Specifying additional connection time - how should this work?

[AlterEgo]

You can plan a journey using those tools, but they will not sell you an itinerary for your planned journey.

Indeed. So buy separate tickets if you want. The railway is not going to open itself up to risk of expensive obligations because one wants it to agree to a connection that involves a fit and tall man power walking and optimising every door and escalator.

 

[redreni]

That's not quite right, and is my point entirely.

There is no Minimum Connection Time between two stations. The MCT applies to the station.

The minimum total transfer time between two stations a journey planner should allow is MCT (A) + Fixed Link (A->B) + MCT (B).

Correct.

It's not unreasonable for St Pancras -> King's Cross to have a similar total transfer time allowance as Whitechapel to Fenchurch Street

It is. Have you done both of these transfers?

because you just walk between St Pancras and King's Cross platforms, whereas you're not expected to do that with Whitechapel to Fenchurch Street; you're expected to take the Tube option.

So what? It's about how long it might reasonably take compared with how much time journey planners must allow. It is perfectly feasible to step off the Thameslink at St Pancras and step on a train at Kings Cross 5 or 6 minutes later. I have done so. But I defy anyone to step off a train (Overground or Elizabeth Line) at Whitechapel and step onto a train at Fenchurch Street 5 or 6 minutes later; it can't be done. In the latter case, I wouldn't buy Advances on an itinerary that only allowed the MCTs plus Tube transfer time between those stations: it is too tight to be sure of making it.

The Fixed Links and the MCTs account for that.

Fine; they don't correspond to reality, though. Nor are different interchanges remotely consistent with each other in the manner in which, or the degree to which, they fail to correspond to reality. And there's no easy way for a passenger who is unfamiliar with station layouts to know if they've been given a connection with loads of slack or a tight one.

Comparing the two makes little sense.

Hard disagree. It doesn't matter how the total time is arrived at or how the transfer is expected to be made, what matters is the relationship between time needed and minimum time journey planners must give.

And in any case, they are not the 'MCT' for those transfers. They're the sum of MCTs at either end + the Fixed Link.

True, but I really don't think this affects my point at all.

== Doublepost prevention - post automatically merged: Yesterday at 22:23 ==

Indeed. So buy separate tickets if you want. The railway is not going to open itself up to risk of expensive obligations because one wants it to agree to a connection that involves a fit and tall man power walking and optimising every door and escalator.

I genuinely don't believe it's about that. If it were, the MCTs at stations like Lewisham or Lewes would be more than 4 minutes.

For my money this is about reluctance to change anything combined with a fear that a more consistent relationship between how much time most people need and MCTs would see some of the sillier cross-London transfer times that currently serve as delay repay killers shortened.

 

[styles]

True, but I really don't think this affects my point at all.

Well it does if you are saying there are Minimum Connection Times between the stations you've quoted. That's not how MCTs work - they don't apply to transfers between stations at all so your point is, to be blunt, invalid, because it's based on incorrect data.

You said they were MCTs. They're not.

If we excuse the misuse of the term Minimum Connection Time it is still highly relevant because the point is that nobody is expected to get a Tube train between St Pancras and Kings Cross, but they are between Whitechapel and Fenchuech Street.

If we were to take this example further, the transfer time between between St Pancras and Nottingham could be given, but it would be silly, because the operators don't expect you to walk between these stations. You can if you want, but you'll have to buy separate tickets, otherwise the operators would be liable for bizarre missed connections.

 

[Starmill]

Never going to be acceptable to the industry which has obligations on compensation and accommodation etc if you miss it.

Generally the pressure internally is the other way around. The timetable offers some popular connections as is and its structure means that giving them longer "buffers" is nigh on impossible, without a full recast of the entire timetable. Customers want longer time to connect. Extending the minimum times risks damaging sales so it's rarely done because of this objection.

 

[redreni]

Well it does if you are saying there are Minimum Connection Times between the stations you've quoted. That's not how MCTs work - they don't apply to transfers between stations at all so your point is, to be blunt, invalid, because it's based on incorrect data.

You may not like or agree with my terminology but the minimum time a journey planner must allow between timetabled arrival at St Pancras and timetabled departure from Kings Cross is 31 minutes, exactly as I said originally and others have pointed out to you. That was exactly my point and it's exactly true.

You said they were MCTs. They're not.

If we excuse the misuse of the term Minimum Connection Time

Much obliged, thank you.

it is still highly relevant because the point is that nobody is expected to get a Tube train between St Pancras and Kings Cross, but they are between Whitechapel and Fenchuech Street.

Quite. Which makes the transfer time needed for the fixed link between St Pancras and Kings Cross (a) much shorter than between Whitechapel and Fenchurch Street and (b) much more predictabl. You'd think greater unpredictability would mean more of a safety margin, but the opposite seems to be true here.

The main reason the overall transfer time St Pancras to Kings Cross (MCT + fixed link + MCT) is manifestly excessive has less to do with the transfer time and more to do with only one of the MCTs needing (imho) to be counted. The MCT is supposed to allow for arriving (possibly at the wrong end of the platform), getting off the platform (possibly via a ticket barrier which might be crowded) and finding your way to your departure platform (possibly via another ticket barrier) and boarding. But with a transfer between Kings Cross and St Pancras you only need half the time at each station because at each station you only need to alight and find your way out or find your departing platform and make your way there, but the system is allowing enough time to do both things at both stations.

If we were to take this example further, the transfer time between between St Pancras and Nottingham could be given, but it would be silly, because the operators don't expect you to walk between these stations. You can if you want, but you'll have to buy separate tickets, otherwise the operators would be liable for bizarre missed connections.

You've lost me. A tube transfer isn't timetabled - an assumption is made about how long you need because the stations are close enough together to have defined fixed links. Nobody is calling for the creation of preposterous fixed links between stations in different cities. If such a link were created it would still be possible to analyse the relationship between the time somebody might reasonably require to make the connection and the transfer time plus MCTs that would be specified in the data. I don't understand why you think the comparison of how long a transfer realistically takes versus what the industry data specifies as the minimum - which then enables further comparison of the outcome of those comparisons for different stations or linked station pairs - becomes invalid when fixed links are involved? Maybe you think it's off-topic but I don't think so: passengers who think the journey planner isn't giving them enough time overall - irrespective of whether it's the MCT or a fixed link that they think is too short (they probably wouldn't know or care either way) - can rectify this by adding extra connection time, which is the topic.

 

[megabusser]

As many people now rely on Journey Planners like CityMapper and Google Map to plan the journeys, and the connection time are far less than the MCT set in NR journey planners.

CiryMapper shows a 1-minute connection from Glasgow Queen Street High Level to Low Level in its journey planners. Even being at the optimal door and running, that’s not achievable unless one train is running early/late. Similarly, Google Maps will give very little connecting time for a local bus to long-distance rail connection (eg Edinburgh local bus to Waverley to get a Kings Cross train), which only the very very brave (or those with very expensive flexible tickets) would opt for.

I’m surprised there’s not some kind of machine learning/AI to help CityMapper (and Google Maps) work out what a realistic connection is. I certainly don’t trust them at present.