Central line working timetable.

[bakerstreet]

Anyone know if there is a recent new Central line timetable in force?

I’ve been using the line over the past few weeks and on numerous occasions I’ve had to wait 13-15 mins for an Epping train.

When it arrives - even during the middle of the weekday - it’s packed.

And the exit gates at Epping can - literally - take 4-5 mins to clear.

Having used the line previously there were Epping trains every 5-10 mins with the odd Loughton.

Often during the day both platforms at Epping might have trains or a very short period where no train was in any platform.

Now Epping can be clear of trains for 10 mins or more.

Line reports ‘good service’

Also Westbound I’m seeing more terminating short at Northolt and North Acton than I remember.

Northolt in particular used to be a rarity during the day.

Perhaps I’ve just been unlucky but I can’t recall the trains being this crowded - even at the extremes of the line - during a weekday.

 

[Dstock7080]

Current timetable introduced 26/01/2020.
Northolt reversers every 20min off-peak

 

[bakerstreet]

Thanks - I guess I’ve just been unlucky over recent weeks. Just looking up the station boards on TfL Go and lots of trains to and from Epping

 

[gingertuber]

I had the same thing. It's not a line I use regularly, but I needed the Epping branch to go on a Sunday walk around the forest, and I had to wait 19 minutes for the next one, with about a million Hainault trains running past. Very odd. I imagine they've just crunched the numbers and that is genuinely the optimal branch frequencies?

 

[bakerstreet]

I had the same thing. It's not a line I use regularly, but I needed the Epping branch to go on a Sunday walk around the forest, and I had to wait 19 minutes for the next one, with about a million Hainault trains running past. Very odd. I imagine they've just crunched the numbers and that is genuinely the optimal branch frequencies?

Exactly my experience- plenty of Hainault trains but even then there was at one stage a 7 min gap in eastbound services to any destination

 

[notverydeep]

Don't forget that you will only see multiple trains to the same destination / branch when they are not going to your destination - because on the occasions when there were, you got the first one!

 

[infobleep]

Don't forget that you will only see multiple trains to the same destination / branch when they are not going to your destination - because on the occasions when there were, you got the first one!

Surely you should see multiple trains to your destination too as they would list the trains following your one and not just the next train you are catching.

 

[notverydeep]

Surely you should see multiple trains to your destination too as they would list the trains following your one and not just the next train you are catching.

Yes indeed, this will be true to a degree - but in my experience whenever there is a branch split, passengers always have a tendency to think the other branch is favoured. They will notice occasions when they have to pace up and down as the third (lightly loaded) train to the other branch comes through, before a farily crowded train to their branch arrives. The next day though, they arrive at the platform as their train comes in and get on barely glancing up at the train list...

Interestingly, this split (in passenger number terms) at Leytonstone is the most even across the network - there are plenty of other locations where services diverge with considerably different passenger numbers split - but it is hard to schedule anything that isn't basically 50:50. To see why this is, imagine a service that passes the central section at 2 minute intervals evenly spaced: A B A B A B A B A B etc. Then try and make the split across the 10 trains 60:40 and see what happens to the headways once the services diverge!

 

[infobleep]

Yes indeed, this will be true to a degree - but in my experience whenever there is a branch split, passengers always have a tendency to think the other branch is favoured. They will notice occasions when they have to pace up and down as the third (lightly loaded) train to the other branch comes through, before a farily crowded train to their branch arrives. The next day though, they arrive at the platform as their train comes in and get on barely glancing up at the train list...

Interestingly, this split (in passenger number terms) at Leytonstone is the most even across the network - there are plenty of other locations where services diverge with considerably different passenger numbers split - but it is hard to schedule anything that isn't basically 50:50. To see why this is, imagine a service that passes the central section at 2 minute intervals evenly spaced: A B A B A B A B A B etc. Then try and make the split across the 10 trains 60:40 and see what happens to the headways once the services diverge!

Could one not do A B B A B B A B B? Where A is the less used service.

 

[Basil Jet]

Could one not do A B B A B B A B B? Where A is the less used service.

They do ABBA at Waterloo.

 

[notverydeep]

Could one not do A B B A B B A B B? Where A is the less used service.

Yes it is easier to do a 2/3 1/3 split than a 60:40 split, but to illustrate why this is not optimal, we can do a basic waiting time calculation. Let us assume that there are 60 passengers per minute who want to travel to the busier branch and 40 per minute to the less busy branch and that they all arrive at the station randomly which will be true of this frequency. On the busier, the train intervals are 4 minutes then 2 minutes repeating compared to every 4 minutes previously. On the less busy branch the intervals have increased from every 4 minutes to every 6 minutes.

In the original scenario everyone has an average waiting time of 2 minutes (half the train interval). With the A B B A B B A B B service, everyone to branch A (less busy) now has an average waiting time of 3 minutes. Passengers on branch B have an average waiting time of 2 minutes, two thirds of the time, and 1 minute one third of the time, which resolves to 1.67 minutes for all branch B passengers (2/3*2)+(1/3*1).

So to compare the total waiting time for all passengers with each option:
A B A B A B A B etc.
Branch A: 2400 passengers per hour (40*60) waiting 2 minutes = 4800 minutes
Branch B: 3600 passengers per hour (60*60) waiting 2 minutes = 7200 minutes
Total = 12000 minutes

A B B A B B A B B etc.
Branch A: 2400 passengers per hour waiting 3 minutes = 7200 minutes (an increase of 2400 minutes)
Branch B: 3600 passengers per hour waiting 1.67 minutes = 6000 minutes (a reduction of 1200 minutes)
Total = 13200 minutes (an increase of 1200 minutes)

So you can see with the 2/3 1/3 split, even though it seems a better fit for a 60:40 branch passenger split, causes a greater increase in journey time to the 'losing' branch than the improvement in journey time to the 'gaining' branch.

To be fair, there are some factors that make the answer a bit more complicated - the concept of "weighting" elements of journey time - this is the process of applying a penalty to the parts of the journey that passengers dislike more than others. Waiting time, crowded in vehicle time and interchange time all attract a penalty (a multiplier of the standard value of time or a specific time penalty). The 'weighting' attached to waiting time is impacted by frequency and how randomly passengers arrive at the platform. You don't turn up at random for an hourly service, some people will and others won't for a 20 minute service, but most do for intervals under 10 minutes. To reflect this, the weighting of waiting time declines as the interval increases reflecting that the time is spent in a comfortable environment (at home or the office for example before setting out for a specific train), but there still is a penalty, reflecting the reduced opportunity to travel at the exactly desire time...

I know this answer has gone on a bit, but I hope it gives some insight about how these sort of decisions are made.

 

[2192]

It has to be an equal split on each of the branches, otherwise the branches don't have the same time interval between trains.

 

[FormerBritInMe]

What is the current arrangement for the Hainault loop? Do through trains to Hainault run via Newbury Park and a shuttle run between Woodford and Hainault? (I believe that's how it was at one time). Are there peak hour trains to Hainault via Roding Valley?

 

[notverydeep]

What is the current arrangement for the Hainault loop? Do through trains to Hainault run via Newbury Park and a shuttle run between Woodford and Hainault? (I believe that's how it was at one time). Are there peak hour trains to Hainault via Roding Valley?

There is now a shuttle service of 3 trains per hour (tph) using two sets, that will at some point later in the year become 4 car trains, though the pandemic has set that back considerably. This operates all day, but one of the sets is changed over in the early evening (and will remain an 8 car set once 4 car operation starts). There are a small number of trains from the western end of Hainault Depot that enter service at Grange Hill and travel to Central London via Woodford and these are extra over the normal 3 tph. One train comes the other way in the PM peak, terminating at Grange Hill and stabling at the western end of Hainault Depot. There are a number of positioning moves to / from Leytonstone or Woodford early and late to get trains between Hainault Depot and the Epping Branch.

Previously (before WTT70 in January 2020) the majority of the Hainault to Woodford service (again 3 tph) was provided by extending trains between (generally) Ealing Broadway and Hainault via Newbury Park on towards Woodford, which returned to Hainault before continuing to Central London via Newbury Park.

 

[FormerBritInMe]

There is now a shuttle service of 3 trains per hour (tph) using two sets, that will at some point later in the year become 4 car trains, though the pandemic has set that back considerably. This operates all day, but one of the sets is changed over in the early evening (and will remain an 8 car set once 4 car operation starts). There are a small number of trains from the western end of Hainault Depot that enter service at Grange Hill and travel to Central London via Woodford and these are extra over the normal 3 tph. One train comes the other way in the PM peak, terminating at Grange Hill and stabling at the western end of Hainault Depot. There are a number of positioning moves to / from Leytonstone or Woodford early and late to get trains between Hainault Depot and the Epping Branch.

Previously (before WTT70 in January 2020) the majority of the Hainault to Woodford service (again 3 tph) was provided by extending trains between (generally) Ealing Broadway and Hainault via Newbury Park on towards Woodford, which returned to Hainault before continuing to Central London via Newbury Park.

Thanks @notverydeep for your detailed response. I have always found the Hainault Loop an interesting aspect of LU operation.

 

[Mike99]

I hope this works and is a help to you
https://tfl.gov.uk/cdn/static/cms/documents/cen-wtt-70.pdf
Full Working Timetables are on the TfL site just a bit tucked away

 

[notverydeep]

I hope this works and is a help to you
https://tfl.gov.uk/cdn/static/cms/documents/cen-wtt-70.pdf
Full Working Timetables are on the TfL site just a bit tucked away

@FormerBritInMe Look for trains 171D and 172 plus Trains 23D (weekdays), 146D (Saturdays) or 74D (Sundays) which take over from 171D (172 on Sundays) on shuttle workings in the evenings. Because of the change of direction at Hainault, it was a bit harder to follow the through trains between Woodford via Hainault and Newbury Park and Ealing Broadway in previous timetables, which used to have footnotes for this. The occasional through trains to Leytonstone or Central London, such as train 65 (07a58 from Grange Hill on weekday mornings, the last train shown on page 24) can be seen with timings at the loop stations continuing on to the rest of the line)...

 

[infobleep]

Yes it is easier to do a 2/3 1/3 split than a 60:40 split, but to illustrate why this is not optimal, we can do a basic waiting time calculation. Let us assume that there are 60 passengers per minute who want to travel to the busier branch and 40 per minute to the less busy branch and that they all arrive at the station randomly which will be true of this frequency. On the busier, the train intervals are 4 minutes then 2 minutes repeating compared to every 4 minutes previously. On the less busy branch the intervals have increased from every 4 minutes to every 6 minutes.

In the original scenario everyone has an average waiting time of 2 minutes (half the train interval). With the A B B A B B A B B service, everyone to branch A (less busy) now has an average waiting time of 3 minutes. Passengers on branch B have an average waiting time of 2 minutes, two thirds of the time, and 1 minute one third of the time, which resolves to 1.67 minutes for all branch B passengers (2/3*2)+(1/3*1).

So to compare the total waiting time for all passengers with each option:
A B A B A B A B etc.
Branch A: 2400 passengers per hour (40*60) waiting 2 minutes = 4800 minutes
Branch B: 3600 passengers per hour (60*60) waiting 2 minutes = 7200 minutes
Total = 12000 minutes

A B B A B B A B B etc.
Branch A: 2400 passengers per hour waiting 3 minutes = 7200 minutes (an increase of 2400 minutes)
Branch B: 3600 passengers per hour waiting 1.67 minutes = 6000 minutes (a reduction of 1200 minutes)
Total = 13200 minutes (an increase of 1200 minutes)

So you can see with the 2/3 1/3 split, even though it seems a better fit for a 60:40 branch passenger split, causes a greater increase in journey time to the 'losing' branch than the improvement in journey time to the 'gaining' branch.

To be fair, there are some factors that make the answer a bit more complicated - the concept of "weighting" elements of journey time - this is the process of applying a penalty to the parts of the journey that passengers dislike more than others. Waiting time, crowded in vehicle time and interchange time all attract a penalty (a multiplier of the standard value of time or a specific time penalty). The 'weighting' attached to waiting time is impacted by frequency and how randomly passengers arrive at the platform. You don't turn up at random for an hourly service, some people will and others won't for a 20 minute service, but most do for intervals under 10 minutes. To reflect this, the weighting of waiting time declines as the interval increases reflecting that the time is spent in a comfortable environment (at home or the office for example before setting out for a specific train), but there still is a penalty, reflecting the reduced opportunity to travel at the exactly desire time...

I know this answer has gone on a bit, but I hope it gives some insight about how these sort of decisions are made.

Thanks for such a detailed response. Very interesting.

It has to be an equal split on each of the branches, otherwise the branches don't have the same time interval between trains.

In response to both of these and this may not be the correct analogy, so do pull it a apart, take a bus service such as the 281 in London. This doesn't run at precisely the same intervals.

I'm can't remember what is on the bus stop timetable but it's something like running every 7 to 12 minutes. So people have to wait varying amounts of time for it.

Then let's factor in the K3. This runs every 15 minutes, which is less than the 281. They share a common route during part of both journeys. People do howvwer have to wait different amounts of time if they are travelling to a point where they diverge.

Could that not be the case for the tube or is that inefficient use of rolling stock?

 

[Basil Jet]

It has to be an equal split on each of the branches, otherwise the branches don't have the same time interval between trains.

Not at all. Suppose the Central ran 30 tph to North Acton, with 20 tph proceeding to Ealing and 10 tph proceeding to West Ruislip. This would create alternate 4 & 2 minute gaps on the Ealing Branch, but it's only two stations so LUL might not care. If they do care, they only have to make half of the Ealing trains wait one minute at North Acton (in both directions) to have even 3 minute gaps on the branch. On most if not all Underground lines the frequencies are so high that the added waits are barely noticeable, whereas if you were running, say twice as many Overground trains to Enfield Town as you were running to Cheshunt and you wanted even headways all around, you couldn't do it without painful waits.

 

[rebmcr]

Not at all. Suppose the Central ran 30 tph to North Acton, with 20 tph proceeding to Ealing and 10 tph proceeding to West Ruislip. This would create alternate 4 & 2 minute gaps on the Ealing Branch, but it's only two stations so LUL might not care. If they do care, they only have to make half of the Ealing trains wait one minute at North Acton (in both directions) to have even 3 minute gaps on the branch. On most if not all Underground lines the frequencies are so high that the added waits are barely noticeable, whereas if you were running, say twice as many Overground trains to Enfield Town as you were running to Cheshunt and you wanted even headways all around, you couldn't do it without painful waits.

This of course is why the Southern Region operated so many splitting services, despite the additional operational needs to support them.

 

[Ken H]

Thanks @notverydeep for your detailed response. I have always found the Hainault Loop an interesting aspect of LU operation.

Only time i did hainult - woodford it was ATO. It was the test bed for the victoria line trains and they left it ATO. I went there specially to ride the ATO trains there.
BTW there was a time when the loop was called the Fairlop loop.

 

[Bigchris]

On the Woodford-Hainault loop services the current service appears to use 3 sets rather than 2 as above. It's 3 tph with a round trip of about 30 minutes, on arrival at Woodford the set goes off into the siding for a short while then comes back into the platform ready for the next trip to Hainault. Funny this should come up as I did that line last weekend for the first time to try and see the old stock stored at Hainault.

 

[notverydeep]

On the Woodford-Hainault loop services the current service appears to use 3 sets rather than 2 as above. It's 3 tph with a round trip of about 30 minutes, on arrival at Woodford the set goes off into the siding for a short while then comes back into the platform ready for the next trip to Hainault. Funny this should come up as I did that line last weekend for the first time to try and see the old stock stored at Hainault.

A sample round trip: train 171D departs Hainault for Woodford at 10:48 arriving at Woodford WB platform 10:58:30 where it has a 1 minute (b) stand for de-trainment before shunting forward to 21 road to reverse (east of both running lines). It then takes a layover (roughly 8.5 minutes allowing a minute each way for the shunt). It then leaves 21 road, crosses to the Woodford EB platform departing at 11:10 for Hainault, where it arrives at 11:19:30. Here it takes the layover at this end, ready to form the 11:28 departure for Woodford. The round trip, including layovers at both ends thus totals 40 minutes. The other train 172 will have worked the 11:08 departure from Hainault between the above train 171D departures and would be back to work the next trip at 11:48 from Hainault to maintain the 20 minute (3 tph) headway with two sets. As mentioned in earlier posts train 171D gets swapped over after the evening peak (train 172 on Sundays).

Hope that clarifies...

 

[Bigchris]

That makes sense. I think a bit of confusion is caused by using the Intertube website as it shows (rightly or wrongly) additional set changes with the leading car identifiers, although there are certainly occasions when the info on there is wrong.

 

[Route115?]

I think that the problem is cancellations of which there are quite a few at the moment creating gaps. West Ruislip where I line has a 10 - 5 - 5 minute service off peak (5 minutes east of Northolt). If the wrong train is cancelled you have a a 15 minute wait.

Looking at the WTT there are essentially eight 20 minute services overlapping off-peak to provide 24 trains per hour. This makes regulation very difficult compared with other lines. Five minute West Ruislip - Epping & Ealing Broadway - Hainault would make life easier (but involve more mileage).

In the evening you often have successive trains to White City & North Acton which then makes the folowing trains very full. Overall the timetable could be better.

 

[rebmcr]

Five minute West Ruislip - Epping & Ealing Broadway - Hainault would make life easier (but involve more mileage).

Not just more mileage, but more trains to run those extra miles. The current Central timetable has been designed to leave several trains out of passenger service, so that they can be run through the upgrade workshop instead.