recommended connection time.

You would be right to summarise that I know what I am doing. I would do 5 minutes at Birmingham New Street if the computer would let me. However, connections are rarely priced for that and so unless on a flexible ticket it is a long wait for some journey pairs.

You could. But the point of the time is to provide an average that an average person can use as the basis for planning. I regard my wife as a good test case - alert, capable of planning, but ill at ease with being rushed or having to process sudden changes of plan.

That needs to be meaningful for the majority of the able-bodied population and should not assume that they have any prior knowledge - including the sort of prior knowledge that would require applying a longer connection time in a planner.

And as a mean average can conceal very broad variances, that means that those like you are a poor guide to what is possible.

 

[Topological]

You could. But the point of the time is to provide an average that an average person can use as the basis for planning. I regard my wife as a good test case - alert, capable of planning, but ill at ease with being rushed or having to process sudden changes of plan.

That needs to be meaningful for the majority of the able-bodied population and should not assume that they have any prior knowledge - including the sort of prior knowledge that would require applying a longer connection time in a planner.

And as a mean average can conceal very broad variances, that means that those like you are a poor guide to what is possible.

There seems to be a big misrepresentation here.

Why does the minimum connection time need to be increased?

The minimum connection time is the time used to calculate connections that are possible. All I am arguing is that delay repay can be paid on the current minimum connections.

 

[35B]

There seems to be a big misrepresentation here.

Why does the minimum connection time need to be increased?

The minimum connection time is the time used to calculate connections that are possible. All I am arguing is that delay repay can be paid on the current minimum connections.

Your suggestion is that the minimum connection time does not need to be increased. I am suggesting that they need review to assess whether they remain valid.

Delay Repay is a useful mechanism, but is a fallback when things do not go right - journey planners should operate on the basis of what a normal member of the public can reasonably achieve, and give them a result that allows them to make their journey.

 

[The exile]

The danger is that many journeys will suddenly appear to take 30 or even 60 minutes longer than at present and thus become hopelessly unattractive. Indeed , There would be some cases where extending the connection time would preclude a “connection” that actually entails remaining seated in the same train.

 

[Bletchleyite]

Indeed , There would be some cases where extending the connection time would preclude a “connection” that actually entails remaining seated in the same train.

All such services should be advertised as through services to prevent this issue occurring.

 

[mmh]

Currently, people can make their own judgement. Those of us who want the fast connection are offered it, whilst those who want to allow longer can.

In many cases people have no choice.

Take for example the typical connection between the North Wales Coast and Chester and the north WCML to Glasgow at Warrington Bank Quay. It's a 5 minute minimum connection, and I'm sure theoretically the 7 minute connection northbound is fine given it's cross platform. In practice, it's too short and relies on the TFW train being on time, which often it isn't - not helped by some trains having a ridiculously optimistic short dwell time at Chester. Nobody will make their own judgement to plan to spend an extra hour on the platform at Warrington (surely one of the least appealing major intercity interchange stations with the poorest of facilities)

Southbound, the typical connection is 50+ minutes, which is far too long.

Fiddling with minimum connection times won't help with poor timetabling.

 

[Russel]

Only the railway would have anything to gain from extending the minimum connection time.

For those who would like a longer personal minimum connection time, then surely plan appropriately?

No need to invest in complicated data-driven solutions on platforming etc.

Short minimum connection times only work when everything is running to time, even a couple of minutes delay, which is far from an uncommon occurrence can make things uncomfortable.

The current trend of sloppy performance and tight timetabling means whenever a journey planner gives me a 10 minute connection, I'll also be looking at what fallback options there are when things inevitably start to go wrong.

 

[Purple Train]

I can see both sides to this. My regular commute has a 1-minute connection at Slough, which is via a footbridge. I would say I make it about 90% of the time, but then I am a relatively quick walker, and it's from a terminating train which usually arrives bang on time onto a train which is reliably 1-2 minutes late. Contrast this to when I had a six-minute connection at Ashford International, from a coast-bound HS1 service running about 30 seconds down onto a Marshlink train starting from Ashford, was at the wrong end of the train etc. - that felt very tight indeed. Ultimately I think those of us who know all the shortcuts, tricks, and so on are precisely not the people for whom MCTs were made. If the railway was run for RailUKforums then "normals" would probably be sworn off it for life. Likewise, the minimum connection time should be for the average person - not for the hypothetical slowest person of all time. It is a balance - assume the train is running to time as otherwise you may as well make it infinite, but assume that the passenger got off the wrong end of the train as there will be someone in that position near enough to 100% of the time, and so on.

 

[TheGrew]

I often have a short 5ish-minute connection at Warrington Bank Quay, which thankfully normally involves either a cross-platform or a same-platform interchange. However, it doesn't take much of a delay on Avanti to miss it in the northbound direction, so I have done quite well out of delay repay. If it's a long time until the next service, I have often jumped in a local cab using the delay repay money to pay for my onward travel to Newton-le-Willows. Southbound, I normally aim for the train before the one I actually need to get on because of Castlefield/Calder Valley delays.