England and Wales High Court (Commercial Court) Decisions

IN THE HIGH COURT OF JUSTICE
QUEENS BENCH DIVISION
COMMERCIAL COURT

		Royal Courts of Justice Strand, London, WC2A 2LL
		15 July 2003

Mr Justice Gross:

INTRODUCTION

1. The Claimant ("GNER") is and was at all material times the operator of the InterCity East Coast rail service, running services to and from London King's Cross and Edinburgh.
2. The Defendant ("Railcare") is and was at all material times a company in the business of the overhaul and/or reconditioning of, inter alia, parts used by the railway industry.
3. In or about May 1997, Railcare overhauled and balanced a wheelset (no. 9549, "the wheelset"); as part of this work, Railcare drilled a hole in the web of the wheel through which a balance weight was bolted on. The wheelset came to be fitted on a Mark IV coach (no. 11224) on the 3^rd June, 1997. That rolling stock was owned by Forward Trust Rail Ltd. (an HSBC subsidiary, "Forward Trust") and was leased to GNER. In the event, it came to be used on the InterCity East Coast rail service.
4. It may be noted that there was no contractual relationship between GNER and Railcare. The details do not matter but I have been told that Railcare had a contract with an entity called Railpart Ltd. to supply rail parts and it was Railpart Ltd. who supplied the wheelset to Forward Trust/GNER.
5. As admitted in the course of these proceedings, after drilling the balance weight hole in the web of the wheel, Railcare negligently failed to deburr and chamfer the edges, leaving a burr which was a stress raiser and which, in turn, resulted in the initiation of a fatigue crack.
6. Over time, the fatigue crack propagated to the rim of the wheel, causing it to break up while in service on the 16^th June, 1998. The result was the derailment of the 17.30 passenger service from King's Cross to Edinburgh, while travelling at 125 mph ("the derailment"). Mercifully, there were no fatalities or serious injuries.
7. Railcare admits that its negligence in failing to debur and chamfer the wheelset caused the derailment.
8. In these proceedings, GNER claims damages from Railcare in tort, arising from the derailment. As recorded in the Order of David Steel,J., dated 18^th January, 2002 and as already foreshadowed, liability was admitted. Accordingly, this trial has been concerned with issues of quantum, including causation, remoteness and contributory negligence.
9. The principal Issues were as follows:

i) The ascertainment of GNER's lost passenger and other revenue attributable to the derailment ("Issue (I): Lost Revenue");

ii) The ascertainment of the expenses incurred by GNER attributable to Railcare's negligence ("Issue (II): Expenses");

iii) The question of whether GNER was negligent or contributorily negligent in failing to detect the crack prior to derailment; if so, what loss and damage resulted from such negligence or contributory negligence; if GNER was contributorily negligent, what apportionment should be made ? ("Issue (III): Contributory Negligence").

Pausing here, Issues (I) and (II) are somewhat differently formulated because Issue (II), as will become apparent, requires a more precise focus on Railcare's negligence. Likewise, again for reasons which will become apparent, although Issue (III) includes a question of whether GNER was negligent (as distinct from contributorily negligent), it is convenient to give this Issue the title of Contributory Negligence; realistically, it is contributory negligence with which Issue (III) is concerned.

10. On the first day of the trial, I heard argument on the question of whether Railcare was (in short) entitled to introduce a case that much of GNER's (alleged) loss should in any event be disallowed on the ground that it constituted irrecoverable economic loss. On the second day of the trial, I ruled that Railcare was not entitled to do so; the reasons for that ruling are set out in the Transcript for the 26^th March, 2003 and need not be repeated here.
11. Oral evidence was given from the following witnesses of fact, all called by GNER: Messrs. Parsons, Cooke, Bond, Lumb, Buxton, Tham and Montgomery. I shall come to their evidence in due course.
12. No fewer than 10 experts had been retained by the parties, five apiece. These experts and their respective fields were as follows:

i) Metallurgists and engineering experts, messrs. Rasaiah and Warder, for GNER, Baker and Cooper for Railcare. In the event, it proved unnecessary to call messrs. Rasaiah and Baker, although their reports were in evidence.

ii) Safety and maintenance experts, Mr. Ward for GNER and Mr. Etwell for Railcare. As will be seen, Mr. Etwell's remit was somewhat wider than this description indicates.

iii) Railway consultants, Mr. Segal for GNER and Dr. Harris for Railcare.

iv) Forensic accountants, Mr. Andrews for GNER and Ms. Winspeare for Railcare.

13. Unavoidably, given the nature of the case, the relationship between the experts and the principal Issues does not lend itself to a tidy summary. In very broad terms, the metallurgists and engineers, together with the safety and maintenance experts, were concerned (at least primarily) with contributory negligence (Issue (III)). As will be seen, the railway consultants to some extent likewise were involved at least with the ramifications of contributory negligence, if established. Matters of revenue and expenses (Issues (I) and (II)) were dealt with by the forensic accountants and the railway consultants.
14. As indicated during the course of the trial, I was troubled by certain aspects of the expert evidence:

i) Underlying my concern was the consideration that, in the end, damages are a question of fact, traditionally a question proper for a jury; the law gives guidance by way of general principles but these are to be applied so as to do substantial justice between the parties: see British Westinghouse Co. v Underground Rys. [1912] AC 673, at pp. 688-689; Monarch Steamship v Karlshamns [1949] AC 196, at p.232; McGregor on Damages (16^th ed.), at paras. 2065-2067. Against this background, the court must do the best it can to compensate the injured party, bearing in mind where necessary any applicable burden of proof. Inevitably, there will be a rough and ready element in the award of damages and a search for precision is likely to prove illusory. Viewed in this light, it did strike me that, faced with a proliferation of experts, it was of the first importance to maintain perspective, keeping well in mind the essentially factual or "jury question" nature of the inquiry. Otherwise there was a danger that the trial could be skewed rather than assisted by the expert witnesses.

ii) The Court is always anxious to benefit from new or developing disciplines but difficulties can arise where there is at least as much "art" as "science" in the field in question. In the present case, the railway consultants, Mr. Segal and Dr. Harris, were enthusiastic, knowledgeable in their field and anxious to assist. The problem lies in converting their informed but somewhat generalised thinking into sufficiently concrete conclusions to prove of assistance in a trial such as this. As it seemed to me, their evidence was ultimately only of real assistance (1) in providing a cross-check on other views and (2) in alerting the Court to relevant factors to be taken into consideration.

iii) A problem of a different nature arose in connection with Mr. Etwell's evidence. He was plainly a knowledgeable and experienced witness, with a broad background in the railway industry. Put neutrally, his evidence principally concerned (1) the prospects for detecting the fatigue crack prior to derailment and (2) the question of whether, had the crack been detected in routine maintenance, GNER might have adopted a different course from that followed subsequent to the derailment – by its nature, a somewhat ambitious exercise - and, if so, what savings would likely have been made. It will be necessary in due course to consider whether Mr. Etwell's evidence (on both (1) and (2)) went far enough to lend support to Railcare's allegation of contributory negligence on the part of GNER. The not unfamiliar problem of determining whether an expert was saying no more than that things might have turned out differently or was in truth lending his support to a charge of (contributory) negligence, arose in acute form here.

iv) With regard to Mr. Ward, I regret that his evidence contained too many changes of view, insufficiently reasoned, to permit me to place reliance on at least much of it.

Subject to the above observations, the experts assisted the court within the limits of their respective disciplines and, helpfully, aided the parties in narrowing the ambit of the dispute.

15. I come next to the principal Issues.

ISSUE (I): LOST REVENUE

16. It was common ground that a railway operator's year is divided into periods ("Ps") of 4 weeks. The derailment took place in P6, week 3 of 1998. There was no, or no real dispute that the derailment led to disruption in Ps 6-8. Although GNER suggested that the disruption continued into P9, for various reasons (the details do not matter), the claim was limited to lost revenue in Ps 6-8. The claim consists largely of lost passenger revenue but, before leaving this topic, it will be necessary to reach a conclusion on the relatively small item of lost catering revenue. The nature of the dispute under this heading appears from an outline of the rival cases, to which I now turn.
17. The case for GNER: For GNER, Mr. Schaff QC submitted that the lost passenger revenue attributable to the derailment was to be calculated by subtracting actual revenue for Ps 6-8 from the best estimate of revenue for that period, making such allowance (or deductions) as appropriate for any other factors (not related to the derailment) which might also have contributed to the shortfall in revenue.
18. At once the need arises to establish the estimated revenue for Ps 6-8. There were two candidates; the first was GNER's budget for 1998; the second was an updated forecast, prepared by GNER, on the evidence, in P5, late in May. The question was, which provided the better estimate of likely revenue for Ps 6-8 ? The difference between the two was substantial. As agreed by the accountancy experts (Mr. Andrews and Ms. Winspeare), use of the budget yields a lost passenger revenue claim of £1,865,000 (before any allowances) whereas use of the forecast would result in a claim of £3,400,000.
19. Unsurprisingly, perhaps, GNER contended that the forecast produced the better estimate. It was more up to date. Further, the basis of the forecast was that GNER was performing some 1.8% ahead of budget; this was a justified assumption, given that in Ps 4 and 5, GNER had performed ahead of budget. Moreover, Weekly Revenue Reports ("WRR") for weeks 1 and 2 of P6, predicted that GNER would be 2.3% ahead of budget at the end of P6. Accordingly, the forecast was more likely to be right and should be chosen instead of the budget.
20. If, contrary to its primary case, the forecast did not prevail or not prevail throughout, then GNER was not in any event limited to the figures yielded by a claim based on the budget. That, submitted Mr. Schaff, would under-estimate the claim. Instead, the forecast should be relied on for P6 (bearing in mind that there was no separate claim for weeks 1 and 2, which ante-dated the derailment) leaving the claim for Ps 7 and 8 to be assessed in accordance with the budget. In practical terms, adopting such an approach would yield a claim in the order of £2.5 million. Alternatively, at the least, the budget figures for weeks 1 and 2 of P6 should be uplifted, in an amount of approximately £200,000 or, at the least, not less than £150,000.
21. As to allowances for extraneous matters, unrelated to the derailment, Mr. Schaff accepted that a figure of £32,000 was to be taken into account, in respect of a London Underground strike, which might have contributed to a fall-off in traffic in the relevant period. That apart, the extraneous factors were not quantifiable and/or had been taken into account in the budgeting and forecasting process. In particular, Mr. Schaff urged that if the claim was limited to the budget figures or to the budget plus an uplift, then it would be wrong to make any mechanical, further deduction for extraneous factors. It was a matter of swings and roundabouts and it was important not to penalise GNER unfairly.
22. The case for Railcare: For Railcare, Mr. Popplewell QC's primary case was that lost passenger revenue should be determined in accordance with the approach suggested by Dr. Harris: first, average daily passenger revenue is ascertained for the whole of the year for periods unaffected by the derailment, namely, Ps 1-5 and 9-13; secondly, that figure is then "unsmoothed" to reflect different travelling patterns for the days of the week; thirdly, that figure is further "unsmoothed", to reflect the different revenue derived from different ticket types; judgment is then applied to determine how much revenue is lost by ticket type day by day. This approach was to be applied from the 16^th June until the following Sunday, the 21^st June; thereafter, to the figure thus obtained, a long term factor of 1.6 to 1.7 was to be applied, to reflect the longer term effect. The application of this methodology would result in a lost passenger revenue claim, prior to allowances, of some £1.6 - £1.7 million. This approach could not fairly be criticised as theoretical; it was employed in the Passenger Demand Forecasting Handbook ("PDFH").
23. If, contrary to his primary approach, it was considered appropriate to assess lost passenger revenue by way of a comparison between estimated and actual revenue, less allowances, then Mr. Popplewell submitted that the forecast could not be relied upon and the GNER 1998 budget should be used instead. There were two ways of testing the reliability of a forecast: (1) by analysis of its methodology; (2) by a hindsight consideration of its reliability. Here, neither (1) nor (2) were encouraging; altogether too little was known of the methodology employed to compile the forecast and nothing in Ps 9-13 of 1998, unaffected by the derailment, supported the assumption that the forecast was a reliable predictor of lost revenue.
24. Having disposed of the forecast, there was no proper basis for uplifting the calculation of lost passenger revenue, supported by Ms. Winspeare and based on the budget, yielding, as already noted, a claim of £1.865 million before allowances. As to the suggestion of a general uplift, Ms. Winspeare's calculation had already allowed for that. With regard to weeks 1 and 2 of P6, again, no uplift was justified: (1) there was no or no proper basis for extrapolating from Ps 4 and 5 in order to justify an uplift; (2) the WRR, properly interpreted, said nothing about whether actual performance in P6 weeks one or two had been ahead of budget (as distinct from a prediction for P6 as a whole); (3) the figure of 2.3% ahead of budget was in any event within the margin of error. If wrong on all that, then Mr. Popplewell submitted that any uplift for P6 weeks one and two should be limited to the order of £75,000.
25. Whatever the figure for lost passenger revenue (at least unless Dr. Harris's methodology was employed), it was then necessary to make deductions to strip out of the equation the extraneous factors which had nothing to do with the derailment. While accepting that precision was not possible in this regard, Mr. Popplewell submitted that it would be appropriate to make a sufficient allowance to bring the claim down from the "Winspeare" level (£1.865 million) to the "Harris" level (£1.6 - £1.7 million).
26. Looking at the matter in overall terms, the "Harris" calculation provided a cross-check for the "Winspeare" calculation and vice versa. Furthermore, powerful support for a figure in the bracket of £1.5 - £2 million was provided by GNER's own contemporaneous internal documents, which estimated a loss of that order. The nature of these documents was such that they were unlikely to understate the loss. These documents undermined the notion that the true level of claim was the forecast based £3.4 or £2.5 million advanced by GNER. In colloquial terms, GNER was being greedy.
27. Decision on lost passenger revenue: (1) The burden of proof: While, as already discussed, a search for precision would be illusory, it remains the case that the burden rests on GNER to prove what revenue it has lost by reason of the Sandy derailment.
28. (2) The methodology: I begin with methodology. I am amply satisfied that the right approach is that favoured by GNER; namely, to subtract actual passenger revenue for Ps 6-8 from the best estimate of revenue for that period; the shortfall thus obtained may then require adjustment in order to make such allowance as appropriate for any other factors, unrelated to the derailment, which have also contributed to the shortfall in revenue. To my mind, this is the obvious and straightforward method to adopt, albeit that there are difficulties in ascertaining (1) the best estimate of revenue for Ps 6-8 and (2) the allowance to be made for unrelated factors. I have, nonetheless, no hesitation in preferring this approach to that advocated by Dr. Harris (summarised earlier). As Dr. Harris accepted in cross-examination, his approach had never, to his knowledge, been applied in the context of seeking to put a reliable value on lost revenue arising out of an incident such as this. Notwithstanding such use to which Dr. Harris's approach has been put in the context of the PDFH, here I do regard it as over-theoretical and unduly complex. The approach of Dr. Harris does, however, have a residual value in that it permits a cross-check on the outcome reached by way of the preferred methodology. It follows that I reject Railcare's primary case as to methodology and prefer the approach contended for by GNER.
29. (3) Budget versus forecast: I turn next to the application of that approach. In practical terms, this involves the issue of budget versus forecast. I accept Mr. Schaff's submission to this extent that interest lies not in the budget or forecast in themselves but rather in establishing which provides the best estimate of revenue that would have been earned but for the derailment in the relevant period. Thereafter, I part company with Mr. Schaff. While, at least at first blush, the fact that the forecast post-dates the budget gives rise to an inclination to prefer a forecast based approach, any such inclination rests on having confidence in the forecast. Here, such confidence is lacking. Instead, to my mind the right approach accords closely if not entirely with that adopted by Ms. Winspeare; it is an adjusted, budget based approach, with the first two weeks of P6 stripped out and uplifted.
30. My reasons for rejecting an approach based on the forecast are these:

i) Too little is known of the methodology used in producing the forecast to justify placing weight upon it as a reliable estimate of revenue that would likely have been earned in Ps 6-8 but for the derailment. As emerged from the cross-examination of Mr. Montgomery and Mr. Segal, the forecast was based on unknown judgments of unknown GNER employees, or at any rate judgments to which these two witnesses could not speak. Those judgments had dictated a forecast outcome for 1998 as a whole of £317 million; the process was result-led; having chosen the desirable result, the £317 million was then allocated across the periods. Moreover, those responsible for the forecast had based themselves not on the 1997 actual figures (as might have been expected) but, for reasons which remain unexplained, on the 1996 figures. To those figures, a given percentage uplift, ultimately of 21.194%, had been applied to reach the pre-selected forecast outcome of £317 million. In short, the witnesses could do no more than explain arithmetically what had happened; they could not assist on the underlying reasoning nor were any documents available for scrutiny by them, by Railcare or the Court. The most these witnesses could do was to speculate. Against this background, at least unless verified by the closest correlation between the forecast outcomes and the actual results for periods unaffected by the derailment, I am quite unable to accept the forecast as a reliable basis for predicting the revenue which GNER had been likely to earn but for the derailment. I remind myself of the incidence of the burden of proof in this regard.

ii) GNER's next line of argument was that, whatever the methodology, the forecast was likely to be a more reliable predictor of actual outcome than the budget; moreover, that the budget would significantly under-estimate GNER's claim. The forecast was reliable and conservative; for Ps 6-8, GNER had forecast, on average, revenue 1.8% over budget. However, as demonstrated by Mr. Andrews, for 1998, Ps 1-5, GNER's actual passenger revenue was on average 2.4% over budget. For Ps 4 and 5, as Ms. Winspeare's own table revealed, GNER's actual performance had outstripped the budget by (about) 7% and 5% respectively. I am unable to accept this reasoning, depending as it does on extrapolation from 1998, Ps 1-5. First, for Ps 1-3, actual performance was broadly at budget levels. Secondly, granted that actual performance for Ps 4-5 exceeded the budget predictions, I am not persuaded that those two periods taken by themselves furnish sufficient data to permit the extrapolation of a trend through Ps 6-8 with any or sufficient confidence. Thirdly, when it comes to actual performance in (unaffected) Ps 9-13, the evidence points to the fact that the forecast proved to be a much less reliable guide than the budget; in my judgment, this outcome seriously undermines the claims made by GNER in favour of a forecast based approach.

iii) In similar vein, GNER placed much reliance on the WRR for P6, weeks 1 and 2, i.e., the two weeks immediately preceding the Sandy derailment. The argument here was that the budget performance indicator (contained in the WRR) uses the weekly revenue figures (on a cumulative basis within the period) to predict how the period will compare as against budget at the end of the period. For P6, week 1, the WRR budget performance indicator was 3.5% and for P6, week 2, it was 2.3%, i.e., ahead of budget. This consideration too pointed towards the budget under-estimating lost revenue, not least given that Mr. Montgomery in his evidence spoke of the WRR themselves under-stating the actual figures as against budget. Accordingly, the forecast rather than the budget was, at the least, the correct starting point for the estimation of lost revenue. Once again, I am not persuaded. First, the budget itself had no weekly breakdown with which to compare actual results. Secondly, the WRR do not show performance for those weeks as being ahead of budget; on the basis of partial sources only of GNER's revenue, they do no more than predict the outcome, compared with budget, at the end of the period in question. Thirdly, as demonstrated by Ms. Winspeare, very little reliance can be placed on predictions based on the cumulative position as at the WRR for week 2 of any period; so, for P1, the WRR prediction over-estimated actual performance by about 1.4%; for P2, the WRR contained an over-estimate of about 2.4%; for P3, an over-estimate of about 1.3%; for P4, an under-estimate of about 5% and for P5, an under-estimate of about 7.5%. In the circumstances, the data contained in the WRR do not persuade me to rely on the forecast, rather than the budget, as a starting point for the estimation of lost revenue over Ps 6-8.

For all these reasons, I cannot treat the forecast as a reliable guide to lost revenue in Ps 6-8.

31. (4) The Winspeare approach: As already foreshadowed, I favour Ms. Winspeare's budget based approach, with some adaptation, as giving the best approximation of estimated revenue which GNER stood to earn for Ps 6-8 but for the derailment. The budget was itself based on the actual 1997 figures. Putting weeks 1 and 2 of P6 to one side for a moment, Ms. Winspeare's approach accommodates GNER's legitimate concern that some adjustment is required to the budget figures to allow for GNER's performance exceeding budget predictions. Subject to such adjustments, the budget correlates relatively well with actual trading results for unaffected periods in 1998, taken as a whole. Turning to weeks 1 and 2 of P6, the logic of Ms. Winspeare's approach requires that they be treated the same way as other unaffected periods in 1998; that I shall seek to do. By way of elaboration:

i) Ms. Winspeare's methodology involves comparing actual and budgeted performance for 1998 as a whole, excluding Ps 6-8, to obtain an expected performance percentage. That percentage is then applied to the budgeted revenue for Ps 6-8; the difference between the expected revenue thus calculated and the actual revenue for Ps 6-8, yields the revenue lost by reason of the derailment – subject to such further deductions as are appropriate to strip out of the equation extraneous factors unrelated to the derailment (not a matter for Ms. Winspeare). Ms. Winspeare's methodology has much to commend it. First, she uses the data available for the year as a whole; to my mind, there is no good reason to ignore the data available for Ps 9-13 and extrapolation based on a greater volume of relevant data is more likely to be reliable. Secondly, it is to be noted that Ms. Winspeare realistically breaks the calculation down into ticket types (eg., discount, first and standard), thus recognising that expectations differ as between them, as did the impact of the derailment. Thirdly, Ms. Winspeare's approach does not rest content with the budget figures and thereby understate the calculation of lost revenue; to the extent that the actual figures for 1998 performance (in periods unaffected by the derailment) disclose that GNER was performing ahead of budget, an adjustment for such over-performance is built into Ms. Winspeare's calculation. As already recorded, if Ms. Winspeare's approach is followed, then (subject to any adjustments for extraneous factors), the revenue lost for Ps 6-8 by reason of the derailment was £1.865 million.

ii) Logically, it follows from Ms. Winspeare's approach that weeks 1 and 2 of P6 ought to be treated in the same manner as Ps 1-5 and 9-13; these two weeks were of course unaffected by the derailment. However, Ms. Winspeare's £1.865 million figure includes weeks 1 and 2 of P6 with the rest of Ps 6-8, not with Ps 1-5 and 9-13. At the hearing, there was some debate as to the correct treatment of weeks 1 and 2 of P6, in particular whether any uplift should be made and, if so, how it was to be quantified. Mr. Popplewell opposed uplifting Ms. Winspeare's calculation in respect of weeks 1 and 2 of P6; with respect, the logic of Ms. Winspeare's approach renders such opposition untenable. The real question to my mind, concerns the quantification of the uplift – not an easy matter as the discussion during the hearing revealed, even allowing for the fact that any such exercise would be rough and ready. Ultimately, Mr. Popplewell submitted that if (which he denied) an uplift was to be made it should be in the order of £75,000; for his part, as will be recalled, Mr. Schaff contended for a figure in the order of £200,000 or, at all events, not less than £150,000. On this point, I prefer Mr. Schaff's application of Mr. Popplewell's calculation. What needs to be done is to strip out weeks 1 and 2 of P6 from the affected periods and insert those weeks into the calculation in respect of the unaffected periods. In crude terms, both the top and bottom lines of the calculation needs adjustment. If such an exercise is conducted, the correct uplift is in the order of £150,000.

iii) Pulling the threads together, in my judgment, Ms. Winspeare's approach provides the best guidance as to lost revenue in Ps 6-8 (leaving aside any allowance for extraneous factors, to which I turn next). The logic of that approach requires an uplift for weeks 1 and 2 of P6. Doing the best I can, I allow that uplift in the amount of £150,000. It follows that before any allowance is made for extraneous factors, I assess GNER's lost passenger revenue attributable to the derailment at £2.015 million (£1.865 million + £150,000).

32. (5) Extraneous factors: I come now to the question of extraneous factors. To recap, these were matters unrelated to the derailment, canvassed in evidence and argument. To the extent, if any, that they resulted in GNER not achieving expected revenue in Ps 6-8, then, logically, they ought to be deducted from GNER's gross passenger revenue loss of £2.015 million; plainly, any loss flowing from these factors was not attributable to the derailment.
33. As it seemed to me, the parties' more extreme positions were unpersuasive. GNER at one stage suggested that if there was uncertainty as to the true extent of its loss because a figure could not be put on it, then I should make no deduction for extraneous factors; it was a matter of swings and roundabouts. With respect, such an approach ignores the fact that GNER must prove its loss, albeit that the figure arrived at will necessarily be somewhat rough and ready. The fact that GNER cannot "put a figure" on its loss does not justify a refusal to make a deduction for extraneous factors if (but only if) such deductions fall properly to be made. Conversely, a passage in Dr. Harris's evidence appeared to suggest that extraneous factors could amount to some £700,000, a figure to my mind lacking credibility. Realistically, the Railcare submissions did not press for a deduction in that amount. Railcare did, however, contend that deductions should be allowed of an order which would bring down Ms. Winspeare's figure for gross revenue loss of £1.865 million (without the uplift for weeks 1 and 2 of P6), to Dr. Harris's figure of about £1.7 million. I do not think that can be right either. It over-estimates the precision of Dr. Harris's approach and under-estimates the need for at least some justification (in terms of an evidential burden on Railcare) for the deductions in question.
34. As it seemed to me, the correct approach was to allow a modest deduction in the amount of £40,000 for extraneous factors. My reasons are these:

i) The extraneous factors canvassed included such matters as a London Underground ("LU") strike, weather, sporting events, competition from air travel and a tragic German rail crash, together with others which it is unnecessary to mention.

ii) Caution is obviously necessary. First, as agreed between Mr. Segal and Dr. Harris, budgets or forecasts prepared by GNER would have included an implicit level of "normal" disruption; no deduction should be allowed in respect thereof. Secondly, given that the issue relates to a revenue shortfall present in Ps 6-8 in circumstances where there is no shortfall in other periods, the search is for extraneous factors specific to those periods. Thirdly, to the extent that the impact of any extraneous factors cannot be quantified, then no deduction is to be made.

iii) Against this background, it is noteworthy that by the end of the hearing there was a measure of agreement or at least no sensible dispute that a figure of some £32,000 was to be allowed in respect of the LU strike. That was time specific, could not be budgeted for or forecast and was broadly quantifiable. By contrast, simply for example, even if competition from air travel was not taken into account in the budget, there is nothing to suggest that its impact in Ps 6-8 was any different from its impact on expected revenue in other periods.

iv) I do not, however, think it right to treat the £32,000 figure as exhaustive of the deduction to be made for extraneous matters. In the nature of things, extraneous factors are bound to have had some further impact on expected revenue. In my judgment, adopting a term employed by Mr. Andrews in a slightly different context, the correct "ballpark" allowance for such matters is to round up the £32,000 deduction to £40,000 and to make a deduction in that amount. In the present context, I am satisfied that a deduction of that order does practical justice.

35. The upshot is that I find GNER's net passenger revenue loss attributable to the derailment proved in an amount of £1.975 million (i.e., £2.015 million - £40,000). It is further agreed between the parties that there should be an additional deduction of £70,902 to take account of savings in retail commission and telesale unit commission. The overall net passenger revenue loss is, accordingly, £1,904 million.
36. (6) Cross-checks: I am fortified in reaching this conclusion by the following matters:

i) First, in very broad terms, the figure of £1.975 million is within the same "ballpark" as Dr. Harris's estimate of (about) £1.7 million.

ii) Secondly – and to my mind, more significantly – the figure of £1.975 million is very much in the same bracket as the assessments made of the likely level of passenger revenue loss, contained in a wealth of relatively contemporaneous GNER internal material. On the 30^th July, 1998, a sales and marketing report estimated the revenue loss for Sandy to be between £1.5 and £2 million. On the 16^th September, 1998, another sales and marketing report put the revenue loss up to the end of P8 at £1.65 million. The 1999 Budget Compendium dealt with the derailment in terms of a £1.8 million adjustment to 1998 revenue. In a Business Summary, dated 25^th January, 1999, GNER's Chief Executive spoke of £2 million of lost sales as a result of Sandy. A Business Update, of 19^th March, 1999, contained an estimate of a £1.9 million loss due to the Sandy derailment. While it is fair to say that Mr. Montgomery and Mr. Andrews in their evidence and some documentary materials spoke of losses in excess of £3 million being justifiable, I am satisfied that the true contemporaneous assessment of likely losses was of the order set out in the documents detailed above. For completeness, I am not dissuaded from this conclusion by Mr. Montgomery's assertion in evidence that at least some of these documents were, for reporting purposes, simply dealing with variations from budget; logical as that may be for certain accounting purposes, that is not how I read these documents in the context in which they came to be compiled.

37. Lost catering revenue: Mr. Andrews suggested a figure of £63,000 for lost catering revenue as a result of the derailment; Ms. Winspeare contended for £29,600. There was, however, no dispute between these experts as to a £0.94 per passenger yield loss. All that remained was the question of the number of passengers lost. That question hinged on the determination of the issue of lost passenger revenue, already discussed. In the light of my conclusion (broadly) in favour of Railcare on the issue of lost passenger revenue, it follows that I accept Ms. Winspeare's figure of £29,600 for lost catering revenue.

ISSUE (II): EXPENSES

38. This Issue concerns the ascertainment of expenses incurred by GNER attributable to Railcare's negligence. During the course of the hearing the good sense of the parties served to narrow the gap between them on this Issue; in the event, the remaining matters in dispute may conveniently be considered under the following headings: (A) Interfleet consultancy fees; (B) Key Services and Track Access payments; (C) Miscellaneous expenses.
39. (A) Interfleet consultancy fees: It is common ground that GNER is entitled to recover from Railcare £79,000 of the fees paid to consultants, Interfleet Technology ("Interfleet"), in the aftermath of the derailment. Beyond that, Railcare denies liability to GNER for the balance of the fees paid to Interfleet, amounting to about £515,000.
40. The facts: I start with the facts. Following the derailment, the fleet was made subject to an 80mph speed limit. Later on the night of the 16^th/17^th June and acting in conjunction with experts in rail safety and Her Majesty's Railway Inspectorate ("HMRI"), GNER withdrew their entire Electric 225 fleet from service. Visual checks were conducted of all similar wheelsets with balance holes, to look for any other significant cracks in the area around the balance weight holes; further, all similar wheelsets were subjected to Magnetic Particle Inspection ("MPI") testing with a view to ensuring that no such cracks remained undetected after visual inspection. Still further, some immediate remedial deburring and chamfering of balance weight holes was undertaken, so as to minimise the prospect of further crack development in the short term. After the initial visual inspection, trains were returned to service but at a reduced speed of 80mph. As and when sets had been subjected to MPI testing and the remedial deburring and chamfering work, they were allowed to operate again at the normal operating speed of 125mph; in the event, GNER resumed operating its normal timetable on (Monday) 22^nd June, 1998.
41. Pausing here, as already recorded, it is common ground that the derailment was caused by Railcare's negligence in failing properly to deburr and chamfer the balance weight hole in the wheel in question. The outcome of the investigations into the derailment was disconcerting; the problem was not one-off in nature. For example, the Railtrack Formal Inquiry ("the Railtrack Inquiry") recorded that, lamentably, of the 811 balance weight holes drilled by Railcare, only 2 were found to be properly chamfered; a Lloyd's Register document puts the figures at 765 and 4; the difference cannot matter. Plainly, Railcare was answerable for the costs associated with the remedial work necessary to deburr and chamfer holes properly. But the investigations also disclosed other unsatisfactory features, not the result of negligence on the part of Railcare of which complaint is made in these proceedings; for instance, wheels were found with balance weight holes drilled in the wrong position or wrongly angled. Indeed, the matter does not quite rest there, because so far as holes in the wrong position are concerned, an allegation of negligence was originally pleaded by GNER against Railcare in this regard and subsequently deleted by amendment. The present issue as to Interfleet consultancy costs may be seen to emerge from matters such as these.
42. Returning to the narrative, notwithstanding the initial visual inspections, MPI testing and remedial work, concern remained as to the safety of the fleet. MPI testing was accordingly continued at 10-day intervals. An important review of the safety and maintenance regime was conducted at a meeting held on 30^th July, 1998 ("the 30^th July meeting"), attended by GNER, Forward Trust and various technical experts including Messrs. Ward (from Serco Railtest), Rasaiah (from AEA Technology) and a number of representatives from Interfleet. In preparation for the meeting both AEA and Interfleet had produced reports.
43. According to the Notes of the 30^th July meeting, Mr. Lumb, GNER's Head of Engineering, stated that he was looking for "clear advice" as to what needed to be done. The Notes provide, inter alia, as follows:

" 2.3 The conclusions which both AEA and Interfleet independently reached were:-

It is possible to propagate a crack in a wheel with a pre-existing crack-like defect in it and hence this explains the wheel that failed at Sandy.

Cracks are extremely unlikely to initiate in wheels with 'good' holes. (See 3.1 for definition/criteria for a 'good' hole

There is no need to continue with wheel MPI provided the holes are to an acceptable standard (defined as a 'good' hole – see 3.1).

2.4 There is a requirement however to carry out on-track tests to measure real wheel web strains in order to underpin both the AEA and Interfleet theoretical work which has been carried out to date.

3. Definition of a 'Good' Balance Weight Hole

3.1 The following definition of a good hole was agreed following discussions involving all parties (See Appendix 1 for details).

A good hole is a balance weight hole which no longer requires MPI in service. To be designated a good hole it must satisfy the following criteria…."

There followed a list of criteria, including the requirement for a minimum chamfer, the absence of burrs between the hole and the chamfer and a description or definition of the acceptable position for the hole; a further note recorded that acceptable angles for the holes in the wheel were yet to be defined. The Notes then continued:

" 3.3 A key piece of information given by D. Ward at the meeting was that MPI would in this application guarantee to find a 1mm or greater crack size, if the MPI process were repeated at the same hole on 4 separate occasions. From this it was concluded that if the chamfer size was then increased by 1mm, following 4 MPI's this would ensure that any smaller undetectable defects were eliminated.

3.5 GNER will be developing a plan to re-assess all balance weight holes and implement measures to bring holes up to the standard which satisfies the criteria of a 'good' hole. This may require replacement of some wheels. GNER's proposal was for Interfleet to manage this task."

44. Thereafter, GNER embarked on the "good hole" programme, some times termed in the evidence, the "fit and forget solution", thereby dispensing with the need for further ongoing MPI testing. The good hole programme was carried out by Interfleet in two stages; stage 1, a trial period, took place between August 1998 and February 1999; stage 2, the implementation phase, took place between February and September 1999. The work was carried out over an extended period to minimise disruption. A relatively few balance weight holes were regarded as incapable of satisfying the "good hole" definition and the relevant wheelsets had to be replaced. It was common ground that the "good hole" programme made economic sense, obviating the need for the continuation of MPI testing.
45. In the meantime, MPI testing continued in August 1998 on the basis of a 10-day regime. By November 1998, the expert advice was unanimous that the regime could be relaxed and MPI testing conducted at 30-day intervals. Mr. Lumb, as he made plain in evidence, was content to do so. Forward Trust, however, refused to agree and GNER, as Mr. Lumb explained, could not act independently of the owners of the rolling stock. Ultimately, the change from a 10-day to a 30-day regime took place at the end of January 1999. Thereafter, 30-day MPI testing continued until September 1999.
46. The matters in dispute: Two issues arise in connection with the Interfleeet consultancy fees: (1) As to the disputed balance of the Interfleet fees, were such fees incurred by reason of Railcare's negligence ("causation") ? (2) Even if the answer to (1) is "yes", is Railcare answerable for the continuation of the 10-day MPI testing regime beyond November and the 30-day MPI testing regime beyond March ("reasonableness") ?
47. Reasonableness: It is convenient to take this issue first and to dispose of it at once. With respect, Railcare's resistance here lacked substance. The reasonableness overall of the "good hole" programme was, as already foreshadowed, not in dispute. If causation is established (the premise on which these questions of reasonableness arise), it is hardly unforeseeable that GNER might be subject to external constraints, as to how it addressed the problems flowing from the derailment. Nor was it unreasonable for GNER to decline to act independently of the views of Forward Trust – even if it was capable of so doing. Accordingly, persisting from November to January with the 10-day MPI testing regime was neither unreasonable nor unforeseeable. There is even less merit in the point taken as to the continuation of 30-day MPI testing from March to September. Such testing was required until the completion of the "good hole" programme; no criticism was or could sensibly be made of the scheduling of that programme, designed as it was to avoid disruption to GNER's service. Accordingly, if GNER is correct in its arguments on causation, then it is entitled to succeed on the topic of reasonableness.
48. Causation: I turn at once to the rival arguments. For GNER, Mr. Schaff's argument proceeded as follows. He started with the consequences of Railcare's negligence; such consequences were not limited to the derailment; they went much further. Railcare's negligence gave rise to the investigation of the whole fleet; it resulted in the need for GNER to put right the obvious defects relating to chamfering and deburring; it also caused the need to go on MPI testing. Indeed there is no issue as to that, at least up until (about) the end of July 1998. But the causative effects of Railcare's negligence did not stop there either. Following the 30^th July meeting, it was apparent that MPI testing could be discontinued only if balance weight holes fulfilled the "good hole" criteria. Hence, GNER undertook the "good hole" programme, which was more economical than indefinitely continuing with MPI testing. Accordingly, said Mr. Schaff, the disputed Interfleet costs of the "good hole" programme are recoverable by GNER from Railcare in damages, as a matter of causation or by way of mitigation. Put another way, when addressing the need to remedy defects as to chamfering and deburring, insofar as the holes exhibited other obvious defects "arising out of the shoddy drilling which has given rise to the burr or the chamfer", the only reasonable thing to do was to remedy those other defects at the same time. In short, the work and supervision relating to other defects in the holes, not individually the subject of allegations of negligence against Railcare, could not be separated from the work and supervision related to the remedying of defects as to chamfering and deburring. In this latter regard, Mr. Schaff placed much emphasis on Appendix 1 to the Notes of the 30^th July meeting, containing details of the definition of a "good hole". He further underlined what he said were Mr. Etwell's difficulties in separating out work required in respect of remedying deburring and chamfering defects from other work forming part of the "good hole" programme.
49. For Railcare, Mr. Popplewell submitted that the correct starting point was to identify those costs which would have been incurred if the problems with the wheels had been limited to Railcare's negligent failure to deburr and chamfer; such costs were recoverable; any other costs were irrecoverable, whether as a matter of causation or remoteness. Mr. Schaff's argument had confused the cause of the derailment with what was done in response to the defects disclosed by the derailment; it was necessary to distinguish cause from opportunity. Railcare's negligence had caused the derailment; such negligence and the derailment had furnished the opportunity to remedy other defects in the wheels for which Railcare was not responsible. Furthermore, on the evidence, the work relating to remedying deburring and chamfering defects was separable from other work related to the good hole programme. Had the only defects in the wheels consisted of those for which Railcare was liable (relating to deburring and chamfering), there would have been no good hole programme; all that would have been required were some 3 or 4 repeat MPI tests: see the evidence of Mr. Cooper. In this regard, Mr. Popplewell underlined that much of the "good hole" programme had nothing to do with the negligence of Railcare; for instance, work relating to incorrectly positioned and wrongly angled holes. While such work might have involved reaming, deburring and chamfering, it was not attributable to Railcare's negligence. As to Mr. Etwell's suggested difficulties in cross-examination on this point, these were not borne out by the transcript. Most tellingly, Mr. Popplewell drew attention to the following passage in his cross-examination of Mr. Lumb:

" Q. Now, would you accept that if the advice that you had been getting about these wheels was that the only problem, or potential problem about them was that they had not been properly deburred and chamfered – I appreciate that you were getting different advice and there were other problems with them but I would like you to put yourself in [that] position: if the advice you had been getting was the only potential problem on the wheels was that they needed to be deburred and chamfered, would you accept that the good hole programme would have been complete by about 5^th August ?

A. Well, if that was the – those were the only two things that needed to be done, there would have been no necessity for a good hole programme

…

Q. I think everybody agrees it probably would have been appropriate to have some MPI testing at 10-day intervals for, I think on your side they say four periods, and on our side they say three periods ?

A. Yes, the advice we had was that if you did four – and it was possibly three and four to be absolutely sure – then you should pick up a latent small crack. And the reason for the four is that if you miss it on one of them, you probably get it on the next one.

Q. And the rectification that had taken place between 16^th and 22^nd June ought to have remedied any of the original deburring or failure to chamfer problems on the wheels because that is what you were rectifying over that period, was it not ?

A. What we were trying to do over that period was to make the wheels and the holes as safe as we could in the time and with the resources we had available. At that stage, we were right at the beginning of this exercise we had, as it were, ad hoc advice in terms… advice which was being developed as the situation became more understandable, and clearly, we wanted to get our trains back into service as soon as possible."

Pulling the threads together, Mr. Popplewell's submission was that all the work attributable to Railcare's negligence was complete by the 5^th August; it involved the initial deburring and chamfering undertaken between the 16^th and 22^nd June, plus 3 or 4 repeat MPI tests; Railcare was not liable for Interfleet costs beyond the £79,000 agreed in respect of the period up until the 5^th August; if wrong in that and some additional works were needed after the 5^th August, or if the 3-4 additional tests then remained to be completed, Railcare was not liable for more than the £26,000 odd relating to Interfleet's invoice in respect of August, 1998.

50. In reply, Mr. Schaff emphasised the advice tendered to GNER and the reasonableness of its decision to undertake the "good hole" programme. Furthermore, he cautioned against over-emphasising Mr. Lumb's evidence, by questioning, in the light of an exchange in the course of his evidence-in-chief, how much consideration had gone into any such breakdown of the works involved; on any view, it was important not to read Mr. Lumb's answers too literally. It would not be right to assume that a guillotine would have come down at the end of July or on the 5^th August. Not least, it might well have been necessary to re-visit some of the work done in such haste between the 16^th and 22^nd June. It is convenient to record here Mr. Schaff's argument in his principal closing submissions that if GNER was not entitled to the whole of the Interfleet costs relating to the "good hole" programme, then it should at the least recover: (1) £26,000 in respect of repeat MPI testing or other works spilling over into August 1998; plus (2) £140,000 in respect of doing what he submitted was deburring and chamfering rectification work itself; (3) £90,000 of supervision costs relating to such rectification work. On this (reduced) footing, GNER would recover some £256,000 in addition to the £79,000 agreed for Interfleet costs. In fairness to Mr. Schaff, he made plain that this was a broadbrush approach, based on a helpful note prepared by Mr. Picken, doing his best with figures which had not been broken down in any such way.
51. Decision on causation: Having set out the arguments at some length, I can state my decision and give my reasons almost summarily. In broad terms, I preferred Mr. Popplewell's argument on the causation issue, albeit with some modifications. My reasons are these:

i) The reasonableness of the "good hole" programme is not in dispute. That factor, though not irrelevant to the present issue of causation, cannot, however, determine it.

ii) I think Mr. Popplewell was right to take as his starting point the costs properly attributable to Railcare's negligence in failing to debur and chamfer. Other costs, if separable, are irrecoverable at least as a matter of causation. For my part, with respect, Mr. Schaff's analysis tended to elide cause and opportunity; Railcare's negligence and the derailment provided no more than the occasion for other defects in the wheels to be remedied; provided only that they were separable, those defects and the costs relating thereto were not caused by Railcare's negligence.

iii) As to whether it is possible and appropriate to separate out the costs relating to remedying defects in deburring and chamfering from other costs related to the "good hole" programme, this is a matter turning on the evidence. It is unnecessary to lengthen this judgment with references to the transcript of Mr. Etwell's cross-examination; suffice to say, with respect to Mr. Schaff's argument to the contrary, that I do not accept his submission as to the significance of some relatively opaque exchanges. To the contrary, however, I do regard the evidence of Mr. Lumb, GNER's head of engineering, as of the first importance. To my mind, his answers were clear; had the only problems with the wheels related to a failure properly to deburr and chamfer, there would have been no necessity for a "good hole" programme. I accept that evidence.

iv) It does not, however, follow from Mr. Lumb's evidence that a "guillotine" at once came down on the work which needed to be done at the end of July or the beginning of August. I agree with Mr. Schaff that Mr. Lumb's evidence should not be read too literally. Accordingly, first, I have no doubt that GNER is entitled to recover an additional £26,000 (to the £79,000 agreed) in respect of 4 additional MPI tests or other Interfleet costs in August, 1998. Secondly, I am persuaded by Mr. Schaff that it is probable that some further works would have been required (even on this hypothesis), together with related supervision costs, in respect of revisiting or rounding off the remedial work undertaken as a matter of urgency between the 16^th and 22^nd June. The difficulty is putting a figure on it. In effect, Mr. Schaff's fallback case involved a claim for some £230,000 in this regard; to my mind, a figure of that order cannot be justified. Doing the best I can having regard to the costs of the "good hole" programme as a whole, I am satisfied that a figure of £50,000 is unlikely to have been avoidable, even if the only defects in the wheels had related to Railcare's failure properly to deburr and chamfer. Such a figure is realistic if (as on this hypothesis) the initial remedial work was not to become part of a larger "good hole" programme.

v) I therefore conclude that in addition to the agreed £79,000 in respect of Interfleet's consultancy fees, GNER is entitled to recover £76,000 (£26,000 + £50,000), as explained above.

52. For completeness:

i) In his submissions, Mr. Schaff contended that if GNER was to be deprived of the disputed Interfleet costs of the "good hole" programme, it was incumbent on Railcare to point to some novus actus interveniens. Mr. Popplewell countered by saying that this was not a case of novus actus or supervening event at all. Suffice to say that, in effect for the reasons already set out, I preferred Mr. Popplewell's argument in this regard.

ii) With a view to the analysis of liability for the disputed Interfleet costs, reference was made in the course of argument to the discussion in McGregor, at paras. 1344-1348 and to Carslogie SS Co. v Royal Norwegian Government [1952] AC 292. With respect, I do not think that there is a ready analogy from that discussion or authorities such as Carslogie, which assists in the resolution of the problem encountered in the present case.

53. (B) Key Services and Track Access payments: These two items can be taken together – and shortly. I accept the GNER claim in this regard.
54. The nature of the Key Services claim: As explained by Mr. Montgomery, there is a Key Services performance agreement which, broadly, provides for a payment by GNER to Railtrack when, sufficiently often, trains arrive at their destinations within a specified variance from the scheduled arrival time. When trains are cancelled through no fault of Railtrack, the agreement has been operated on the basis that the trains are, for these purposes, deemed to arrive on time. The claim is for the amount - £42,000 - which GNER has paid to Railtrack in respect of deemed punctual arrivals in excess of that which GNER would have expected to have paid had the services not been cancelled. The payment has actually been made. The GNER calculation of the sum due enjoys the support of Mr. Andrews.
55. The Railcare objections: These were twofold. First, that the Key Services agreement made no express provision for payment when trains were cancelled. There was no proper basis for the implication of a term. The payment was gratuitous and is irrecoverable. Secondly, that the amount of the payment was speculative so that GNER had not proved its loss; further or alternatively, the loss was too remote.
56. Decision as to Key Services payment: I reject the Railcare objections. I accept the evidence that a payment in the amount of £42,000 was made. I further accept Mr. Montgomery's evidence as to the operation of the agreement. If the choice lies between a gratuitous payment and one made by virtue of legal obligation, I am amply satisfied it was the latter. The source of that legal obligation was any one or more of an implied term, estoppel by convention or an implied variation. As to proof of the claim, I accept Mr. Montgomery's and Mr. Andrews' evidence. As to remoteness, I am satisfied that such a payment to Railtrack was reasonably foreseeable by and eminently within the contemplation of Railcare. I keep in mind that, as the tortfeasor, all that Railcare is required to foresee is the type of loss. I allow the £42,000 claimed in full.
57. The nature of the Track Access claim: This is a similar claim. In essence, GNER makes payment to Railtrack when trains are delayed due to its (GNER's) fault. The claim here is for the increased amount - £102,000 – paid by GNER in respect of delayed trains consequent upon the derailment, compared to the payment it estimates it would have made in respect of delays to trains had there not been a derailment.
58. The Railcare objections: Again, these are, first, that there was no legal liability to make such payments; the payment made was accordingly gratuitous. Secondly, that the amount of the payment was speculative, the loss was not proved and/or it was too remote.
59. Decision as to Track Access payment: I reject the Railcare objections. I accept that the amount in question was paid by GNER to Railtrack. As to legal liability, my reasons are, mutatis mutandis, the same as those given in respect of the Key Services payment. As to the calculation of the amount to be paid, I accept Mr. Montgomery's evidence, based on experience. As to remoteness, Mr. Popplewell's suggestion that this was a payment for a loss of opportunity to earn a contractual payment was, with respect, artificial. In my judgment, this was a compensatory payment to Railtrack, consequent on the derailment. I regard as unreal the submission that it was outside the relevant contemplation of Railcare, an operator within the railway industry. I allow the £102,000 claimed in full.
60. (C) Miscellaneous Expenses: To my surprise and some regret, there remained other items of expense in dispute. I shall deal summarily with them:

i) Season Ticket Vouchers: Ultimately, as summarised in the helpful schedule prepared by Mr. Cooper, GNER claimed £60,860 and Railcare was prepared to accept liability for £54,994. The dispute is as to encashment rates. I am attracted to Mr. Montgomery's oral evidence in this regard and accordingly allow this claim in the amount of £60,000.

ii) Vouchers given to delayed passengers: Ultimately, GNER claimed £96,811 and Railcare was prepared to accept liability for £87,173. Consistently with i) above, I tend to favour GNER on this point. I allow this claim in the amount of £95,000.

iii) Catering wastage: Ultimately GNER claimed £24,524 and Railcare was prepared to accept liability in the amount of £10,000. Broadly speaking, I preferred GNER's case to Railcare's objections. I allow this claim in the amount of £20,000.

ISSUE (III): CONTRIBUTORY NEGLIGENCE

61. Introduction: The derailment was caused by a fatigue crack in the circumferential web of the wheel propagating to failure. On the evidence, such a crack was unprecedented in Mark IV wheels, although it is also fair to say that the possibility of fatigue cracks occurring in wheels from holes was not an unknown phenomenon.
62. As is to be expected, GNER had a maintenance regime, involving regular inspections of train wheelsets. The relevant inspections here were "B" and "A" examinations. In broad terms, "B" examinations were required every 90 days, "A" examinations every 10 days. A "B" examination took place on the 24^th March, 1998; between that date and the derailment, 7 "A" examinations were conducted. The crack was not detected during any of these examinations.
63. The requirements of a "B" examination were as follows:

" 1) Inspect wheels visually for cracks, flats, pitting, damage or any sign of possible failure in tyre, rim and hub.

2) Inspect brake discs visually for cracks, damage or any sign of possible failure.

3) Inspect visible parts of axle for cracks, damage or any sign of possible failure.

4) Check that the balance weights are not loose. Refer to Maintenance Procedure U-34 for re-torquing. Record any that are found loose on the exam/repair card.

5) Record and report any defects to the Supervisor."

64. The requirements of an "A" examination were, insofar as relevant, as follows:

" 5) Visually check the wheelsets for signs of damage or defect. …

6) Record and report any defects to the Supervisor."

65. There was a considerable volume of factual and expert evidence on the present Issue. It is unnecessary to recount all of it; instead it is convenient to record central items of evidence in the course of outlining the parties' cases. It is further convenient to divide this Issue into the following sub-issues: (1) The charge of negligence; (2) The financial consequences; (3) Apportionment.
66. The charge of negligence: (A) The Railcare case: Railcare alleged that GNER was negligent in failing to detect the crack prior to the derailment. Such negligence was causative of the derailment, alternatively contributed to it. In its final form, Mr. Popplewell's case was that the crack should have been detected (1) in the course of the "B" examination on the 24^th March, 1998, alternatively, (2) at some time in the course of the 7 subsequent "A" examinations which took place between then and the derailment. No case was advanced in respect of any examinations prior to the 24^th March.
67. Mr. Popplewell submitted that this case did not depend on hindsight. His argument was developed as follows. He began with the length of the crack at the time of the "B" examination on the 24^th March. By the first day of the trial, this had become common ground, as follows:

Clockwise (cm) Anticlockwise (cm) Total (cm)

Inner 12.8 15.6 28.4

Outer 7.8 14.0 21.8

68. On the basis of these dimensions, Mr. Popplewell next submitted that the evidence supported his case. First, there was the evidence of Mr. Parsons, the GNER depot manager, with responsibility for train inspections and maintenance at Bounds Green Depot, where the "B" and "A" examinations took place. It is not unfair to say that Mr. Parsons was somewhat defensive in his evidence, which makes the following passage in his cross-examination, on "A" examinations, noteworthy:

" Q. … if there was a crack through the wheel which had penetrated both onto the inner and outer surfaces there ought to be two opportunities to see it, both underneath the train and along the side of the train ?

A. That is correct.

Q. Now, paragraph 12 of your statement …. You say at the end of Paragraph 12: 'If there had been a very obvious crack then our A examiners would certainly have been expected to note and report it.' How long would a crack be before it would be something that you would describe as 'very obvious' ?

A. Using the terminology 'obvious' in that the crack would be visible and open. The length is relatively immaterial if, in our view, the crack is obvious.

Q. So, if it is an open crack, how long or how short would it be before you would describe it as something your A examiners would have been expected to note ?

A. It is very difficult to put a figure to that.

Q. Well, let me try to help you. Suppose it is a total of 85 centimetres long, both sides of the balance weight. A competent inspector ought to notice that on an A exam ?

A. I would have expected that to be so.

Q. The same is true of anything above 30 centimetres, is it not?

A. Again, I would expect the crack to be visible. I would not only be persuaded that the crack was visible by its length.

Q. You have made that point but assuing that it is a crack which is visible and an open crack, if it were 30 centimetres or more, you would expect someone to see it on an A exam if they were doing their job properly..?

A. I would say yes. "

69. Secondly, Mr. Popplewell made considerable play with Mr. Ward's changed stance. Mr. Ward had shifted his ground from (1) saying that the dimensions of the crack must have been much smaller than those ultimately agreed because otherwise it would or should have been seen, to (2) saying that the crack was, if not a "closed crack", then a fine, hairline crack which could be difficult to see. Plainly, submitted Mr. Popplewell, the crack could not be a closed crack; had it been such, it could not have propagated. Not only did these changes of stance discredit Mr. Ward; they also permitted the inference that the crack ought to have been detected in the course of the relevant examinations. Even if on certain occasions, depending on the rest position of the wheel, it was difficult to see, it should have been detected in the course of some examination between the 24^th March and the derailment. Moreover, relying on the evidence of Mr. Cooper, external signs of a crack would have been present.
70. Thirdly, Mr. Popplewell placed much emphasis on the evidence of Mr. Etwell. This was a matter of substance not labels; the substance of Mr. Etwell's view was that there had been negligence in failing to detect the crack, both in relation to the 24^th March "B" examination and the subsequent "A" examinations; these were failures in the inspection regime and the standards of inspection had been lacking. Even the "A" examination required the inspector to examine the wheelsets for signs of damage and defects; such damage or defects included cracks: see item 5) thereof (set out above); all the more so, in the case of a "B" examination. Had he been depot manager, such failures would have resulted in disciplinary action.
71. Building on all this evidence, Mr. Popplewell underlined that the "B" examination allowed about one minute per wheel for checking – sufficient for a detailed inspection of the wheel. Reverting to Mr. Etwell's evidence, when the detailed "B" examination was combined with the frequency of the "A" examinations, the crack should have been detected and the derailment prevented.
72. As to the 24^th March "B" examination itself, Mr. Cooke had conducted it and had recorded the discovery of a loose balance weight on the wheel in question. Mr. Bond had returned on the 27^th March and tightened the balance weight, or claimed to have done so. Accordingly, the attention of both men had been drawn to the balance weight hole. As expressed in Railcare's closing written submissions:

" There can only be one of two explanations for why the crack was not spotted at this B examination. Either neither man looked properly at the area surrounding the balance weight hole or they did not look at all."

In short, Messrs. Cooke and Bond should have found the crack; their failure to do so constituted negligence.

73. Mr. Popplewell acknowledged that there had been three investigations into the derailment and that none had criticised GNER. That said:

i) The HMRI investigation had proceeded on the erroneous assumption that the crack was at the time not of a length so as to be visible beyond the balance weight hole;

ii) A Lloyd's Register audit had focussed on very different matters and did not purport to address any questions relevant to the present inquiry;

iii) The Railtrack inquiry was convened very promptly but at that time did not have and could not have had a true understanding of the actual level of crack length at the time of the prior examinations; further, the panel included messrs. Tham and Lumb, both of whom gave evidence in this case and were at the time employees of GNER; Mr. Tham had been Production Director, with responsibility for both operations and engineering; Mr. Lumb, as already recorded, had been GNER's head of engineering; the conclusions of such a panel could not weigh too heavily in the scales.

74. Overall, GNER's position was unattractive and unacceptable. It argued that this crack could not reasonably have been detected by visual examination. Yet, even today, its inspection regime remained based on visual examination.
75. (B) The GNER case: In dealing with this Issue, Mr. Schaff pointed to the curious evolution of the Railcare case. As pleaded, it had involved criticisms of the failure to detect the crack in both the "B" and "A" examinations. For much of the trial, however, it had appeared that the sole focus rested on the 24^th March "B" examination; ultimately, however, Railcare had had second thoughts and now sought to develop an independent case based on the "A" examinations subsequent to the 24^th March, even had GNER not been negligent in relation to the "B" examination. Mr. Schaff submitted that the charge of negligence or contributory negligence was not made good, whether based on the "B" or the "A" examinations.
76. In Mr. Schaff's submission, the Railcare case was expert-led and hindsight-based. At the time, there had been no criticism of a visual system of inspection. Whatever might be said of the composition of the panel, Mr. Schaff placed reliance on the observations contained in the Railtrack inquiry report:

" 3.2.1. It is clear that the only relevant problem with the vehicle was the very poorly finished balance weight fixing hole.

3.2.5. Inspection and maintenance of the vehicle had been carried out exactly as laid down at the time.

3.2.6. With hindsight, the maintenance regime which only ever provided a visual inspection of wheelsets, must be questioned.

3.2.7. It seems certain that, at 24 March 1998 'B' examination, the cracks were already quite well developed, having probably been propagating since June 1997. While dirt and debris particles on the wheel combined with the effects of artificial lighting from the pit, may well have made the defects invisible to the naked eye, magnetic particle inspection at this stage would surely have revealed the problem.

3.2.9. By 'A' examination dated (10 June 1998), the cracks must have been very well advanced. However, wheel inspection (even visual) was not part of the process…."

Pausing there, Mr. Schaff reminded me that both the Railtrack inquiry and the HMRI Investigation knew of the wheel having failed and the "A" examinations which had taken place – whatever they did not know of the crack length as at the 24^th March. Yet neither report contained anything by way of criticism of GNER. Further, it was plain from the Railtrack inquiry that GNER's system of inspection was not criticised; the case against GNER could, accordingly, only rest on the implementation of that system.

77. Turning to implementation, Mr. Schaff argued that there was no or no proper foundation for the allegation that Messrs. Cooke and Bond had been negligent; indeed, the allegation was unfair. These witnesses were straightforward; there was no reason to doubt their evidence that they are likely to have done what each said he had done, in accordance with his usual practice and the documentary records; Mr. Cooke, that he had noticed the balance weight being loose and Mr. Bond, that he had tightened it. They had been cross-examined to the effect that they had not inspected the area closely; but that went nowhere, unless (1) the system of inspection was under attack; or, (2) in the case of Mr. Cooke, that he had reason to anticipate a circumferential fatigue crack in way of the balance weight hole; or, (3) in the case of Mr. Bond, that a failure to tighten the balance weight was causative of the derailment (which it was not). They had not been cross-examined on the basis that having inspected the area closely, they had missed an obvious and visible crack. Mr. Bond, it was to be recollected, was in any event not called in to carry out any examination, let alone to detect cracks in the web.
78. As to the nature of the crack, Mr. Cooper (Railcare's metallurgy expert) had accepted in cross-examination that fatigue cracks were notoriously "fine" and that closure can result from compressive loads – here, the static load of the coach. It followed that much depended on the rest position of the wheel. The conundrum was that: (1) when the crack was in about the 12 o'clock position, it would be most "open" in terms of compressive loads but least visible because of the axle; whereas (2) when the crack was between the 4 and 8 o'clock positions, it would be most visible but most "closed" because of compressive loads. In all the circumstances, including factors such as light, dirt and water, it would have been difficult to see the crack. It was to be remembered that conditions in court did not approximate those in the depot, not least when it was not to be anticipated that there was a crack in the wheel web.
79. With regard to the examinations, Mr. Schaff argued that the "B" examination directed attention to locations other than the web insofar as cracks were concerned: see item 1) (set out above). The only mention of balance weights concerned the requirement to check that they were not loose: see item 4) (set out above). The reason was simple; there was no history of circumferential cracks in Mark IV wheels. As to Railcare's late attempt to develop a case based on the "A" examinations, it was mis-placed. It wrongly assumed that "A" examinations were to be treated as equivalent to "B" examinations. They were not; "A" examinations were cursory, involving no more than a 30-40 minute walk down the train once in every 10 days looking for obvious defects. In any event, there was no specific requirement in the "A" examination to look for cracks.
80. Too much had been made by Mr. Popplewell of the evidence of Mr. Parsons. The premise underlying the answers given by Mr. Parsons in the passage of his evidence set out above was that the crack was "visible and open". If but only if so, Mr. Parsons would have expected it to be noted. Properly considered, such evidence did not advance the inquiry, let alone Railcare's case.
81. Mr. Etwell's evidence required close scrutiny. Much of his opinion, submitted Mr. Schaff, was based on hindsight. Instructively too, his view that the crack should have been detected during the "A" examinations was a matter which had evolved later on in his reports. Perhaps most significantly, in evidence, Mr. Etwell had been unwilling to state in terms that GNER had been negligent; instead, Mr. Etwell had viewed the matter in terms of possible disciplinary implications not, or not necessarily, equating to negligence. In this regard, two passages in Mr. Etwell's cross-examination merit quotation; first:

" Q. So, maybe you do not like the use of the word 'negligence' but at the end of the day, it is your opinion that there has been an unprofessional B inspection and an unprofessional series of A inspections which do not live up to the standards of competence that you would expect ?

A. I do not feel comfortable agreeing to those words … I would use different words, but if this is the implication, I repeat the wording that I would prefer is that to me, it is an indication that the inspection regime has failed and that the standards of inspection have been lacking.

Q. Both on the B and also on the A ?

A. Both on the B and on the A."

Secondly:

" Q. … what I have to suggest to you, Mr. Etwell, is that this is a pretty serious allegation, is it not, that is being run in this case against GNER ? The case is that GNER's negligence – admittedly along with the defendants – caused the derailment. That is the allegation, is it not ?

A. I have not made that allegation …."

Finally, whatever Mr. Etwell's earlier observations as to "B" and "A" examinations, his later remarks told against there being negligence on the occasion of the "B" examination taken by itself:

" Q. … What I am putting to you is: the reason why you have frequent inspections, frequency being something that you have particularly stressed, is that because there is a risk that an operator may not detect a crack on a particular inspection?

A. Yes.

Q. And an operator may not detect a crack, even a significant crack, on a particular occasion, for all sorts of reasons which have nothing to do with professional culpability ?

A. On a once-off basis… yes, but you rely on it being done and that is with the frequency. The B exam cannot be looked at independently of the A because the B exam does include the A exam content and the frequency of the safety inspection is heavily dependent upon this general visual examination on the A exam frequency. So on a once off, one may say yes to answer that question, but the general answer, I would have to say no.

Q. … on a once-off basis on 24^th March, 1998, you cannot say, can you, that the failure to detect the crack on this one-off occasion is only consistent with negligence ?

A. Well, I can only repeat the context of my previous statement, that I could only look at that B exam really in the context of regular A examinations as well as the B with the A examination over a pit in a similar inspection situation in visibility. So, I would expect it to be found at one of those examinations."

If so, making good the charge of negligence with regard to the "A" examinations was central to Railcare's case.

82. As to the continued use of visual inspections, Mr. Schaff's response was as follows:

i) Balance weight holes were phased out;

ii) There was MPI testing for a very considerable period;

iii) The current requirements for "A" examinations now specifically direct attention to checking for cracks in the areas around the holes in the wheel webs.

83. (C) Decision: Having explored the rival cases and the evidence at some length, I move directly to my conclusions and the reasons for them. In summary, I have not been persuaded that GNER was negligent in failing to detect the crack prior to the derailment. My reasons follow.
84. To begin with, it is right to mention that I found certain aspects of GNER's evidence disquieting. So:

i) First, GNER promulgated a maintenance regime documented in terms of days but in fact mileage based. In the event, the "B" examination on the 24^th March was on any view overdue in terms of days but not in terms of mileage. For my part, I had no difficulty in accepting that: (1) having a specified interval in terms of days between examinations provided a ready guide to when examinations were due; (2) from an engineering perspective, what really mattered was mileage between examinations; (3) other operators used a mileage based system. What was less satisfactory was that the documentation and the underlying philosophy did not appear to be at one. Matters were not improved by the defensiveness displayed both by Mr. Parsons and Mr. Lumb in their evidence, when dealing with this topic; moreover, neither was in a position to explain precisely why the "B" examination was late (in terms of days). Manifestly, any maintenance regime must be simple to understand and implement if sloppiness is to be avoided.

ii) Secondly, after the derailment, it was ascertained that of some 811 holes, 270 were empty in the sense that balance weights were missing. In his evidence, Mr. Bond had said that if any empty hole was discovered on an "A" or "B" examination, it would be recorded. That evidence was contradicted by Mr. Lumb, who had been unaware of this problem, although head of engineering for some three years at the time of the derailment. Although he said that this ranked lower down the list of pressing safety issues than a number of other matters, it would have been a matter of concern to him had he known about it. This topic accordingly does not provide reassurance as to GNER's management grip; additionally, it suggests that, at least in this aspect of his evidence, Mr. Bond was (so to speak) gilding the lily.

That said, I do not regard either of these matters as in any way causative of the derailment. To my mind, the leap which Mr. Popplewell sought to make from the lateness of the "B" examination and the confusion surrounding the question of days or miles, to sloppy conduct of the work undertaken on the examination, is not justified. Likewise, while it is, to say the least, unfortunate that Mr. Bond's evidence was inaccurate in connection with the reporting of empty holes and that Mr. Lumb was unaware of this problem, there were in fact (as the subsequent investigations tended to suggest) good reasons for not replacing balance weight holes. Other than observing that there are managerial lessons for GNER to learn from these matters, it is unnecessary for me to say more about them.

85. Returning to the main theme of this Issue, it is plain that no proper criticism can be made of GNER having operated a visual system of inspection. There is no hint of any contemporaneous criticism in this regard and, indeed, Railcare did not or not seriously pursue any such complaint. The question of implementation of the visual system is accordingly central.
86. In my judgment, the allegation of negligence in relation to the "B" examination of 24^th March, 1998, falls well short of its mark. First, I am not persuaded that either Mr. Cooke on that day, or Mr. Bond when he returned on the 27^th March to tighten the balance weight (and assuming, without deciding, that such work formed part of the "B" examination) was negligent. I see no reason to doubt that they did what the contemporaneous documentation records them as having done. I am not deterred from this conclusion by the separate criticism which I have already thought it appropriate to make as to another aspect of Mr. Bond's evidence. I also think that there was some force in Mr. Schaff's submission that it would be unfair to hold either Mr. Cooke or Mr. Bond negligent for having failed to detect the crack in the light of the slant or emphasis of their cross-examination. Be all that as it may, secondly, given the lack of any history of circumferential web cracks in Mark IV wheels, I do not think that Messrs. Cooke and Bond were negligent in failing to detect the crack here. There is a world of a difference between detecting a crack which is anticipated to be present and detecting a crack for which there is no precedent and hence unanticipated. Furthermore, there are simply too many unknowns as to lighting conditions and the like, not to mention the rest position of the wheel, to justify any conclusion of negligence in this regard. Significantly to my mind, Mr. Etwell, as has been seen, was manifestly unwilling to lend support to an allegation of negligence in respect of the "B" examination of 24^th March on a "once-off" basis and without having regard to the subsequent "A" examinations. Accordingly, Railcare's principal allegation of negligence as pursued through the trial, going to GNER's negligence in the conduct of the "B" examination of 24^th March taken by itself, must fail.
87. There remains the question of the seven "A" examinations, subsequent to the 24^th March and preceding the derailment. As already recorded, the Railcare case in this regard developed somewhat curiously, almost by way of afterthought. For my part, I think that Railcare's and Mr. Popplewell's earlier thoughts were more soundly based. While my concern lies with substance, not with any literal construction of the requirements for "A" examinations, essentially for the reasons canvassed by Mr. Schaff, I am not persuaded that "A" examinations are to be equated with "B" examinations. I am satisfied that for this, then unprecedented, crack, the Railcare criticisms depend on hindsight. Furthermore, though I accept Mr. Popplewell's submissions that it is the substance of Mr. Etwell's evidence which matters rather than the labels which he attached to his conclusions, I am not persuaded that Mr. Etwell was willing to support Railcare's case of negligence in this regard. Having carefully considered Mr. Etwell's evidence on this topic, in my view, he was concentrating on matters which, with the standards upon which he would have insisted, he would have hoped would and should have been detected; hence, his focus on disciplinary matters; he was not, however, prepared to take the extra step of branding GNER's conduct of the "A" examinations as constituting negligence or contributory negligence causative of the derailment. Most tellingly, when Mr. Schaff bluntly put to him the Railcare case that GNER's negligence had caused or contributed to the derailment, Mr. Etwell's plain and unqualifed answer was "I have not made that allegation ….".
88. I have not in all this overlooked the passage in the evidence of Mr. Parsons much emphasised by Mr. Popplewell. Having read and re-read it, I am persuaded by Mr. Schaff that it was indeed premised on the assumption that the crack was "visible and open". The passage will not therefore bear the weight that Mr. Popplewell sought to place on it. In any event, in the light of my conclusions as to Mr. Etwell's evidence and given all the circumstances of the Railcare case on the "A" examinations, I would have been reluctant to reach a conclusion adverse to GNER on this issue, simply on the basis of Mr. Parsons' evidence. For completeness, I was not persuaded that Mr. Ward's change(s) of stance added much, if anything, to the Railcare case.
89. As to the various investigations following the derailment:

i) I accept Mr. Popplewell's submission that the Lloyd's Register audit was irrelevant to the matters presently under consideration.

ii) While caution must be exercised with regard to the HMRI investigation and the Railtrack inquiry - both because of the then imperfect knowledge as to the dimensions of the crack and because the Railtrack inquiry panel, including as it did Messrs. Lumb and Tham, can hardly be regarded as independent – it is not without some significance that neither even hinted at criticism of GNER. At the least, this fact tends to reinforce the conclusion to which I have come on the Railcare charge of negligence against GNER.

iii) In this connection, it is to be noted that the HMRI is a part of the Health and Safety Executive ("HSE"). The cause of the derailment and the existence of the "A" examinations must have been known at the time of this investigation. Had there been proper scope for criticism of GNER, I cannot but think that at least the emphasis of the HMRI report would have been different. In the event, the decision was taken to prosecute Railcare but not GNER.

iv) The Railtrack inquiry was conducted, as I understand it, in accordance with Railway Group Standard GO/RT3434/3, "Accident Investigations and Formal Inquiries", a part of Railtrack's "safety case", issued pursuant to The Railways (Safety Case) Regulations 1994. Whatever misgivings there might be as to the independence of certain panel members, it would be wrong to devalue the importance of prompt industry inquiries such as this, involving individuals with obvious practical experience. Nothing in the conclusions, the terms of which are set out above, lends any support to the case against GNER here.

90. For completeness, nothing turns on the system of inspections currently in use.
91. Looked at overall, GNER's failure to detect the crack prior to the derailment is admittedly troubling. But for all the reasons already discussed, I am not satisfied that the charge of negligence advanced by Railcare has been made good. In a nutshell, I think it is essentially a hindsight case. A finding of negligence would be harsh and, on all the evidence, unwarranted.
92. It follows that Railcare must fail on the Issue of contributory negligence. It follows further that sub-issues (2) and (3) (namely, the financial consequences and apportionment) are academic. Given, however, that these sub-issues were fully argued and that (2) took up much time at the hearing, I shall seek to deal with them, if far more briefly than would otherwise have been the case.
93. The financial consequences: (A)Introduction: On the assumption (contrary to the conclusion to which I have come) that there was negligence on the part of GNER in failing to detect the crack prior to the derailment, then Railcare must prove what loss and damage would have been avoided had GNER not been negligent. The hypothesis here is that the derailment would not have happened. In the event, the following questions arise:

i) What loss and damage would have been avoided had GNER's reaction to the discovery of the crack, absent the derailment, been substantially the same as its reaction when the crack was discovered after the derailment ? ("minimum consequences");

ii) (a) Would GNER's reaction to the discovery of the crack, absent the derailment, have been the same as its reaction following the derailment? (b) If "yes", would that reaction have been unreasonable and, if so, what consequences flow therefrom? (c) If the answer to (a) would or should have been "no", what loss and damage would have been avoided by a hypothetical different reaction ? (so-called, "Scenario C").

94. (B) Minimum Consequences: On the hypothesis that GNER had discovered the crack, that the derailment had been avoided and that GNER's reaction to the discovery of the crack was substantially the same as its reaction following the derailment, there was by the end of the hearing no or no serious dispute as to the financial consequences. The total saving would have been of the order of £300-400,000, by way of saved expenses. Had this point not been academic, I would have invited the parties to attempt to agree the precise figure, alternatively to furnish brief further argument, taking into account my decision on principal Issues (I) and (II). As the point is academic, I do not propose to say anything more about it.
95. (C ) Scenario C: (a) Overview: Much time was devoted at the hearing to Scenario C; namely, the hypothetical response which Railcare submits GNER would and should have taken to the discovery of the crack without the derailment.
96. The Railcare thesis hinges on Mr. Etwell's proposed response to the discovery of the crack without the derailment. It postulates GNER discovering the crack in the course of routine maintenance and reacting to it freed from (what was termed) the "derailment mindset". No criticism is made of GNER's actual response, given the derailment. But freed from the derailment mindset, Mr. Etwell proposed a different, hypothetical, response generating savings of £1 to £1.6 million, depending on whose figures are preferred. Without, it was submitted, any compromises on safety, these savings could have been achieved by, inter alia, reducing the speed not to 80mph but to 100 or 110mph and, in particular, by not undertaking the immediate reaming and chamfering work at the same time as the initial MPI testing. Instead, rolling stock would have been restored to service after visual inspection and MPI testing; thereafter rolling stock could have been brought back in a staged programme, with the repair work carried out over a 14 day period. Such an approach would have reduced the risk that the reaming and chamfering work was not done properly and, importantly, would have restored the fleet to a higher, if not full, speed sooner; hence the significant savings. Given savings of this order, Railcare submitted that GNER would have adopted Mr. Etwell's solution had the crack been detected during a routine inspection. Alternatively, if GNER had nonetheless chosen the more expensive option, it would have been unreasonable for it to have done so and the difference cannot be recovered from Railcare.
97. To my mind, this is an ingenious thesis or submission but one which I have no hesitation in rejecting. As it seems to me, it breaks down in two critical respects. First, I am amply satisfied that GNER would not have adopted Mr. Etwell's solution; in the contemplated hypothetical situation, GNER's reaction would have been much the same as it was following the derailment. Secondly, I do not think that GNER's reaction would have been unreasonable nor, ultimately, did Mr. Etwell in his evidence say that it was. In reality, Scenario C reflected a cleverly constructed possible course of action; doubtless, had Mr. Etwell or anyone else suggested it to GNER, it would have merited consideration. But it goes no further than that. It is tempting to say no more of Scenario C; out of deference, however, to the arguments and evidence at the hearing, I shall set out my reasoning in brief.
98. (b) GNER's likely hypothetical reaction to the discovery of the crack without a derailment: I begin with the context. It is true that the pressures, customer and public relations considerations following a derailment are, as Railcare contended, very different from those facing operating companies when a problem is discovered in the course of routine maintenance. That said, there can be no gainsaying the fact that the discovery of a crack of (say) 22/28cm would have raised significant safety concerns, not least because of ramifications for replication elsewhere in the fleet. To my mind, it is simplistic to characterise such a discovery as little more than a demonstration of successful routine maintenance procedures; instead, even though the crack had not propagated to failure, it was unprecedented for Mark IV wheels and would and should have caused alarm bells to ring.
99. Against this background, I come to the GNER evidence. The decision to impose the 80mph speed limit was taken on the night of the derailment by Mr. Tham. He was concerned for the safety of the fleet. The 80mph speed limit related to a standby timetable which GNER was accustomed to use in the case of bad weather; here, of course, the reduction of speed was not general but was particular to GNER and so there were real disadvantages, which Mr. Tham acknowledged in his evidence. However, as he put it, he would initially have had to "err on the side of safety" until he received technical advice. For my part, I cannot see why Mr. Tham's initial process of reasoning should have been any different had the crack been detected prior to the derailment. It may be that such a reduction in speed was not strictly necessary; whether, however, HMRI would thereafter have agreed to an increase of speed at least until conclusion of MPI testing, must be speculative; HMRI did not agree to an increased speed in the situation which actually prevailed.
100. I turn next to the evidence of Mr. Buxton. He was a consultant from Interfleet, called in to advise following the derailment. On the basis of his advice, the fleet was "stopped" in the early hours of the 17^th June; i.e., trains were directed to proceed to depots. There was some debate at the hearing as to the import of his evidence in respect of the hypothetical situation. To my mind, his evidence was clear. Had a crack of this nature been encountered in the course of "routine" maintenance, Mr. Buxton would have given similar advice. The concerns of which he spoke in relation to the actual situation, namely, the risk of cracks in other wheels and not knowing their rate of propagation, are equally applicable to the hypothetical situation. All that remained was some debate as to whether, in the hypothetical situation, the fleet would have been stopped at once, say in the early morning if that was when the crack was discovered, or at the end of the day. My impression is that Mr. Buxton favoured giving such advice as soon as possible; to the extent that it matters, I do, however, regard it as unreal to enter into conjecture as to when precisely in the course of a day such advice would have been given or executed.
101. Once the trains were in the depots, visual inspections would have taken place. Some time would have elapsed before MPI procedures were organised. Following visual inspections, the fleet may have been returned to service but, for the reasons already given, if so, probably at a reduced speed.
102. I come next to the nub of the argument on this part of Scenario C. It was a key constituent of Mr. Etwell's reasoning that following MPI testing, trains would have been returned to service and then subsequently recalled for rectification work. While such a course of action was not unfeasible and, as Mr. Tham fairly accepted, would have been considered if suggested, I am wholly unable to accept that that is what would probably have happened in the hypothetical situation. It was contrary to both Mr. Lumb's engineering instinct and Mr. Tham's commercial instincts. To Mr. Lumb, it contravened his "gut feeling" to put the trains back in service without undertaking remedial work; that feeling would have been the same in the hypothetical situation. Mr. Lumb was responsible for GNER's safety case and therefore the decision was his. For my part, even if, with the benefit of later objective consideration it can be seen that safety is not compromised by undertaking rectification work later rather than immediately following (successful) MPI testing, I can well understand Mr. Lumb's visceral reaction; it is only necessary to contemplate what Mr. Lumb's position would have been had a train been restored to service without remedial work and another casualty had been suffered because of some slip-up in MPI testing. So far as concerns Mr. Tham, his view was that the approach adopted in the actual situation was the least of all evils. He could not run a 125mph timetable until (essentially) all the fleet was capable of doing so. A short sharp attack on the problem was preferable to having the issue drawn out over a period of weeks. He did not see Mr. Etwell's proposal having advantages in terms of disruption. Additionally, it was to be kept in mind that the MPI testing and rectification work were done by separate teams; MPI testing in fact took place throughout much of the period from the 18^th June to the 21^st/22^nd June; in short, no time would have been saved by not undertaking the remedial works during the same period. Pulling the threads together, in GNER's written closing submissions, the matter was put this way:

" The idea that either Mr. Lumb or Mr. Tham would have done MPI testing, without the initial rectification work, is fanciful. Given the opportunity to do the work, with the wheels jacked up, balance weights removed and paint stripped off, and faced with obvious stress raisers or initiators in the form of burrs and lack of chamfering, it is quite clear that GNER would have done exactly what in fact they did at the time."

I agree.

103. (b) Was GNER's likely hypothetical reaction unreasonable ? Intuitively, this did not suggest itself as a promising line of attack for Railcare. Mr. Etwell himself accepted both (1) that from a safety or technical perspective there was "very little difference" between the actual and hypothetical situations; and (2) that no criticisms could properly be made of GNER's response to the actual situation following the derailment. Why then, it might be asked, would broadly the same response be unreasonable in the hypothetical situation ?
104. Ultimately, in my judgment, Mr. Etwell was unwilling to suggest that it was. His views appear from a somewhat extended passage in his cross-examination:

"Q. But you accept that it was an entirely reasonable thing for them to have done what they did at the time ?

A. I accept it was reasonable for them, but I accept that I myself would not have done that.

Q. Forgive me, Mr. Etwell. It may be that we are at cross purposes. I want to make it clear. It may be that you would have done things in a different way and you have made that clear but you are here giving evidence … in support of a case that if GNER had acted in a particular way, that would have been unreasonable for them to have done so.

So… the issue is not how you would have done it but if you are prepared to say the way in which GNER say they would have approached this matter would have been beyond the bounds of acceptable professional conduct, would have been unreasonable.

…Are you saying – can we cut through this – that if in fact GNER had decided to deburr and chamfer at the same time as MPI testing on scenario C, are you saying that it would have been an unreasonable thing for them to have done?

A. I would have considered it not unreasonable but a risk. Because, I repeat, at that point in time there is a high likelihood that there would be a second stage necessary because to do the rectification immediately, you are in a risk where experts will come back and redefine their standards.

Q. Yes. Not unreasonable but a risk of having to improve the quality of the work that you are doing at short notice when you have the wheelsets there being MPI tested, is that how you would put it ?

A. It is not unreasonable to do an immediate ream and chamfer to recondition the holes to acceptable state. It is not unreasonable.

Q. No, that is very fair and it must be right, must it not? If you see a burr, as you have the wheelset jacked up and you strip the paint off, and you know that a burr is a stress raiser or a stress initiator, to say to GNER it is unreasonable to remove that burr would be very hard to say, would it not ?

A. I repeat, it is not unreasonable but it is also a risk and I would be aware of a risk of experts coming back in time to say we have rethought and we think there is additional work that you need to undertake.

Q. Yes, but you would not criticise them for doing it?

A. No, I would not. "

105. In re-examination and in final submissions, Mr. Popplewell made, if I may say so, a valiant attempt to suggest that these answers related only to possible engineering solutions; by contrast, an earlier answer properly expressed Mr. Etwell's opinion that, doing the rectification work during the same period as the MPI testing, was unreasonable from an operational point of view. With respect, such contentions are untenable. The earlier answer relied upon almost immediately preceded the passage set out above. In the above passage, there is no hint that Mr. Etwell was confining his views to engineering as distinct from operational matters. Having seen and heard Mr. Etwell give evidence, I am satisfied that his views generally are as expressed in the above passage. As such, they are fatal to Railcare's charge that if GNER had reacted to the hypothetical solution in much the same way as it did to the derailment, that would have been an unreasonable course to adopt. Here too, Railcare has needed Mr. Etwell to go further and, ultimately, he was not prepared to do so. Perhaps, he never intended to. In his Second Supplementary Report, all that Mr. Etwell said was this:

" I would comment (in case it is not clear from my first report) that I would not myself have found it necessary for GNER to carry out deburring and chamfering while doing the initial MPI examination after the derailment."

106. I accordingly conclude that had GNER's reaction in the hypothetical situation been much the same as its reaction following the derailment, such a course of action would not have been unreasonable.
107. For completeness, various other matters were canvassed in evidence and argument as to Mr. Etwell's hypothetical solution, such as timetables, turnaround times and available train sets. In the light of the conclusions to which I have come it is unnecessary to take up time with them.
108. (c) Loss and damage avoided had GNER followed Mr. Etwell's proposed solution ? No useful purpose would be served by exploring this question further. As already indicated, the figures canvassed ranged from some £1 to £1.6 million, depending, as I understood it, on the outcome of Issue (I) and a measure of disagreement between Mr. Segal and Dr. Harris.
109. Apportionment: There was some debate at the hearing as to the correct approach to apportionment under s.1(1) of the Law Reform (Contributory Negligence) Act 1945 (in the event that contributory negligence had been established). The argument was prompted by the fact that even had GNER detected the crack prior to the derailment, very substantial losses would in any event have been caused by Railcare's (admitted) negligence.
110. The matter may be illustrated as follows. Let the loss which GNER would in any event have sustained be termed X; let the additional loss which GNER would have avoided had the crack been detected and had there been no derailment, be termed Y. The argument then ranged over whether the apportionment should take into account two "pools" (X and Y) or whether there was only one pool (X plus Y). There was further argument as to whether GNER's recoverable damages should be reduced by the whole of Y.
111. Given the wholly academic nature of this debate in the light of the conclusions to which I have already come, I propose to say no more than this. First, it would not have been right that GNER should have its otherwise recoverable damages reduced by the whole of Y. That, in effect, would be to treat GNER as solely to blame for the derailment and, hence, Y; at most, to my mind, GNER would have been contributorily negligent; Railcare's negligence remained an operative cause of the derailment. Secondly, it seemed to me unnecessary and over-theoretical to decide whether there were two "pools" or only one pool. On no view could it be right to reduce GNER's recovery of X. However approached, the apportionment could accordingly never reduce GNER's entitlement to damages by more than a proportion of Y. Whether such apportionment is expressed as a fraction of Y or as a fraction of X+Y, matters not. Thirdly, having reached the conclusion that GNER was not (contributorily) negligent, it is invidious to express a final view as to the apportionment I would have made had that conclusion been different. I prefer to go no further than to say that my inclination would have been to reduce GNER's recovery of the £300-£400,000 expenses (which would have been avoided had there been no derailment) by something of the order of 20 - 25%. Provisionally (I express no final view), such an apportionment would have struck me as just and equitable, having regard both to causative potency and blameworthiness. For completeness in this regard, it will be recollected that having in any event rejected Railcare's case on Scenario C, no question would have arisen of savings in the order of £1 - £1.6 million.

POSTSCRIPT

112. I would not wish to leave this case without expressing my gratitude for the exemplary assistance received from counsel for both parties and their respective teams.
113. I shall seek the assistance of counsel in drawing up the appropriate order and with regard to all questions of costs.