Swindon - London journey times - What's Going Wrong?

[RichW1]

I have been geeking out at the GWR timetable and noticed the Swindon to London time is 56mins thereabouts. I think the track mileage is 76miles or so? This means something between London and Swindon is seriously wrong! Ipswich is 69 miles and that's an hour... how is GWR managing such times? Where's the slow section? To average 80mph or so on 125mph trunk route is abysmal.

What's going on, where and why?!

 

[yorkie]

I have been geeking out at the GWR timetable and noticed the Swindon to London time is 56mins thereabouts. I think the track mileage is 76miles or so? This means something between London and Swindon is seriously wrong! Ipswich is 69 miles and that's an hour... how is GWR managing such times? Where's the slow section? To average 80mph or so on 125mph trunk route is abysmal.

What's going on, where and why?!

Paddington to Swindon is 77 miles 20 chains
Liverpool St to Ipswich is 68 miles 62 chains

In order to reach the claimed 1 hour journey time for Ipswich, you can only achieve this in the 'Down' direction; the 1700 from Liverpool St runs non-stop to Ipswich in 59min (69.9mph average), but a more typical journey time for a fast train is 1hour 7min, so you're comparing a headline Ipswich non-stop journey time with an average Reading & Didcot stopping Swindon journey time.

As you are using a faster than normal journey time for Ipswich, you need to do the same for Swindon to be fair. The 1645 from Paddington does the journey in 52min with a call at Reading (89.1mph average, even taking into account the dwell time at Reading).

In the 'Up' direction, there is no evidence of 1 hour journey times from Ipswich, while the quoted Swindon journey time can be beaten - slightly. If you depart Swindon at 0801, the journey time to Paddington is 53min, while from Ipswich the 0820 takes 1hour 4min, and that seems to be the fastest possible time. So it's an extra 11 minutes from Ipswich, and the distance is shorter.

If you ignore headline times and look at more typical journey times to London, it gets even worse for Ipswich! There are typically 4 off-peak trains per hour for each, but Swindon has more uniform gaps between trains, and from Ipswich journey times are typically 15 to 20 minutes longer than from Swindon.

 

[pdeaves]

In addition to all that yorkie says, remember that people like to get on and off trains at intermediate points and this costs time on each occasion. Lots of time. Far more than just the dwell time. This drags the average speed down.

 

[Mag_seven]

the average journey time on 125's on the GWML has declined markedly since the 125's were introduced mainly due to:

1. more stops being introduced (e.g. at Didcot and Reading in particular)
2. more congestion due to more trains on the network (e.g. HEX)
3. more recovery time to meet externally imposed targets

(on 3 above I'm sure that some of the recovery time introduced post Hatfield has never been removed!)

 

[DaleCooper]

It's about 52 min. from Peterborough to Kings Cross (just over 76m I think). I seem to remember that was under 50 min. in the past.

 

[RichW1]

King's Cross to Peterborough was always 45mins. I was not aware it had changed! Timetables get slower every time I look, with a couple of minutes here, a couple of minutes there. Network Rail tap about track capacity, but they have been busy ripping track up for miles alongside mainlines that could have been used for years, so it's a bit rich.

It is a little sad that 80mph and 90mph to Parkway is about the average for the core only GWR route Parkway to Paddington, when WCML and ECML exceed this comfortably. The GWR has really suffered over the recent decades hasn't it.

Anyway, there we are, nothing's going to change for many years to come unfortunately. Arbitrary 125mph limits, and stations receiving Intercity trains that should not have them. The market led railway is not helpful to city to city journey times.

 

[glbotu]

King's Cross to Peterborough was always 45mins. I was not aware it had changed! Timetables get slower every time I look, with a couple of minutes here, a couple of minutes there. Network Rail tap about track capacity, but they have been busy ripping track up for miles alongside mainlines that could have been used for years, so it's a bit rich.

It is a little sad that 80mph and 90mph to Parkway is about the average for the core only GWR route Parkway to Paddington, when WCML and ECML exceed this comfortably. The GWR has really suffered over the recent decades hasn't it.

Anyway, there we are, nothing's going to change for many years to come unfortunately. Arbitrary 125mph limits, and stations receiving Intercity trains that should not have them. The market led railway is not helpful to city to city journey times.

I mean, that's nothing to do with "Market Led", that's to do with "providing a useful service to passengers". Headline times are less useful than increasing capacity on an increasingly constrained railway. Sure, 45mins Peterborough - King's Cross would be great, but people at Stevenage being able to physically board trains is better.

 

[DaleCooper]

King's Cross to Peterborough was always 45mins. I was not aware it had changed!

I've just checked and KGX to PBO at 45 min. is faster than PBO to KGX, it must be downhill.

 

[yorkie]

It's about 52 min. from Peterborough to Kings Cross (just over 76m I think). I seem to remember that was under 50 min. in the past.

That's correct in terms of the typical timetabled journey time from Peterborough to King's Cross.

King's Cross to Peterborough was always 45mins. I was not aware it had changed!

King's Cross to Peterborough is indeed typically timed for around 45 minutes and has not changed.

It's rather like Liverpool Street to Ipswich, or many other examples! Booked times are faster in the 'down' direction out of London terminals than in the 'up' direction.

Of course, the reverse is typically true between the penultimate stop and the final destination at the 'country' end of the lines

(and we are not allowed to use that word beginning with "p" and ending in "g"!)

 

[berneyarms]

Recovery time is generally added on the last leg of journeys - hence the differential.

 

[Greenback]

Station dwell times at Reading seem to have increased since I worked there. I suppose that's the result of ever increasing usage. The journey times between Reading and Paddington also seem to have been extended by a few minutes across the board.

 

[D1009]

Station dwell times at Reading seem to have increased since I worked there. I suppose that's the result of ever increasing usage. The journey times between Reading and Paddington also seem to have been extended by a few minutes across the board.

The huge investment in infrastructure at Reading should have led to reduction in journey times, and many trains sit there waiting time at present. It did not make sense to accelerate the trains while electrification work is still going on however. Once the full IEP timetable starts, there will be improvements.

 

[Wilts Wanderer]

At present, the Paddington-Swindon start-stop journey time with stops at Reading and Didcot Parkway generally breaks down as follows (from memory);

Paddington-Reading sectional running time: 23 mins
Engineering allowance at Kennet Bridge Jn: 1 min
Reading station dwell: 2 mins
Reading-Didcot Parkway sectional running time: 11.5 mins
Engineering allowance at Didcot Parkway: 1 min
Didcot Parkway station dwell: 1.5 mins
Didcot Parkway-Swindon sectional running time: 13 mins
Engineering allowance at Swindon: 1 min

The above totals 54 mins. Removing the Didcot stop will give you about 4.5 mins back on paper. So this suggests you can do Padd-Swindon calling at Reading in about 50 mins with a clear run, of which 3 mins is engineering 'padding' for TSRs etc.
Where the reality gets complicated is the way in which individual sectional times are calculated by Network Rail. There are four types; pass-pass, start-pass, pass-stop and start-stop. For a given timing load (for instance HST2+8) SRTs aee generally calculated between all adjacent mandatory timing points, plus all conditional points that are of use to the planners.

Take Didcot PW-Wantage Road-Challow as an example of three consecutive mandatory points. The Wantage Road-Challow pass-pass SRT is (I recall) 3 mins. This SRT will be used by both trains that have, and have not, called at Didcot PW. However, the train that has stopped will not be travelling at 125mph due to the slow acceleration profile of an HST. So the p-p calculation takes the worst case scenario and is most likely modelled against a train averaging 100-110mph over this short section. The difference might be 0.5 mins which doesn't sound like much, but many similar instances through an end-to-end run adds several minutes to the on-paper non-stop run. The worst in my experience is non-stop through Reading on the down, which can result in 5-6 min early running by Didcot. (1748 Padd-Cheltenham is first stop Didcot, and waits time every night.)

Really the flaw is software-based planning and the sheer volume of data meaning that tailored SRT adjustments based on each individual train's characteristics just isn't viable any more. And padding = improved performance (in theory) so there's little incentive to change matters.

 

[A Challenge]

Is there scope to allow for example the 17:48 to have a few minutes taken off its running times post Swindon (dwell time removed on paper and then it just arrives early if you don't want to commit to an earlier advertised arrival) in cases where this is known?

 

[Wilts Wanderer]

Is there scope to allow for example the 17:48 to have a few minutes taken off its running times post Swindon (dwell time removed on paper and then it just arrives early if you don't want to commit to an earlier advertised arrival) in cases where this is known?

It is possible (although not always a good idea) to advertise subsequent calling times earlier than 'working', and sometimes there are extremely practical reasons for doing so. A long chain of request stops such as the Heart of Wales for example, where few of them will actually be served on a given day and the train risks running early. However with Intercity trains it is not advisable and often not possible due to preceding services on the timetable graph. So the 1748 for example, chances are if it left Didcot earlier (say minus-advertised by 3 mins) it would just catch up with the 1745 Padd-Swansea, which is itself following the 1730 Padd-Taunton. Etc etc.

 

[A Challenge]

It is possible (although not always a good idea) to advertise subsequent calling times earlier than 'working', and sometimes there are extremely practical reasons for doing so. A long chain of request stops such as the Heart of Wales for example, where few of them will actually be served on a given day and the train risks running early. However with Intercity trains it is not advisable and often not possible due to preceding services on the timetable graph. So the 1748 for example, chances are if it left Didcot earlier (say minus-advertised by 3 mins) it would just catch up with the 1745 Padd-Swansea, which is itself following the 1730 Padd-Taunton. Etc etc.

I see, yes, swap it with the previous path Paddington to Reading?

 

[Wilts Wanderer]

That would be changing the question, though. The position of each train schedule on the graph (particularly the down evening peak) is one of extremely long discussion and strategic importance! Generally it is preferable to flight fast trains ahead of slower ones but in the peaks it is simply about trying to fit as many trains through the Paddington throat as possible. Also bear in mind that chances are the HST set for the 1748 arrives Padd at about 1720-30ish and to run earlier would reduce the turnaround time, and also have an impact on the platforming plan, which is a full time job in of itself.

I don't intend to belittle the question, but generally the timetable is the best it can be with the current set of 'givens' and every possible option to improve it (balancing performance, capacity and revenue) has been considered.

Introducing electric IETs will change one existing 'given'; they accelerate so quick up to 125mph that the padded SRT issue will disappear.

 

[The Planner]

Where the reality gets complicated is the way in which individual sectional times are calculated by Network Rail. There are four types; pass-pass, start-pass, pass-stop and start-stop. For a given timing load (for instance HST2+8) SRTs aee generally calculated between all adjacent mandatory timing points, plus all conditional points that are of use to the planners.

Take Didcot PW-Wantage Road-Challow as an example of three consecutive mandatory points. The Wantage Road-Challow pass-pass SRT is (I recall) 3 mins. This SRT will be used by both trains that have, and have not, called at Didcot PW. However, the train that has stopped will not be travelling at 125mph due to the slow acceleration profile of an HST. So the p-p calculation takes the worst case scenario and is most likely modelled against a train averaging 100-110mph over this short section. The difference might be 0.5 mins which doesn't sound like much, but many similar instances through an end-to-end run adds several minutes to the on-paper non-stop run. The worst in my experience is non-stop through Reading on the down, which can result in 5-6 min early running by Didcot. (1748 Padd-Cheltenham is first stop Didcot, and waits time every night.)

Really the flaw is software-based planning and the sheer volume of data meaning that tailored SRT adjustments based on each individual train's characteristics just isn't viable any more. And padding = improved performance (in theory) so there's little incentive to change matters.

In which case they are fudging it, if there is that sort of difference then there should be an adjustment value in the TPR approaching Challow. There is no reason why SRTs should be like that. Its even worse if people are calculating SRTs individually between timing points and not considering the cumulative effect. I would argue that individual allowances are more important than ever.

 

[Greenback]

The huge investment in infrastructure at Reading should have led to reduction in journey times, and many trains sit there waiting time at present. It did not make sense to accelerate the trains while electrification work is still going on however. Once the full IEP timetable starts, there will be improvements.

I hope we can get back to the sort of times we had in the late 1990's. From what I cna remember, HST's were scheduled to depart Reading 25 minutes after leaving Paddington.

 

[Richard_B]

I hope we can get back to the sort of times we had in the late 1990's. From what I cna remember, HST's were scheduled to depart Reading 25 minutes after leaving Paddington.

Did the regularly make that though or were they often a minute or two late?

 

[455driver]

I hope we can get back to the sort of times we had in the late 1990's. From what I cna remember, HST's were scheduled to depart Reading 25 minutes after leaving Paddington.

Only if you remove the data recorders!

 

[D1009]

Did the regularly make that though or were they often a minute or two late?

IIRC when the HSTs started the Paddington to Reading start to stop time was 22 mins, which was regularly achieved, though the trains did not necessarily keep to 125 mph. It was increased to 23 mins when the trains were lengthened to 8 coaches, whenever that was. Nowadays we have a more restricted first mile or so from Paddington and "professional" driving.

 

[Greenback]

Did the regularly make that though or were they often a minute or two late?

They regularly made it!

 

[Wilts Wanderer]

The advent of Central Door Locking increased minimum dwell times by 30 seconds. Defensive driving also plays a major part these days as well, particularly when it comes to braking profiles. No more hitting platforms at 50mph with full brake applications.

 

[Mintona]

The advent of Central Door Locking increased minimum dwell times by 30 seconds. Defensive driving also plays a major part these days as well, particularly when it comes to braking profiles. No more hitting platforms at 50mph with full brake applications.

Oh I don't know, there are a few platforms you can hit at 50 without any need for a full brake application.