Great Western Coffee Shop

Is it me or are things getting worse.  Here are the Plymouth to London Paddington train performance figures for the last 12 weeks:

TOC Scheduled Times           % Arrivals               Actual Arrival
        d PLY    a PAD   Dur          On Time               Average
GW   05:53   09:00   3h 7m     2%            09:14 14L
GW   05:30   09:21   3h 51m   26%            09:29 8½L
GW   05:09   09:44   4h 35m   46%            09:52 8½L
GW   06:55   10:03   3h 8m     11%          10:15 12L
GW   07:48   11:24   3h 36m   18%             11:33 9½L
GW   08:53   12:24   3h 31m   36%             12:33 9L
GW   09:48   13:39   3h 51m   16%             13:50 11L
GW   10:44   13:44   3h 0m     7%             13:57 13½L
GW   12:01   15:21   3h 20m   14%             15:35 14L
GW   12:56   16:22   3h 26m   40%             16:31 9L
GW   15:03   18:21   3h 18m   7%             18:32 11½L
GW   16:02   19:24   3h 22m   8%             19:35 11L
GW   16:57   20:39   3h 42m   7%             20:51 12L
GW   18:03   21:24   3h 21m   27%             21:33 9½L
GW   19:42   23:41   3h 59m   18%         00:09 28L

Data from http://www.recenttraintimes.co.uk

I've been noticing at Liskeard the "up" is often 5 minutes or more late by reaching here, fairly early in its journey. This is observations and often doesn't reflect what the CIS is showing.

Pretty worthless having a 3h 00m headline timing if it's only achieved 7% of the time.

I wonder what the figures are from the South West to Reading?  A lot of delays are incurred with signalling problems that have plagued the stretch from Reading to Paddington.

(Not that the Berks & Hants doesn't have its moments!)

Roughly 10 minutes late on average at Reading with around 50% 'on time' margin for those same services.

Quite a few "slippery rails" over the past couple of weeks

Not much better:

TOC       Scheduled Times       % Arrivals       Actual Arrival
        d PLY   a RDG   Dur           On Time         Average
GW   05:53   08:32   2h 39m   2%                   08:42 10L
GW   05:30   08:47   3h 17m   2%                   08:55 8½L
GW   05:09   09:14   4h 5m     34%                 09:21 7L
GW   06:55   09:32   2h 37m   0%                   09:42 10½L
GW   07:48   10:50   3h 2m     2%                   11:00 10L
GW   08:53   11:50   2h 57m   14%                 11:58 8½L
GW   09:48   13:08   3h 20m   23%                 13:18 10L
GW   10:44   13:16   2h 32m   25%                 13:26 10½L
GW   12:01   14:50   2h 49m   10%                 15:02 12½L
GW   12:56   15:49   2h 53m   2%                   15:58 9L
GW   15:03   17:49   2h 46m   21%                 17:57 8½L
GW   16:02   18:51   2h 49m   2%                   19:01 10½L
GW   16:57   20:08   3h 11m   2%                   20:21 13½L
GW   18:03   20:50   2h 47m   2%                   21:02 12½L
GW   19:42   23:04   3h 22m   8%                   23:11 7½L

Data once again from http://www.recenttraintimes.co.uk

Agreed, although there won't be the padding built into the Reading times as there is at a final destination all too often.

Sorry, I was referring to 'on time' up until 5 minutes late, rather than exactly on time as has been posted by SandT, which shows very grim reading indeed.

I've carried out a similar exercise for Bristol TM, Cardiff and Cheltenham and the figures are at least 50% better than those for the Southwest.  Think that GWR/NR need to do some work here ::)

Despite the GWR overall November 2016 performance figures (http://www.firstgreatwestern.info/coffeeshop/index.php?topic=16825.30) the actuals for the far Southwest seem to be getting worse.

Figures for the past four weeks:

TOC  Scheduled Times             % Arrivals  Actual Arrival
         d PLY   a PAD   Dur           On Time    Average
GW   05:53   09:00   3h 7m     0%              09:25   25L
GW   05:30   09:21   3h 51m   10%            09:41   20½L
GW   05:09   09:44   4h 35m   20%            09:57   13½L
GW   06:55   10:03   3h 8m     5%              10:28   25½L
GW   07:48   11:24   3h 36m   0%              11:39   15½L
GW   08:53   12:24   3h 31m   20%            12:38   14L
GW   09:48   13:39   3h 51m   0%              13:59   20½L
GW   10:44   13:44   3h 0m     5%              14:14   30L
GW   12:01   15:21   3h 20m   15%            15:41   20½L
GW   12:56   16:22   3h 26m   25%            16:36   14L
GW   15:03   18:21   3h 18m   5%              18:35   14L
GW   16:02   19:24   3h 22m   5%              19:42   18L
GW   16:57   20:39   3h 42m   5%              21:03   24½L
GW   18:03   21:24   3h 21m   16%            21:42   18L
GW   19:42   23:41   3h 59m   33%            23:44   3½L

Data from http://www.recenttraintimes.co.uk

Thanks for posting those latest figures, SandTEngineer.  ::)

Here are the latest figures. Monday to Friday data for the last 8 weeks (mid-January to mid-March 2017):

TOC       Scheduled Times            % Arrivals     Actual Arrival
        d PLY    a PAD   Dur             On Time        Average
GW   05:53   09:00   3h 7m         5%                09:13   13L
GW   05:30   09:21   3h 51m       15%              09:31   10L
GW   05:09   09:42   4h 33m       17%              09:49   7½L
GW   06:55   10:01   3h 6m         0%                10:15   14½L
GW   07:48   11:24   3h 36m       41%              11:31   7L
GW   08:53   12:21   3h 28m       13%              12:26   5L
GW   09:48   13:40   3h 52m       55%              13:45   5L
GW   10:44   13:44   3h 0m         30%              13:53   9L
GW   12:01   15:21   3h 20m       46%              15:24   3L
GW   12:56   16:22   3h 26m       48%              16:23   1½L
GW   15:03   18:21   3h 18m       32%              18:24   3½L
GW   16:02   19:21   3h 19m       20%              19:29   8½L
GW   16:57   20:39   3h 42m       25%              20:44   5½L
GW   18:03   21:22   3h 19m       28%              21:30   8½L
GW   19:44   23:41   3h 57m       38%              23:45   4½L

Data from http://www.recenttraintimes.co.uk

The first time, during your date periods, that a train (only the one mind) has been on time more often than not!

I've been doing some research into where the delays are occuring and, not supprisingly, its mostly poor regulating between Slough and Paddington.  It would appear that the signalling system 'Automatic Route Setting' (ARS) equipment is not functioning to its optimum and this is resulting in conflicting movements being signalled that it eventually takes up to 15 minutes to settle down again only for the next conflict to trigger it all off again.  If you watch the Paddington map on Open Train Times (http://www.opentraintimes.com/maps/signalling/d3_1) you can see this happen. Quite often you will see Paddington departing trains being signalled down Line 2 when Line 1 is free, resulting in no alternative to Line 3 for inbound services.  Line 3 has some very restrictive signalling controls that were implemented at the time of the Ladbroke Grove accident and quite often trains are held as far out as Old Oak Common waiting for Line 3 to become available when Line 2 could have been available.

Signallers are instructed not to intervene as they are at risk of discipliniary action if they delay other trains by altering the priorities.

So looks like we are stuck with very rarely having an 'on time' arrival at Paddington (and of course, its all likely to get much worse in 18 months time)  ::) :P

Yes, and that's the one that takes the 'pretty' route via Bristol and Bath, so pleanty of padding built into the timings  ::) :P

I haven't posted an update on here for a while but things seem to be worse since the new January 2018 timetable started...

TOC       Scheduled Times   % Arrivals        Actual Arrival
         d PLY   a PAD   Dur     On Time    Average
GW   23:54   05:03   5h 9m   30%           05:11 8L
GW   04:51   08:37   3h 46m   0%           08:43 6½L
GW   05:53   09:00   3h 7m   0%           09:14 14L
GW   05:29   09:23   3h 54m   0%           09:34 11L
GW   06:53   10:02   3h 9m   10%           10:15 13½L
GW   07:45   11:21   3h 36m   40%           11:31 10½L
GW   08:51   12:21   3h 30m   30%           12:41 20L
GW   09:49   13:14   3h 25m   22%           13:24 10½L
GW   11:00   14:01   3h 1m   20%           14:16 15L
GW   12:01   15:21   3h 20m   12%           15:36 15L
GW   12:55   16:21   3h 26m   40%           16:39 18½L
GW   13:56   17:21   3h 25m   40%           17:24 3L
GW   15:00   18:16   3h 16m   0%           18:30 14L
GW   16:00   19:20   3h 20m   0%           19:29 9L
GW   16:57   20:37   3h 40m   0%           20:53 16L
GW   18:03   21:22   3h 19m   13%           21:35 13L
GW   19:44   23:39   3h 55m   0%           23:44 5L

Data from http://www.recenttraintimes.co.uk

Still no better over the past four weeks..... ::)

TOC       Scheduled Times       % Arrivals       Actual Arrival
        d PLY   a PAD   Dur          On Time       Average
GW   04:51   08:37   3h 46m   0%                   08:49 12L
GW   05:53   09:00   3h 7m   0%                   09:14 14L
GW   05:53   09:02   3h 9m   50%                   09:06 4L
GW   05:53   09:04   3h 11m   0%              09:22 18L
GW   05:29   09:23   3h 54m   10%                   09:33 10½L
GW   06:53   10:02   3h 9m   5%                   10:15 13L
GW   07:45   11:21   3h 36m   25%                   11:28 7L
GW   08:51   12:21   3h 30m   35%                   12:31 10½L
GW   09:49   13:14   3h 25m   5%                   13:22 8L
GW   11:00   14:01   3h 1m   20%                   14:08 7½L
GW   12:01   15:21   3h 20m   15%                   15:24 3½L
GW   12:55   16:21   3h 26m   50%                   16:27 6½L
GW   13:56   17:21   3h 25m   32%                   17:23 2L
GW   15:00   18:16   3h 16m   0%                   18:27 11L
GW   16:00   19:20   3h 20m   0%                   19:29 9L
GW   16:57   20:37   3h 40m   5%                   20:44 7½L
GW   18:03   21:22   3h 19m   20%              21:27 5½L
GW   19:44   23:39   3h 55m   0%                   23:41 2½L
GW   19:44   23:42   3h 58m   12%                   23:51 9L
GW   19:44   23:44   4h 0m   50%                   23:47 3L
GW   19:44   23:53   4h 9m   50%                   23:54 1½L

Apologies for not being able to get the data columns to line up.

Data as always from http://www.recenttraintimes.co.uk

.....and the Down Direction is no better.  So much for the much wanted 'early' morning arrival in Plymouth.

TOC       Scheduled Times       % Arrivals       Actual Arrival
       d PAD   a PLY   Dur           On Time       Average
GW   06:33   09:33   3h 0m   5%                   09:39       6L
GW   07:30   11:11   3h 41m   45%                   11:16       5L
GW   09:03   12:19   3h 16m   35%                   12:31       12L
GW   10:03   13:03   3h 0m   25%                   13:11 8½L
GW   11:03   14:30   3h 27m   15%                   14:40 10½L
GW   12:03   15:05   3h 2m   50%                   15:11 6½L
GW   13:03   16:18   3h 15m   20%                   16:22 4½L
GW   14:03   17:20   3h 17m   40%                   17:22 2½L
GW   15:03   18:37   3h 34m   20%                   18:44 7½L
GW   16:03   19:28   3h 25m   60%                   19:30 2½L
GW   17:03   20:26   3h 23m   45%                   20:29 3L
GW   18:03   21:17   3h 14m   10%                   21:24 7½L
GW   18:33   22:11   3h 38m   0%                   22:21 10L
GW   19:03   22:22   3h 19m   6%                   22:35 13L
GW   19:03   22:24   3h 21m   0%                   22:39 15L
GW   20:03   23:36   3h 33m   10%                   23:51 15L
GW   21:03   00:40   3h 37m   19%                   00:49  9½L

Data from http://www.recenttraintimes.co.uk

Not very impressive, however from my personnel point of view, I am less concerned by regular MODERATE lateness, than by occasional extreme delays.

If a train is routinely 10 minutes late, then that is a poor performance, but from my perspective, this  is actually preferable to being on time 19 times out of 20, but delayed by 200 minutes on one journey in 20.

My most frequent journey from London was on the Golden Hind from Paddington to Taunton, This was often a few minutes late, which was irritating "on principle" but not in fact of any practical consequence.
However an occasional arrival several hours late was much more serious, the last bus had departed hours before, taxis not readily available, and so on.

I would in most circumstances, prefer to be 12 minutes late 100 times, than to be 1,200 minutes late just once.
And yes I have been about 1,200 minutes late!
(I intended to depart London on the 18-03 one evening, waited at Paddington for hours before following the advice "don't travel with us today, try tomorrow". I returned home by taxi at a cost of £30, and tried again the next day, eventually boarding a dangerously overcrowded train the next day. For my improved first class fare of over £150 I got to stand to Taunton, and had to firmly resist efforts by a TTI to "rob" me of another £150. So I lost £35, and a days holiday, and had to stand despite paying the full first class fare, and the total delay must have been about 20 hours)

The Real Time Trains site allows you to look at other metrics too - for example I have asked about the number of services over an hour late into Plymouth (4 weeks , Monday to Friday, so 20 days) - will add the screen capture of the results to this post as (yes), it's messy to get RTT to line up for copy and paste onto the forum.

Worst performer - 09:03 from PAD, over an hour late 3 times.  11:03 was over an hour late twice. and 4 other services were over an hour late once.   Rest  of daytime trains were within the hour each time they ran.

Grahame, just a small correction to your post above.  The site from which I extracted the data is Recent Train Times and not Real Train Times  :)

Noted ... thank.  2018 has started FAR too busy and occasionally (  ;D ) my check back fails

Certainly work to be done on those timings, and it’s great that detailed information is readily available nowadays.

Is punctuality of long distance trains worse than it used to be though?  I don’t have data on West Country services back in the more distant past, but I remember keeping a log for several months on the average lateness of, what were then, InterCity services into Reading from the Birmingham direction in the very early 90s, based on automatic reports into the TRUST system, which back then were accessible through a TOPS terminal for up to a month previous.  It was more to test out the new spreadsheet program I was using on my Amstrad CPC than anything else!

If memory serves me (the computer and the cassettes the data were stored on are LONG gone)  it was a similar story with delays averaging out at 5-10 minutes for most trains.  Those days it was all loco-hauled, HSTs didn’t arrive on the route until just after (almost all 47s with the occasional 31 subbing in) and the trains only ran once an hour, rather than twice an hour, though were usually 7 or 8 coaches long.

Agreed, it wasn't a perfect world in BR days either.  But......I commuted weekly (outbound on Monday, return on Friday) from Plymouth to Paddington between 1995 and 2002. The trains I used were nearly always on time at destination.  In the first year, my commute was extended to Ashford (Kent), and I cannot recall ever missing my connections in London and being late for work or getting home late (other than in the case of the serious incidents that occured during that time period).

Now then, analyising the current data a bit more shows the trains in the tables in the posts above being significantly less late at Reading, so most of the serious lost time seems to be between Reading and Paddington.  Indeed, deeper analysis shows significant loss of time between Hayes and Paddington.  Watching the trains referenced on the RTT Maps shows significant congestion and serious regulating issues on the immediate approach to Paddington, and I think I have mentioned before I blame ARS (Automatic Route Setting) for that.

In the BR pre-ARS days, signallers could make their own regulating decisions to minimise overall delay, but if you watch ARS in action you will see conflicts being created because the programming assumes all trains are running to schedule and they can only utilise their booked route regardless of lateness.  I find that very frustrating considering the numerous alternative routes available, but never used, in the Ladbroke Grove to Paddington signalling layout.  The classic case is all Up Main trains being routed in via Line 3 despite the fact their destination may be the lower numbered platforms, causing conflict in the station throat with trains departing the higher numbered platforms.

Apologies for the whinge again about ARS.  It really bugs me.  I mentioned the same thing in post #14 above..... ::)

I can remember the predecessors to XTC regularly arriving at Reading from Southampton/Bournemouth two or more hours late. For the Newcastle services I would imagine they needed to hand out Red Cross somewhere north of Macclesfield.

Usually the 47/8s failing or other 47' s running nut of fuel, pre the Voyagers, One memorable night an XC arrived with the Reading 08 hauling it the 47 having failed at Tilehurst.

The tone of this thread does suggest reliability is more important than regular lateness so long as it's not excessive.

I think that's probably fair comment - though a lot of things have changed and much comment in the thread is historic.

Changes ...

1. Services have moved 'up' in frequency - I recall 2 Liverpool to Poole trains, and one York to Poole per day, where there are now 2 trains per hour on the core centre of the route, and hourly towards the extremes

2. Customer's metrics have changed from occasional long distance journeys to more frequent but much shorter journeys - "even the new class 800 are providing for predominantly commuter customer base" was a discussion I had this morning.

3. Information systems now report to us (expose) performance in a way that never happened back in the last century.

I have a theory about the frequency at which a service needs to run to encourage most of the passengers available on that flow to use it - and that is that the frequency of a journey being offered should be no less that the time taken on the journey.    So a London to Swindon service - journey time one hour - wouldn't actually loose too much traffic if it was less frequent that an the moment, whereas Ashchurch to Cheltenham Spa services would gain passengers by leaps and bounds if they were improved.   Of course it's not as simple as that, as the London - Swindon trains fan out from Swindon, each line needs its frequency, and there are capacity issues too.  And you will not justify 4 trains an hour from Ashchurch into Cheltenham Spa - the service is sustained and viable because of all the other passengers making longer journeys, where such an intense service isn't necessary to retain the traffic.

Why bring up the theory here?    I suspect that something similar applies to delays.  Provided that connections are not missed, I suggest that people might in general not be too unhappy arriving 5% of their journey time late.   So London to Swindon - 3 minutes late not really a problem.   Penzance to Paddington - quarter of an hour late is no more that the shrug of the shoulders. But of course if that train had picked up at Newbury, an hour out, the passengers from there would be pretty unhappy even though it's the same 15 minutes.

So - I suspect that human nature has not changed, but with a much higher proportion of shorter journeys now, on average late arrivals are far less acceptable than they were 20 years ago.   And information systems trumpet the stats and failures - to a very great extent we don't help on a forum like this when we highlight performance failures that would previously have not made it into people's knowledge zones.

Grahame, I first read that as you thinking it doesn't help the 'cause' if we highlight these failures?  I'm sure that was not the intent of what you actually said.

To me its very important that we have a reliable and efficient service, something at present that we don't have, even by a close call.  There is no point in advertising services to the nearest 1/2 minute if you can't even meet the plan by 10 minutes.

To me, and yes I do admit to having had a reasonably good insider view, is that NR/GWR have some over capacity and infrastructure underuse issues that they need to address.  What's the point of having all that flexible signalling layout at Paddington if you never use it to its full capacity?  I agree that certain lines do not need the service levels that are offered at present and that certain lines do. I am really concerned about what is going to happen when Crossrail starts in 2019 because the railway can't cope even now.....

I'll get off my 'high-horse' now and stick to only posting worthwhile information.

I have a theory about the frequency at which a service needs to run to encourage most of the passengers available on that flow to use it - and that is that the frequency of a journey being offered should be no less that the time taken on the journey.   

Grahame I'm not quite sure what exactly this means Are you saying a 10 minute journey should have a train every 10 minutes or 6 tph, whereas a 2 hour journey needs a frequency of every 2 hours. Also are you talking abut overall train journey time or individual passenger journey time?

It sort of works for Slough to Reading (18 miles) stopping services 4 trains an hour journey time on average 25 minutes. But Slough to Taplow 2tph journey time 7 minutes.

I think that the frequency of a journey being no less than the time taken for the journey, is a reasonable start provided of course that a bit of common sense is applied regarding very short and very long journeys.

I see no need for a train every 8 minutes for a journey that takes 8 minutes, and I also think that Penzance needs a more frequent service than once every 5 or 6 hours. But for many journeys it would be a reasonable starting point.

Of course many busy services need more frequent trains for reasons of capacity. Paddington to Reading needs more than a train every 20 minutes, simply to shift the numbers.

A more complex way of putting this might be

"The interval between trains should be no longer than the time taken on the train, subject to two exceptions.
Firstly the interval between trains should not, in most cases, exceed 2 hours even when journey time does exceed 2 hours.
Secondly, no service need run more often than 4 times an hour unless a more frequent service is required to avoid overcrowding."

A couple of clarifications.

Firstly, I wasn't intending to suggest that I or other members stop bringing systemic and unusual incidents up, which is basically what we do.  But I was commenting that there's an awful low more noise about than there used to be - a feature of our information age - so passengers who would be blissfully unaware of problems in the region but not on their line will be much more alerted these days and could be put off rail and / or worried needlessly when 9 out of 10 journeys they take will be fine.  Balance thing.

In terms of the service frequency, broad brush you folks have read me right.   But whilst a train from PNZ taking 5 hours to PAD may for that flow loose little traffic even if it only runs 4 times a day, it would loose so much intermediate traffic such as PNZ to PLY, PLY to EXD, TAU to WSB, PEW to PAD, etc ... so my suggestion of flow modelling is about journeys between two places and does not easily map onto a train calling at many.

Looking back, perhaps there's tacit acknowledgement of an algorithm along these lines for the past 50 years - an occasional crack express service and lots of trains calling at a bevy of intermediate stations.  The coming of the motor car to everyone in the 1950s and 1960s was the death knell of many of the local services, a prior to that the algorithm really didn't apply because people had no choice.  So either services blossomed - look at how Pewsey or Warminster, Highbridge or Crediton have grown ... or they withered and got closed - Camelford or Portishead, Radstock or Langport, Corsham or Ashburton.