If you register, you can do a lot more. And become an active part of our growing community. You'll have access to hidden forums, and enjoy the ability of replying and starting conversations.

Edward Thompson: Wartime C.M.E. Discussion

Discussion in 'Steam Traction' started by S.A.C. Martin, May 2, 2012.

  1. Selsig

    Selsig Member

    Joined:
    May 6, 2007
    Messages:
    861
    Likes Received:
    365
    Location:
    Coventry, United Kingdom, United Kingdom
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Wouldn't there be an argument against the records data being falsified *just* on Thompson locos, whilst being totally legit on Gresley locos. A lot of Simon's work hasn't just been on the absolute statistics, but also on the comparative ones as well, which is surely a benchmark that should smooth out any inherent bias of individuals.

    John
     
    jnc, MellishR, Fred Kerr and 3 others like this.
  2. Jamessquared

    Jamessquared Nat Pres stalwart

    Joined:
    Mar 8, 2008
    Messages:
    26,100
    Likes Received:
    57,414
    Location:
    LBSC 215
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Sorry - I think that is fanciful. The engine record card represents the summation of hundreds, or perhaps thousands, of man-hours of work on each locomotive. Locomotive repair was one of the major costs faced by railways; as such the company boards paid close attention to the costs; and the workshops themselves had accountancy and audit staff attached to ensure that as far as possible costs were correctly attributed, and represented the work actually done.

    So could an individual card be wrong? Yes, possibly. But could there be a consistent error in one direction - and even worse, an error that penalised the locos of one designer while moving in the other way for those of another? Not a chance, it would be impossible to maintain such a subterfuge as a deliberate act, and even an unintentional error would reveal itself soon enough.

    As for driver's deliberately failing a loco unnecessarily: I'm sure it happened from time to time, though I think that underestimates a degree of pride from many in getting the job done regardless of the condition of the tools they were given. But any driver consistently doing so would rapidly become unpopular: firstly with his mates (by adding extra pressure to the shed staff); and the shed master and shed foreman would rapidly spot such patterns. Remember as well that for many years, employment was precarious so few people would wish to be identified as troublemakers. My sense is that if there was a consistent error in any direction, it was probably locos not being failed when they should have been, rather than the other way round.

    What is frustrating from a historical point of view is that there is relatively little written from the first hand about the systems in use - KJ Cook being an honourable exception. But the records card themselves I just can't see being falsified in a widespread, and consistent, basis.

    Tom
     
    Romsey, jnc, Cosmo Bonsor and 6 others like this.
  3. S.A.C. Martin

    S.A.C. Martin Part of the furniture

    Joined:
    Aug 31, 2010
    Messages:
    5,591
    Likes Received:
    9,325
    Gender:
    Male
    Occupation:
    Asset Engineer (Signalling), MNLPS Treasurer
    Location:
    London
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    Absolutely. The availability statistics which we discussed in this very thread, and I’ve made available at times in the form of my spreadsheet, give the entire data for the whole of the LNERs locomotive stock for 1942-1946, not just the Thompson Pacifics.

    The amazing thing is that the engine record cards and the availability data do in fact validate each other in terms of years and mileages. So happily, we can be reasonably satisfied that we’ve got accurate data derived from good record keeping, given they’re recorded in different ways by different people across the company.

    I’ve since managed to find and start preparing data for years before and after that specific time frame.

    I’ve also had the data validated a few times so that we can all be sure of the accuracy. I’m not infallible after all.
     
    Kje7812 and MellishR like this.
  4. Big Al

    Big Al Nat Pres stalwart Staff Member Moderator

    Joined:
    May 30, 2009
    Messages:
    21,064
    Likes Received:
    20,773
    Location:
    1016
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Yes it did. Slight digression but on the Southern, a London bound service was once failed at Waterloo for no good reason so that a spare (poorly steaming, as it turned out) loco had to be used for the return working - the Belle. It was a 'driver jealousy' issue at the time.

    And whilst the "Gresley good, Thompson bad" view of the world is probably too much of a generalisation, you do begin to wonder whether some authors at the time slipped too much into celebration mode with their books rather than taking a hard nosed and balanced look at the evidence as it seems that @S.A.C. Martin has been doing.
     
    ragl, MellishR, 35B and 1 other person like this.
  5. Monkey Magic

    Monkey Magic Part of the furniture

    Joined:
    Jan 6, 2018
    Messages:
    3,498
    Likes Received:
    6,845
    Location:
    Here, there, everywhere
    Heritage Railway Volunteer:
    No I do not currently volunteer
    I am not saying that data is systematically falsified, in fact I am arguing the reverse. By using large datasets the noise of one individual ie the driver who thought they would 'slip on Musselburgh sands' can be ironed out.

    The Barker quote is about over-reading a single source (or set of sources) and taking for example a comment scrawled in a hurry as reflecting wider views without situating it in everything else. ie not taking one porter's reminisces that one signalman went to the pub as evidence of widespread drinking on the job, or for example, taking a single report from one driver as evidence of wider performance.
     
    Last edited: Apr 25, 2021
    Bluenosejohn, 35B and S.A.C. Martin like this.
  6. S.A.C. Martin

    S.A.C. Martin Part of the furniture

    Joined:
    Aug 31, 2010
    Messages:
    5,591
    Likes Received:
    9,325
    Gender:
    Male
    Occupation:
    Asset Engineer (Signalling), MNLPS Treasurer
    Location:
    London
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    I have given myself some time to consider whether I seriously respond to Julians post here, but on reflection and for clarification it would be sensible to at least point out a few flaws in Julian’s claims.

    The statistics we have from Kew Gardens are part of the LNERs full archives there. They form one part of a wider body of work that also includes:

    • The board minutes
    • The locomotive committee minutes
    • Letters from Gresley, Thompson, Peppercorn, and more, and between them
    • Works reports
    • Statistical data
    • Accident reports
    • Design reports and designs from the drawing offices
    In looking at the war years I’ve gone back to the primary evidence as much as possible, including reports from the contemporary railway press.

    I’ve also bought and paid for a huge amount of archive stuff to be scanned for better examination, together with buying up old LNER pamphlets, some contemporary books, and more.

    The bibliography I have put together runs to about six A4 sheets at the moment. Much of the material I’ve read I’ve had pointers from others to read - Andrew Hardy’s work on the P2s pointed me in directions I needed to go for the views on the original and rebuilt machines.

    The fact of the matter is that the views of the Thompson Pacifics as built, and in service later, are nowhere - and I really do mean “nowhere” - near as bad as that we’ve heard about them after their scrapping. In fact many positive things were said at the time of their introductions, and the reality is that I have the physical evidence to show that.

    My work has the potential to be seriously embarrassing to a number of authors who have gone before me on the basis that their views do not match the evidence available to all of us. I’m working from primary evidence, not secondary or viewpoints given years after the events that we are discussing now.

    That’s important because it’s what we should have all been working from in the first place.

    So I refute entirely your suggestion that the availability statistics go against any other contemporary material - in fact the stats are validated by other sources contemporary to them.

    All steam locomotives had issues. The question here is whether the reported issues are factually accurate. If you - were - being objective Julian, you’d have my point of view based on the primary/contemporary evidence: which is that the Thompson Pacifics in particular had some issues - but not to the extent of that reported, at all.

    Besides which - we are not just talking about Thompson’s Pacifics. The whole point of this thread has to be to open up a debate about him and his work and, happily, we’re having a serious debate that has developed over many years where those interested in the nuances of the research ask pertinent questions instead of sticking their heads in the sand: just because it doesn’t fit with their long held views (which surprise surprise might be entirely wrong).

    That is because you are stubborn, Julian, as opposed to having a counter argument.

    “One” may wish to actually post something of relevance if “one” would like to be taken seriously in the future.

    You pop up every so often, as do a couple of others, looking to undermine my research but rather spectacularly failing to do so by not forming a coherent argument or citing any actual evidence that counters it.

    Happily I’m showing the data and the supporting evidence together in the book so people can make their own minds up.

    “Partial picture” - given the way in which Thompson and his work has been reported in before - what a laugh!
     
  7. 35B

    35B Nat Pres stalwart

    Joined:
    Jun 18, 2011
    Messages:
    25,485
    Likes Received:
    23,718
    Gender:
    Male
    Location:
    Grantham
    Heritage Railway Volunteer:
    No I do not currently volunteer
    I read that comment the other way round - that the driver in this situation would make a fuss about the loco which would, importantly, go into the reputation of the loco and/or class at that shed. By screening out individual occurrences of misreporting, statistical analysis would show the overall performance by removing the “noise” of individual occurrences.

    I’ve heard enough train crew complain about writing up faults in repair books and getting “no fault found” back to be sure that no system was perfect. Likewise, I well recall Fiennes writing of the aftermath of one fatal incident that the influence of the driver (a senior man at New England) made identifying the cause unnecessarily hard as others supported his account of iirc a signal sighting issue.


    Sent from my iPad using Tapatalk
     
    ragl, S.A.C. Martin and Monkey Magic like this.
  8. Monkey Magic

    Monkey Magic Part of the furniture

    Joined:
    Jan 6, 2018
    Messages:
    3,498
    Likes Received:
    6,845
    Location:
    Here, there, everywhere
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Exactly, the other point is that if you noticed a bunch of outliers ie very low/high availability, you'd look for some common variable that might explain it. Are they all at the same shed - so something about that shed, or from a batch that had a design feature that wasn't replicated in other class members.

    By having the data across the years you can also see changes ie did availability at shed x improve/decline with the replacement of the shedmaster, did it improve/decline in response to modifications or things wearing out (in response to high usage?).

    You can start to control for things: ie unfamiliarity when new, type of work demanded, route demands, staff turnover, coal supply, water supply - anything you think that might skew the results or explain why you might have a bunch of outliers. Cross reference with availability for other classes (is it a class or geography cause - ie are all locos at this shed outliers, is this class an outlier when new/old compared to other classes). ie @Jamessquared point about the Britannias at Cardiff Canton.

    The advantage of large n data like this is that you can get the macro view, iron out potential individual biases, but also identify outliers for deeper investigation.

    In other words, low availability in the early days might well boil down to one shed (ie Canton), unpack it further and there is something (in the case of the Britannias the switch from RHD to LHD and change in firebox). Conclusion, no evidence that Britannias were 'bad' locomotives but that they took time to bed in because of the changes in the way people worked. Or you note that low availability is based around six shed but that shed X had very demanding turns, or that shed Y and Z had union problems or that sheds A,B,C had very high staff turnover, so on and so forth and that there is no mono-causal explanation for low availability in our outliers.

    Even with comparison you'd need to control for things. If you are comparing synchronically ie locos in 1955 you'd be comparing a 9 year old loco with a 20 year old loco. If you were comparing diachronically between classes at say 10 years you'd need to control for the fact that 10 years was 1945 (with issues relating to maintenance/availability) vs 1955 (with different issues at play). Likewise relatively small class sizes vs larger class sizes.
     
    Last edited: Apr 25, 2021
  9. MellishR

    MellishR Resident of Nat Pres Friend

    Joined:
    Apr 16, 2009
    Messages:
    8,068
    Likes Received:
    5,164
    Do I see some troll-feeding going on here?
     
    paullad1984 likes this.
  10. jma1009

    jma1009 Well-Known Member

    Joined:
    Mar 16, 2013
    Messages:
    1,381
    Likes Received:
    1,637
    Gender:
    Male
    Location:
    ynysddu south wales
    I dispute the accusation of 'trolling'.

    What I had in mind was Gresley pacifics during WW2 being badly maintained.

    When we come to the Thompson rebuilds of the Greeley Pacifics, they would obviously initially have greater availability and mileages than their worn out forebears?

    The account of Don Young of 'Mallard' in the 1950s being 'knackered' when at Doncaster for a heavy overhaul is illuminating.

    What backlog of heavy overhauls was there during the period under consideration of older Gresley pacifics, that might have given at the time the Thompson rebuilds an advantage of availability and mileage?

    I don't and didn't suggest that the Kew records were false, merely that other factors might affect how they could be interpreted.
     
    Bluenosejohn likes this.
  11. S.A.C. Martin

    S.A.C. Martin Part of the furniture

    Joined:
    Aug 31, 2010
    Messages:
    5,591
    Likes Received:
    9,325
    Gender:
    Male
    Occupation:
    Asset Engineer (Signalling), MNLPS Treasurer
    Location:
    London
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    This in fairness is an entirely sensible query. The answer is not absolutely clear.

    There was such an issue of backlogs at the major works that Thompson put in place some special measures - like additional boiler inspectors - that would examine locomotives and decide on whether or not they could be returned to service without a visit to works. Availability had slumped across the whole of the LNER, not just Gresley locos (though the fact they were the newest classes, with lower availability than some double or more their age, was a huge concern for Gresley, then Thompson, and the board).

    What you do need to bear in mind that we saw a Mikado replaced with a Pacific, and it was the Pacific which was superior by nearly double the availability and mileage.

    Thank you for the clarification Julian, and my apologies for the brusqueness of my original response, which was unnecessary given what you were aiming at.
     
    ragl and jnc like this.
  12. Monkey Magic

    Monkey Magic Part of the furniture

    Joined:
    Jan 6, 2018
    Messages:
    3,498
    Likes Received:
    6,845
    Location:
    Here, there, everywhere
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Apologies in advance for a boring post. :)

    Before heading into the long weeds of methodological approaches, I would just say that I don't think that any one approach is better than another. It depends on the data and what the question is. You can use the wrong approach to analyse data, and you can use the wrong data to answer a question. But I don't think this has happened here.

    I can't speak for Simon, but the way in which I would approach it would be this. Compare synchronically (ie both classes as the same point in time ie 1945-1960), then I'd compare a second time diachronically at same age ie A3 in year 1 vs A1 in year 1. I'd aim to control for known factors that would impact on loco availability ie war situation, strikes, new locos, old locos, sheds, changes in policy. I'd also control for class size as clearly, with a smaller class you can get much bigger variations if say 1 loco has an accident and that is 5% of your class out of action.

    This would give you a lot of different data but if aggregated it pointed in the same direction then I think you would have a stronger argument. You will have also countered the 'but what about x' argument.

    However, you can't have more variables than cases. And as long as you are not selecting off the dependent variable then you are ok. I would also argue that this is appropriate methodology for the data available.

    I take on board the point that adding more variables allows us no closer to understanding the causal mechanisms and merely prolongs the agony. If I were critical of this methodological approach it is that it does not allow us into the 'black box' of causality.

    However, this is important information to have even if it doesn't explain the why. Redressing or critically assessing with an evidence based approach is of vital importance.

    Here's the thing - what Simon is doing is no less valid than if someone went and did an oral history of Top Shed or Haymarket and managed to get a 150 accounts from people who worked there (unlikely now because of age). But let me give you an example of poor oral history methodology - with oral history how you ask questions is vital. So if you went and asked the question 'How bad were the Thompson pacifics?' you would get a negative response, if you asked 'tell me about the Thompson pacifics?' or 'tell me about the Gresley, Thompson and Peppercorn pacifics?' you'd get different answers. (the first invites negativity, the third an explicit invite to compare, the second just about the class). The problems with this approach are obviously the evidence is subjective, small samples and the difficulty in scaling up to the macro. As any anthropologist will tell you, it tells you just about the here and now and no where else.

    At the risk of getting post-modern, I don't think there is an objective truth to be found here. Oral history might give us a different answer than Simon's statistical analysis but both can be right at the same time, because they are asking different questions about the same thing and getting different answers. Simon tells us about availability, oral history tells us what some people remember. As long as the data is good, the methodology is sound and the analysis strong and supported by the available evidence then both can be solid answers.

    Of course, if you were bold and this is probably for someone to else to do later down the line is to synthesize the macro (availability data) and micro (individual responses). You could do this with something like ecological inference (you need very good statistical data for this) - ie how likely is a Thompson pacific based at Top Shed likely to have ride complaints. (When you read a story that says 'Statistics show Trump voters not all poor and rural' this will have most likely used ecological inference). Alternatively, and I think Simon is correct in not throwing out the memories of crew, shed staff, etc, how to synthesise this would be to adopt nearer to process tracing and to use what is known as Coleman's boat.

    (Now to head into the really long weeds)

    [​IMG]


    So Simon's statistical analysis shows that where a loco is shedded impacts on availability - macro cause - macro effect.

    but why?

    Situational mechanisms - by doing a deep dive on a sheds we can observe that alternative employment, poor industrial relations matter

    Micro level event
    - unhappy shed/new shedmaster is a nasty piece of work.

    action formation mechanisms - by looking at accounts from people working in the sheds under question we can see that when the shedmaster was a difficult person people to work under and because there was an alternative job they decided to leave and go and work in another industry. This is where we bring in all the stuff written or talked about by people.

    Micro level event - people leave - high staff turnover, loss of skilled staff

    Transformational Mechanism - poor maintenance/maintenance backlog

    Macro effect - lower availability

    Hence we get the best of all worlds, macro level and micro level. Problem - it is time consuming and difficult to do, so it is easier to do a smaller hit/task. ie demonstrate the macro cause and effect, or to unpack a micro level process and to level the macro implications to someone else. And of course it all depends on the evidence.
     
    Last edited: Apr 27, 2021
    Spinner, 30854, Mandator and 3 others like this.
  13. Eightpot

    Eightpot Resident of Nat Pres Friend

    Joined:
    Aug 10, 2006
    Messages:
    8,077
    Likes Received:
    2,264
    Gender:
    Male
    Occupation:
    Engineer Emeritus
    Location:
    Aylesbury
    Heritage Railway Volunteer:
    Yes I am an active volunteer
    I presume you are referring to the rear-end collision by 20-mile bridge at Welwyn Garden City involving 60520 'Owen Tudor'? Seems the story got confused in its telling to encompass not only other signals on the Up Main, but ones on the Down Main as well!
     
  14. 69530

    69530 New Member

    Joined:
    Nov 20, 2013
    Messages:
    60
    Likes Received:
    75
    Gender:
    Male
    I am pretty sure it was the 83D Laira men that disliked the Brits, bad view forward, draughty dusty cabs, difficult to fire, etc. Eventually they were congregated at 86C Canton and gained acceptance from the crews there, who got some good results on the Paddington trains.
    Strangely why would you allocate Brits to GW sheds when the castles were quite adequate, and Two Brits to 73A Stewarts lane when they had both WC's and MN's allocated there.
     
  15. clinker

    clinker Member

    Joined:
    Oct 8, 2016
    Messages:
    569
    Likes Received:
    348
    Gender:
    Male
    Location:
    romford
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Just holding a candle to the Devil, butI'm sure that Swindon's record sheets would have been meticulously kept, particularly around the time that Rood Ashton Hall was scrapped.
     
    johnofwessex likes this.
  16. Bikermike

    Bikermike Well-Known Member

    Joined:
    Mar 11, 2020
    Messages:
    1,442
    Likes Received:
    1,595
    Gender:
    Male
    Location:
    Thameslink territory
    Heritage Railway Volunteer:
    No I do not currently volunteer
    I presume you have read "Steam in the Blood" by RHN Hardy? (I was just reading it last night, and he has a bit on ET (who hired him, and went to tea with his mother) at the beginning.
     
  17. Jimc

    Jimc Part of the furniture

    Joined:
    Sep 8, 2005
    Messages:
    4,052
    Likes Received:
    4,665
    Occupation:
    Once computers, now part time writer I suppose.
    Location:
    SE England
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Its well established that the WR stopped keeping a lot of records in the last couple of years of steam. Tender allocations for example.
    But 4965/4983 is not too different from 3265/3365 back in 1929. If you leave out the enthusiast theory that loco identity rests with frames I wonder if there were any records not kept that should have been?
     
  18. clinker

    clinker Member

    Joined:
    Oct 8, 2016
    Messages:
    569
    Likes Received:
    348
    Gender:
    Male
    Location:
    romford
    Heritage Railway Volunteer:
    No I do not currently volunteer

    By Whom?
     
  19. Jimc

    Jimc Part of the furniture

    Joined:
    Sep 8, 2005
    Messages:
    4,052
    Likes Received:
    4,665
    Occupation:
    Once computers, now part time writer I suppose.
    Location:
    SE England
    Heritage Railway Volunteer:
    No I do not currently volunteer
    eg RCTS part 12, pM10 "a few [Dean era 3,000 gallon tenders] were still in stock when records ended in September 1964"
     
  20. Mandator

    Mandator Part of the furniture

    Joined:
    Feb 12, 2015
    Messages:
    4,868
    Likes Received:
    3,043
    Gender:
    Male
    Heritage Railway Volunteer:
    No I do not currently volunteer
    Not a boring post at all! In fact extremely informative and illuminating.
    A thank you from me.
     
    Spinner and S.A.C. Martin like this.

Share This Page