Barrow (by his litigation friend and grandfather Hugh Barrow) & Ors. v Merrett & Anor.

Felix Barrow (‘A’) was seriously injured on 7 October 2015 when he was crossing the road outside his house. A car driven by Mrs Rosemary Merrett (‘R’) collided with him. He made a claim for damages for negligence. After a hearing lasting five days (1-5 March 2021), Richard Hermer QC, sitting as a Deputy Judge of the High Court (‘the Judge’), dismissed that claim, in a reserved judgment which he handed down on 31 March 2021. As the Judge said, at the start of his judgment, there was ‘no dispute that what occurred was a tragedy with lasting impact on many involved, but above all on Felix who sustained life-changing injuries’. This is an appeal from the Judge’s decision. Stuart-Smith LJ gave permission to appeal. He observed that the grounds of appeal were ‘just arguable’.

The trial was recorded on Livenote. The parties and the Judge were provided with a transcript of each day’s hearing.

A was represented on the appeal by Mr Weir QC and Mr Burton. R was represented by Mr O’Sullivan QC. All counsel, apart from Mr Weir, appeared below. Ms Susan Rodway QC led Mr Burton for A at the trial. I thank counsel for their written and oral submissions.

There are three grounds of appeal.

Paragraph references are to paragraphs in the Judge’s judgment, or, to the paragraphs in the parties’ skeleton arguments, or in the authorities, as the case may be, unless I say otherwise.

The grounds of appeal give a flavour of the detailed and unsparing attack on the Judge’s reasoning, and on his approach to the trial, which Mr Weir developed in his skeleton argument and in his oral submissions. I make no apology for the fact that I have summarised his judgment in considerably more detail than should be necessary on an appeal.

As the Judge explained (paragraph 11) he divided the substance of the judgment into seven parts: the factual background, the procedural history, the law, an analysis of the approach he should take to the evidence, a summary of the lay and of the expert evidence, and his findings of fact and conclusions.

The Judge noted in his introduction (paragraph 3) that although the parties relied on a good deal of expert evidence, it was soon clear that the ‘core dispute’ turned on ‘two closely related core questions of fact’.

He added that ‘the parties agree[d]’ that the answers to those two questions could unlock the dispute about who was responsible for the accident (paragraph 4). In paragraphs 5 and 6, he described the parties’ answers to those questions.

In paragraphs 7-11 the Judge explained how difficult it is for litigants and the courts to find out what has happened from evidence about fast-moving and traumatic events. In many cases, ‘it is simply not possible to conclude with absolute precision what occurred’. The court did not have to be certain, but had to come to a ‘reasoned view as to the most probable explanation’. In many accidents, there was a range of ‘confounding factors’ which make a ‘precise reconstruction of events impossible’. This was such a case. The event had only lasted a few seconds. It was not recorded in any way. The few eye witnesses saw it from different positions. There was little ‘hard evidence’, such as ‘extensive damage to the car that would enable ready reconstruction’. A’s injuries did not give clear answers to the core questions, nor, for reasons which the Judge was to explain, did the evidence of the accident reconstruction experts (paragraph 8).

In order to find the ‘most probable’ answers to the core questions, the court had to look at the available evidence as a whole. The court had to understand the layout of the scene, and to ‘identify any objective facts which might act as lodestars by which more subjective opinion and recollection can be tested’. The court had to scrutinise the evidence of the witnesses of fact and of the experts, in court and in their written statement. The court also had to apply ‘a fair dose of common sense’ (paragraph 9).

The background facts were that just before 8am on 7 October 2015, A who was 11 years old, had set off to walk to school with his best friend, Nicholas Stannard. He had just started at secondary school. His parents, who had ‘instilled into him a need always to take care when crossing roads’ now allowed him to walk to school with his friends. A lived on a long road with a single carriageway. Near his house the road is called Hill Pound Road, and there is only a pavement on the side of the road opposite his house. The speed limit is 30mph and is clearly marked. A little further south, the speed limit is 40 mph. Nicholas Stannard had been dropped at A’s house that morning. They set off from A’s house by crossing the road to the pavement on the opposite side. When they got to the pavement, A told Nicholas Stannard that he had forgotten his rugby boots, and that he needed to go home to get them. He re-crossed the road while Nicholas Stannard waited for him on the pavement.

As A came back, R was driving north along the road towards A’s house. The pavement was on the left of her car. She noticed Nicholas Stannard on the pavement as she got closer to A’s house. She also saw traffic coming towards her on the opposite side of the road. What happened next was in dispute, but it was agreed that at about the point when R drew level with Nicholas Stannard, her car hit A on the left side of his body, just over the centre line of the road, that is, just on the side of the road on which R was driving (paragraph 15). The impact threw A’s body about 8 metres to the north (that is, in the same direction as R’s car was travelling), and about 3 metres away from the offside of the car (paragraph 16).

Nicholas Stannard told A’s parents what had happened. A was unconscious in the road. They called the emergency services. The police closed the road and started an investigation, led by PC Giles. A was evacuated to hospital by air ambulance. A has been left with ‘the lifelong consequences of the severe physiological and neurological injuries he sustained’ (paragraph 19).

R’s car was not very much damaged. There was some dispute about the cause of some of that damage. There were only three areas of minor damage on the front offside of the car, which the Judge described. Two photographs in the judgment showed that damage (figures 1 and 2). Figure 3 was a diagram by Mr Johnston showing where the front and offside of the car were damaged (paragraph 20).

In paragraph 23, the Judge summarised A’s pleaded claim. It was that as A was walking across the road towards Nicholas Stannard, he slipped. As he was getting to his feet, he was hit by R’s car. The allegations of negligence really amounted to ‘an assertion that [A’s] position in the road should have been obvious such that a reasonable driver would have seen him in time to avoid a collision’. The defence was that a reasonable driver could not have avoided the collision because A ran into the road when R’s vision ‘was largely obscured by oncoming traffic, giving insufficient time to brake’. There was also a plea of contributory negligence.

In paragraphs 27-30 of the judgment, the Judge summarised the relevant law. He commented that the assessment of liability for road traffic accidents turns on the facts in the vast majority of cases, so that this is ‘not therefore an area rich in legal learning’. He added that, ‘Important as it is to keep this standard in mind, it is of less practical application in this case where the parties are essentially agreed that dependent on the findings I reach in respect of the two core questions, the driving was either reasonable or unreasonable’. Mr Weir complains that the Judge misdirected himself by not stating specifically that there is a high burden on a driver to reflect the fact that a car is ‘potentially a dangerous weapon’. He argues that this ‘failure to recognise this standard of care is of a piece with the Judge’s overly benevolent approach to the evidence of [R]’.

The Judge then considered the parties’ arguments about how the evidence in the case should be approached (paragraphs 31-38). A relied on Gestmin. The Judge noted the observation of Leggatt J (as he then was) that ‘in complex commercial claims, the existence of substantial amounts of contemporaneous documentation will often provide a more reliable source of evidence than the recollection of witnesses proffered in a courtroom many years later’. The Judge recorded A’s submission that he should apply Gestmin by analogy, and, because human memory is fragile, treat the evidence of the witnesses as a secondary source, as much more objective, ‘harder’ evidence was available from the experts; in particular, the experts in accident reconstruction. He should ‘place little, if any, reliance at all on the witnesses’ recollection’ (paragraph 32).

The Judge accepted that the objective evidence ‘is always an extremely helpful source both in itself, and as a guide to calibrating the recollection of the witnesses’. He rejected any wider submission, in the context of case like the present, for two reasons.

He accepted that Leggatt J had expressed ‘insightful reflections on the fragility of human memory’ which were ‘a beacon to any court’ trying a case in which there is conflicting evidence from witnesses. He cited four of Leggatt J’s insights (paragraph 36). They reflected ‘what courts have long known’, which includes that accounts given at the time are more likely to be reliable than accounts given later. An honest witness may give wholly inaccurate evidence because ‘their memory may have been subconsciously degraded not just by time but a range of biases’.

Those insights underlined the caution which should be attached to evidence from the witness box, or in statements generated for the purposes of litigation, which deals with events years beforehand, ‘all the more so when the events were highly traumatic and last only a few seconds’. Gestmin reminded the court that ‘often (but not always) accounts given at the scene will be more reliable than versions given some time later as part of the litigation’. That was particularly relevant to the Judge’s assessment of the evidence, ‘including the evidence of Nicholas Stannard, the only witness to the accident itself called by [A]’ (paragraph 38).

The Judge summarised the evidence in paragraphs 39-71. Three people saw the accident; Nicholas Stannard, R, and Alexander Gent, a neighbour of A’s family. Nicholas Stannard gave evidence for A. R gave evidence in her defence, and called Alexander Gent, PC Robert Giles and PC Stephanie Wheeler. PC Giles was in charge of the investigation. PC Wheeler helped him.

The Judge recorded A’s concession that if critical elements of Nicholas Stannard’s evidence were found to be unreliable, the claim must fail, and the R’s concession that, if the core of his evidence (that A fell and tried to get up again) was ‘held to be probable’, the claim would succeed. That made Nicholas Stannard’s evidence ‘central’ (paragraph 42). Nicholas Stannard was recorded as giving ‘at least five accounts’ (paragraph 43). The Judge summarised those accounts in paragraphs 44-49.

The first account was recorded by PC Giles in his Incident Log on the morning of the accident. His evidence was that he recorded the account at the time of his conversation with Nicholas Stannard and his mother. Nicholas Stannard said that he had left A’s house with A. ‘Crossed road. He forgot something so came back to H/A. Came back to road. Nicholas said ‘WAIT THERE’S A CAR’. Then ran across road. Slipped and hit car’. The Judge said that the accuracy of that record was ‘a significant issue at trial’. He added, with some understatement, that ‘The suggestion that [A] ran across the road notwithstanding a warning, that he slipped and then hit the car is difficult to fully reconcile with [A’s] pleaded case’.

The second account (‘albeit in second-hand hearsay form’) was recorded in computer records which were disclosed ‘during the course of the trial’, but which reflected evidence already given in the witness statement of PC Giles. On 30 November 2015, Mr Barrow, A’s grandfather, contacted PC Giles. He had spoken to Nicholas Stannard and to his mother. Nicholas Stannard now wanted to say that A was walking across the road and that he was getting up before he was hit by the car. In his notes, PC Giles said that this was a ‘stark contrast’ with the first account. Mr Barrow had also told PC Giles that the family were considering a civil claim which would help to pay for A’s care (paragraph 46).

On 26 November 2018, just over three years after the accident, Nicholas Stannard gave his third account, in a witness statement for this case. He described how A had returned to the edge of the road after going home, had stopped at the edge of the road, checked both ways and after letting two cars pass had started to walk across the road at a normal pace. He had then slipped on a shiny patch of road, and fallen back onto his bottom. He was hit by the car when he tried to get up (paragraph 47).

Nicholas Stannard’s fourth account was a supplemental statement dated 24 February 2020. He addressed a suggestion by R’s accident reconstruction expert that A was falling forwards when he was hit by the car. Nicholas Stannard was certain that A was not falling forwards, but that he was getting up when he was hit. He was filmed showing how A had been moving when he was hit. I have watched that film, which is in two versions (one at normal speed, and the other slowed down).

Nicholas Stannard gave evidence at trial and was cross-examined. That was his fifth account. He ‘strongly disputed’ having told PC Giles, shortly after the accident, anything other than his name and address. He had not told him any details of the accident. He did, however, accept that some of the information which PC Giles had recorded and attributed to him was correct: A had gone back home, had returned to the road, and he had warned A about cars. He denied telling PC Giles that A ran, slipped, and hit the car (paragraph 49). The rest of his evidence was broadly consistent with what he had said in his third account (paragraph 50).

The Judge summarised the evidence of Nicholas Stannard’s mother in paragraphs 51-54. Her October 2018 witness statement described what Nicholas Stannard said to her just after the accident. He told her that A was not running. He had not spoken to PC Giles until she was present. PC Giles asked whether A had been running and Nicholas Stannard said that he had not. In cross-examination she added that she had told PC Giles some of what Nicholas Stannard had told her: A had slipped while walking back from his home. She suggested that she, rather than her son, was the source of some of the information in the notebook. She denied having told PC Giles that Nicholas Stannard had said that A was running, slipped, and hit the car. Nicholas Stannard had always given a consistent account. She was asked in cross-examination why her statement had not mentioned Nicholas Stannard’s warning to A. The Judge quoted her answer ‘I don’t think you quite understand how traumatic that day was for all of us and how hard it has been to recreate it on so many occasions’ (paragraph 54).

The next topic in the judgment is the evidence of PC Giles (paragraphs 55-60). His evidence came first, but it was also convenient to deal with it first because it was ‘directly relevant’ to the evidence of Nicholas Stannard and of his mother (paragraph 55). He provided logbook entries ‘generated at the scene’ and contemporaneous computer records of further steps in the investigation (paragraph 56). He arrived at the scene about 30 minutes after the accident. He was an experienced officer. When he learnt how serious A’s injuries were, he created a ‘scene log’ in case it later emerged that a criminal offence had been committed. He spoke first to R, who told him that A had run directly into her car without warning. His evidence was that shortly after 9am he spoke to Nicholas Stannard and his mother, leading to the entry in the incident log (paragraph 57).

He was extensively cross-examined about the accuracy of the entry recording his conversation with Nicholas Stannard and his mother. As the Judge observed, ‘The amount of time spent on these few lines may be thought a reflection of their potential materiality to the core dispute (paragraph 58). The Judge summarised the themes of that cross-examination in paragraph 59. The Judge, in a passage to which Mr Weir objected, said ‘PC Giles was having none of it’. The Judge then summarised the responses which PC Giles gave to those lines of cross-examination, ending with: ‘He maintained that the account given by Nicholas of [A] running into the road and slipping was very clear, and to his mind, very relevant’ (paragraph 60).

In paragraph 61, the Judge recorded PC Wheeler’s nearly contemporaneous record in her notebook of R’s account; ‘…from nowhere boy appeared from behind dark car running to other side of road from o/s’. She confirmed that the road was slightly damp and that R was very upset.

The Judge considered R’s evidence in paragraphs 62-64. In her witness statement she described driving along the road at about 30mph. She noticed Nicholas Stannard standing on the pavement on her nearside. As she drove towards him, another vehicle was driving towards her on the opposite side of the road. As she drew level with the back of that car, a child ran from her right into her car. She assumed he was running because he was in a crouched position. She had no chance to avoid the collision (paragraph 62). In cross-examination, she was clear that she could remember some things, but not others. She knew the road very well. There was nothing unusual about children waiting on the pavement for their friends (paragraph 63). She was clear in her recollection that she was concentrating on the road ahead and that she only saw A ‘a split second before the collision’. When she saw him, he was ‘leaning forward as if he was running’ (paragraph 64).

The last witness called by R was Alexander Gent. His statement was dated 24 March 2016. There was also a copy of the statement he gave to the police the day after the accident (paragraph 65). He had joined the road a short distance to the north of A’s house. He was driving in the opposite direction from R, that is, south (paragraph 66). In his police statement he described seeing Nicholas Stannard shortly after he joined the road and while he was accelerating. After a grey car ahead of him had passed a driveway, a boy ran from his left towards the middle of the road, where he was hit by R’s car. The statement said ‘The boy ran straight out into the path of the red car. There was nothing the red car could have done to avoid hitting him (paragraph 67). The statement in the proceedings was shorter than the police statement but consistent with it (paragraph 68).

He was cross-examined ‘in some detail’. He denied that there was ‘no love lost’ between his family and A’s family. He explained that as far as knew there had been one dispute, between his parents and A’s parents, many years ago, about some trees. It was not suggested that he was not an independent witness (paragraph 69). The Judge described Alexander Gent’s cross-examination in paragraph 70. He maintained his evidence that he saw A run out from his drive, immediately behind the car in front of him, and into the oncoming car driven by R (paragraph 70).

The judge summarised the expert medical evidence in paragraphs 72-78. He was able to do so shortly because there was considerable agreement about A’s injuries and the ‘very limited extent to which conclusions could be drawn on body position at the time of impact including whether he was running, walking, slipping, or in the process of getting up’ (paragraph 72).

R’s expert did not give evidence; Dr Hulse, A’s expert, did. Little weight could be given to the reports of R’s expert, but the Judge did receive memoranda of two meetings between the experts (paragraph 73).

The Judge described A’s injuries in paragraph 74. They were mostly to the left side of his body. He suffered a right frontal fracture to his skull, almost at the midline. A reader of the reports might think that their authors were offering opinions on A’s body position on impact, and deriving conclusions from those about what A was doing when he was hit. It soon became clear, when Dr Hulse was cross-examined, that (whatever the reports might suggest) he was not doing that. He said that A could not have suffered those injuries if he had been standing up straight or lying down when he was hit; but that he could go no further than that. The Judge quoted passages from Dr Hulse’s cross-examination in support of that view (paragraph 75).

All Dr Hulse could say was that A must have been in front of the car. The illustrations at figure 4 exemplified, but did not define, the range of positions A might have been in on impact (paragraph 76). Dr Hulse’s own drawing, from his first report, was not intended to show the precise position of A’s limbs on impact, but rather the ‘likely level of [A’s] head and torso relative to the front of the car.’ It was not designed to show how far across the car A was when he was hit. There was little difference between the positions of A’s head and torso in the two images. A’s actual body position could have been closer to one or to the other, or to neither. At the end of Dr Hulse’s evidence, R’s counsel decided not to call his own expert.

In paragraphs 79-100, the Judge considered the evidence of the parties’ accident reconstruction experts. Mr Sorton was A’s expert. Mr Johnston was R’s. Each wrote two reports and there were two joint statements.

The Judge summarised the four reports in paragraphs 80-83. In the first joint statement they agreed that there was ‘sliding contact’ between A and the offside of R’s car. At the second meeting they considered some modelling by Mr Johnston. They agreed that the modelling depended on various assumptions. It depended ‘crucially’ on the extent to which most of A’s mass was positioned across the front of the car, and on his precise posture (paragraph 86).

The Judge observed that Mr Sorton’s evidence began badly. Despite having been in court throughout the trial, he introduced, with no advance warning, a significant change of opinion in his evidence in chief. He had seen digital images two days previously which led him to retract his agreement that the scuff mark reflected direct contact with A. He did not explain why he had kept this thought to himself, and had not let R know. The Judge explained why this was bad practice, and unfair. Mr Sorton apologised and R was able to adjust to this change of position (paragraph 88).

The Judge described Mr Sorton’s explanation for his change of view in paragraph 89. He did not consider that close examination supported Mr Sorton’s view, but it was not necessary to decide what caused the mark. The Judge explained Mr Sorton’s mistake in paragraph 90.

In cross-examination Mr Sorton said that whether A was running or not depended on where the car first made contact with A. If the contact had been towards the centre of the car, A would not have ended up where he did. If the contact was with the front offside, A could have been running. He also agreed that, if that was so, his change of evidence was ‘likely misplaced’ (paragraph 91). The Judge said that there was a ‘degree of unreality’ about this aspect of the evidence because, as Mr Sorton acknowledged, there was no ‘hard’ evidence actually capable of proving to where the first point of bodily contact was, for example, obvious damage to the car such as an imprint of a body (paragraph 92).

It was clear that Mr Sorton was sceptical that either of the images (in figure 4) showed what a fall forward from a run would look like ‘in the “real world” ’, and about R’s description of ‘slipping forward and the likelihood of doing so in a crouching position’. He nevertheless acknowledged that ‘he could not provide expertise on “how people fall”’ (paragraph 94). In paragraph 95 the Judge quoted Mr Sorton’s summary of his position towards the end of his evidence. The issue was less whether A was running or walking, ‘but really turned on “how he fell” not least because that determined the extent to which [R] could reasonably have avoided the accident’.

Mr Johnston confirmed his earlier view that the damage to the offside of the car had been caused by contact with A’s body. He did not agree with Mr Sorton’s reasons for changing his view (paragraph 96). The Judge recorded that A’s line in cross-examination was to suggest that Mr Johnston was partisan. Mr Johnston accepted that he had not initially considered a theory that A had fallen forwards when he was running. He had not had a ‘lightbulb moment’ but the possibility of a slip forward was ‘something that emerged from discussion and thought’. Mr Johnston did not accept that the assumptions he used in his modelling were an attempt to ‘reverse engineer’ a solution for R’s benefit. He had used them primarily to rebut Mr Sorton’s theory that A could not have been running. He denied that ‘there was really any “new” wrap around theory but rather this was all consistent with what he and Mr Sorton had previously agreed, namely that [A’s] body, having been hit by the front offside of the car, would have been rotated so as to also make contact with the offside itself’ (paragraphs 97 and 98).

The Judge did not consider that this was a case in which the accident reconstruction evidence of the experts helped the court very much. There was useful information in the reports about stopping distances, and useful plans and photographs. The reason was that ‘their analysis …[depended] on a range of unverifiable assumptions (eg where on the car [A] was first struck) and was also dependent in large measure o[n] the competing narratives emerging from the witness evidence (eg was he slipping forward or getting up?)’ (paragraph 99).

Importantly (in my judgment) the Judge said, ‘There was nothing in the evidence of either expert that demonstrated that the contentions of one side, or the core evidence of a given witness, was incontrovertibly wrong’. The Judge acknowledged that there was very little for the experts to use: there were very few, minor, marks on the car, and they could not support firm conclusions about where A was hit, or how his body was projected by the impacts. The Judge made a similar observation about the ‘equivocal’ medical evidence. It helped to show where A’s head and torso were at the point of impact, but it did not ‘permit an expert assessment as to whether it is consistent only with a slip forward whilst running, or a return to the feet from a fall to the bottom, or some other mechanism’ (also paragraph 99).

The Judge ended this analysis, in paragraph 100, with a quotation from paragraph 10 of the judgment of Coulson J (as he then was) in Stewart v Glaze [2009] EWHC 704 QB. Coulson J had said that it was ‘the primary factual evidence’ which was most important in a case like the present. The expert evidence was a useful way of testing the factual evidence and the inferences which could be drawn from it, but was not a ‘fixed framework’ or formula for ‘rigidly judging’ the defendant’s actions ‘with mathematical precision’.

In paragraphs 101-115, the Judge made his findings of fact and expressed his conclusions. He decided that, most probably, A ran across the road back towards Nicholas Stannard into the path of the oncoming traffic. The Judge said that this was just after a car passed A ‘from his left thereby obscuring him from [R’s] vision’. A probably slipped and his body fell into the path of R’s car. R had ‘no realistic opportunity of avoiding the collision’.

R accepts that the Judge’s reference to ‘the left’ is a slip, and that the Judge must have meant ‘the right’. It is unfortunate that this slip was not picked up when counsel commented on the draft judgment, but I do not consider that it is a slip which undermines the Judge’s reasoning in any way. Indeed, the very fact that counsel did not pick it up when they read the draft judgment suggests to me that neither considered that it was significant. I did not understand A to submit otherwise.

The Judge explained that he based this finding ‘primarily on my assessment of the lay witness evidence …in the context of the road layout’.

In paragraph 103, he acknowledged that it was possible, but unlikely, that R would have driven along the stretch of road, when the road ahead was easily visible, without seeing A in the middle of the road if he had walked across the road or even if he had run, fallen, and tried to get up in the way that Nicholas Stannard described. A’s counsel had estimated, by reference to the film of Nicholas Stannard (see paragraph 27 above) that it would have taken A about ten seconds to get up from the ground to the point of impact. Even if it had taken much less time, A would still have been an obvious obstacle long before he was obscured by the car which was in front of Alexander Gent and approaching from the opposite direction.

The Judge acknowledged that R could have been distracted, or have been driving dangerously and not paying proper attention to the road ahead (paragraph 104). Indeed, had the only evidence been that of R, it would have been ‘much more difficult’ to make findings of fact. The Judge explained that the ‘key reason’ why he found the evidence of Nicholas Stannard ‘unreliable’ was the ‘corroboration provided by Mr Gent’. If Nicholas Stannard’s account was correct, then Mr Gent ‘must also simultaneously have overlooked an obvious obstruction on the road in front of him whilst driving in the opposite carriageway’. He must also ‘have imagined that he saw [A] running out of the drive’. The Judge considered that it was improbable that A could have been on the ground in the middle of the road for several seconds without R or Alexander Gent noticing him. It was also ‘improbable’ that Alexander Gent was mistaken in his recollection, given to the police the next day, of seeing A run into the road. Moreover, that recollection ‘As set out below’ was consistent with that given to the police by Nicholas Stannard on the day. This is a reference to the Judge’s assessment of that account in paragraph 107 of the judgment. The Judge continued, in paragraph 104, ‘This is a case, like many before it, in which the evidence given immediately after the event tends to be a more reliable source than those created subsequently, particularly when litigation is in contemplation’.

Alexander Gent was not only ‘independent’, but ‘impressive’. He was clear about what he could and could not remember, and ‘acknowledged (credibly) matters that he could no longer be sure of’. His evidence was broadly consistent with what he told the police the day after the accident. ‘He clearly had an excellent angle in which to observe both [A] running into the road and the collision with [R]’ (paragraph 105). The Judge acknowledged that not every aspect of Alexander Gent’s evidence was ‘accurate or comprehensive’ (paragraph 106). He did not remember A slipping, and ‘his description of the mechanism of how his body hit the bonnet, might not fit precisely with the damages or injuries identified by the experts’. This is another understatement by the Judge. The Judge considered that those were ‘hardly fatal flaws in the credibility of a witness to a split second collision’. It was plausible that he did not see a momentary slip, and that ‘he would not be able to accurately recall precisely how [‘A’s] body was propelled in a collision lasting only fractions of a second’.

In paragraph 107, the Judge considered what could be derived from the account which Nicholas Stannard was said to have given to PC Giles just after the accident. It did not matter whether that account was given by Nicholas Stannard, or by his mother in his presence. The point was that the source of the information could only have been Nicholas Stannard. The Judge concluded that Nicholas Stannard (either directly, or through his mother in his presence) told PC Giles that he warned A about the cars, and that A ran into the road and then slipped before being hit. That account did not include any description of falling onto his bottom and being hit while he was trying to get back up again. The Judge accepted that he had to be cautious about an initial account given by an eleven year old child who had just seen a horrifying accident, but ‘the clear evidence of PC Giles satisfied me that the record he made was accurate and was not tainted by any of the criticisms of it levelled by [A]’.

The Judge acknowledged that Nicholas Stannard and his mother disputed the accuracy of part of PC Giles’ note, and that Nicholas Stannard had later given a ‘significantly different account’. The Judge referred to Nicholas Stannard’s mother’s account to A’s family in November 2015, which was then conveyed to PC Giles by A’s grandfather. In that account, A was not running, but walking across the road, and was hit when he was trying to get up. That account was ‘broadly consistent with’ Nicholas Stannard’s witness statement and his evidence at trial. The Judge was ‘satisfied however that the evidence was mistaken in so far as it relates to [A] walking rather than running and also in so far as it depicts [A] slipping as described …and falling onto (or towards) his bottom and then attempting to get up over a number of seconds’. That evidence was inconsistent with the evidence of Alexander Gent, and of R, and with Nicholas Stannard’s initial statement, which, the Judge found, were ‘more reliable sources’ (paragraph 108).

The Judge was careful to say that he was not finding that Nicholas Stannard was lying. The Judge’s impression was that Nicholas Stannard was trying to do his best to remember a ‘truly harrowing experience’. He had been asked to give his account several times and ‘would have been under no doubt of the significance of it to [A’s] case’. The disparity was best explained by the factors identified in Gestmin as ‘capable of degrading the quality of recall’. He also found that Nicholas Stannard’s mother was ‘an entirely honest witness’ but that her recollection of what Nicholas Stannard told her was less likely to be accurate than PC Giles’s contemporaneous record. This was a case, like many others, in which honest witnesses had given conflicting accounts of an event (paragraph 109).

In paragraph 110 the Judge described other factors which made Nicholas Stannard’s recollection less reliable than R’s evidence. The Judge considered that Nicholas Stannard’s account of A’s slip and fall did not ring true, and, at least, was less likely than a slip while running. Nicholas Stannard described a slippery patch in the middle of the road, but no such patch was found by the police, or was visible in the photographs. PC Giles said the surface was damp but drying. It was difficult to understand why a person walking across the road would slip in the way Nicholas Stannard had described. The Judge considered that a person was much more likely to slip while running ‘and/or when appreciating in a split second that they are in danger of colliding with an oncoming car’ (paragraph 110).

The Judge did not consider it likely that if A was in the road for more than a few seconds that Nicholas Stannard would have failed to warn him, or the approaching traffic, of the risk of a collision. Nicholas Stannard was only 11 at the time, and the events took a few seconds. But Nicholas Stannard’s evidence was that when A fell, he was already aware of the approach of R’s car. If A was in the road, struggling to his feet, the Judge considered that it was likely that Nicholas Stannard would have warned A, or the approaching cars (paragraph 112).

The Judge had not overlooked the argument (‘advanced with vigour’) that A’s likely body position on impact was ‘simply incompatible’ with a slip or momentary fall forward and was ‘equally irreconcilable’ with R’s evidence that A was ‘crouched’. It was also argued that this was an inherently improbable explanation because R had advanced it ‘relatively late in the day’ (paragraph 112).

He did not consider that those points negated a slip or fall forward in the ‘split second’ before impact. He gave three reasons in paragraph 113.

The Judge said (paragraph 114) that he had been able to reach those conclusions without the help of the accident reconstruction evidence. His conclusions, nevertheless, fell ‘within the bounds of what both experts considered possible’. Mr Sorton, accepted that, depending on which bit of the car hit him, A could have been running. He could not say which bit of the car hit A. He acknowledged that ‘the question of body position would primarily be derived from whether the Court thought the position he was in was anatomically consistent with a slip forward’. Mr Sorton thought that a forward slip was an unlikely explanation, but he accepted that it was not a question for the experts. Mr Johnston supported the theory that A had slipped forwards.

In those circumstances, the Judge did not have to decide between the evidence of the experts. Neither expert argued that the conclusions the Judge had drawn from the lay evidence and primary facts ‘would be incompatible with (to use the phrase in its most general sense) “the science”’. The Judge said, nevertheless, that he found Mr Johnston’s analysis more helpful. The Judge put to one side Mr Sorton’s lately disclosed change of opinion, and the surprising fact that he did not look at digital copies of damage to the car until the trial, despite their importance to the close analysis of the previous two years. The problem was that Mr Sorton did not ‘convincingly’ explain why he had changed his mind. His unconvincing explanation for the damage was ‘somewhat troubling’ (another example of understatement by the Judge) because his earlier concession was ‘plainly more favourable’ to R’s case. The Judge gave ‘broad examples’ of why he preferred the analysis of Mr Johnston to that of Mr Sorton. He again rejected the suggestion that Mr Johnston had shown he was partisan by putting forward the theory that A was falling forward at a late stage, and the Judge explained why (paragraph 115).

A argued that the ‘core issue’ for the trial was whether, when R’s car hit A in a crouched position, he was trying to get back to his feet, having fallen to the ground on his bottom, or whether he was falling forwards (paragraph 8). A noted that R did not plead that A fell forwards before being hit (paragraph 5). A submitted, that, ‘in a sign of things to come’ the Judge’s description of the two related core issues in paragraph 3 revealed a ‘failure to grasp the core issue’. A contended that A’s body position at impact was ‘set by the evidence of Dr Hulse, and that the issue was whether he was in the process of getting up, or falling down’ (paragraph 10).

In paragraph 11, A criticised the Judge’s summary of the law. It said that the Judge failed to recognise that the standard of care is exacting, and that this was ‘of a piece with [his] overly benevolent approach to the evidence of [R]’.

The Judge was also criticised (paragraph 12) for repeatedly downgrading the evidence of Dr Hulse (A referred to paragraphs 8, 35, 72 and 102 of the judgment). He was said to have failed to recognise that this evidence put A into a position best described, as it was by R in her witness statement, and by Nicholas Stannard (though, as A concedes, significantly, in my judgment, he did not use the word) as ‘crouched’.

In paragraph 13, A summarised the Judge’s reliance on the evidence of the lay witnesses. The Judge is said to have rejected the evidence of Nicholas Stannard and of his mother because it was inconsistent with that of R, Alexander Gent, and Nicholas Stannard’s initial account. That conclusion is said to be based on the premise that it was unlikely that R would have failed to see A in the road if he had been getting up from a fall to the ground. The Judge rejected the submission that it was inherently improbable that a person would fall forwards in a crouched position. ‘He forgave (our expression) Mr Gent for failing ever to describe seeing [A] fall’ before the collision. He did not address the evidence that Alexander Gent’s view ahead was blocked by a car.

A then described the ‘hard’ evidence (both in the skeleton argument and in his oral submissions). It was the evidence of Dr Hulse, and of the damage to R’s car (paragraph 16). The front offside of the car was damaged, low down, but not the bonnet or the windscreen. That evidence was said to be ‘incontrovertible evidence’ of A’s body position at the time of the collision. A accepted that Dr Hulse could not say whether A was stationary, or getting up, or falling down, when he was hit by the car, ‘But it was beyond doubt in this trial that [A] was, highly unusually, in this crouched position when hit’. A described that as ‘the hard evidence on body position’.

A submitted that the grant of permission to appeal meant that this Court was to decide whether the Judge was wrong ‘giving full weight to the advantages enjoyed by the Judge’. A developed this point by suggesting that it was for this Court to ‘make up its own mind about the correctness or otherwise of any findings of primary fact or inferences from primary fact’ challenged by A ‘while reminding itself that so far as the appeal raises issues of judgment on unchallenged primary findings and inferences, the court ought not to interfere unless it is satisfied that’ (in short) the Judge’s view was wrong. I will return to this submission below (paragraph 90).

A submitted that Gestmin shows that it is dangerous to rely on evidence which is given confidently. Witness evidence has to be assessed ‘in its proper place’ with contemporaneous documents, and with other evidence which can be relied on. Witness evidence should be assessed against other evidence which is agreed, or clearly established by other evidence. The Judge recognised this in paragraph 9, but did not follow that approach in practice, A submitted.

In paragraph 25, A returned to his complaint that the Judge’s incomplete summary of the law set ‘the tone for his uncritical approach towards drivers’. A did not criticise the Judge’s approach to the law about evidence. A did, however, criticise the Judge for finding that the objective evidence did not help him. Paragraph 35 ‘encapsulated’ the Judge’s primary error. The ‘hard’ evidence about body position was a ‘lodestar’ against which the evidence of Nicholas Stannard and Alexander Gent had to be tested. That was ground 1. The Judge’s ‘other failings…can be said to fall as easily within ground 3 as any other’.

The Judge’s ‘errors’ are described in five groups. They concern his failure to understand and apply the ‘hard’ evidence on body position, his analyses of the evidence of Nicholas Stannard, Alexander Gent and R, and his failure to understand that A’s ‘crouched position was readily [sic: perhaps ‘really’ is intended] consistent only with’ A’s case.

The first failure was said to have ‘infected’ all the Judge’s reasoning, and his justification for rejecting the evidence of Nicholas Stannard. A referred to paragraphs 8, 35, 72 and 102 of the judgment. The Judge had overlooked the significance of the objective evidence that A was in a crouched position when he was hit. He should have recognised that ‘highly unusually’ A was in a crouched position when hit. The Judge needed to test all the evidence against this fixed point. If he had done that, he would have realised that the evidence of Nicholas Stannard was consistent throughout that A was in a crouched position, long before the medical evidence emerged (an indicator of its reliability); Alexander Gent never gave evidence that A was in a crouched position and gave positive evidence which was inconsistent with the ‘hard’ evidence about body position and R supported this evidence because she described A being in a crouched position in her witness statement. The only reasonable conclusion was that Nicholas Stannard’s evidence should be preferred. That conclusion is bolstered by the Judge’s other errors, which should ‘strip this court of any confidence in the conclusion reached’.

The starting point was said to be Nicholas Stannard’s account to PC Giles which the Judge found was accurate. In paragraph 34, A made several points about PC Giles’s account which, it is very likely, were made to the Judge, and appeared, to that extent, to be re-arguing the case. The Judge had ‘fallen into the Gestmin trap’ by relying on the clarity of PC Giles’s evidence and on the fact that he ‘was having none of that’. It was said the Judge had failed to see the limitations of the evidence of PC Giles and to ‘set it against the evidence of Nicholas or his mother’. Her evidence was that Nicholas Stannard had always shown how A ‘slipped backwards’. Nicholas Stannard has used the word ‘slipped’ all along, including in PC Giles’s note (‘if that is accurate’). That word can mean ‘slip backwards’ and does not normally mean ‘slip forwards’. The evidence of PC Giles’s reference to ‘running’ was disputed, but ‘slipped and hit car’ is consistent with Nicholas Stannard’s evidence although it is incomplete. The Judge’s construction of this evidence and his use of Nicholas Stannard’s failure to describe A falling over and trying to get up again ‘against Nicholas…is very revealing for the partisan manner in which the Judge repeatedly analysed the evidence’.

A suggested that the Judge’s statement that Nicholas Stannard had given at least five accounts was ‘clearly pregnant with criticism’. A claimed that PC Giles ‘had a clear view in his mind long before’ A’s grandfather spoke to PC Giles in November 2015, ‘that [A] had run out, hence his not proceeding to interview Nicholas or take a statement from him’. It was suggested that PC Giles put a ‘spin’ on the giving of the account in November 2015. These are more points which were no doubt put to the Judge in closing submissions, but which are out of place on an appeal. The ‘stark contrast’ recorded by PC Giles may have felt like that to him but ‘should not have to the Judge’ (paragraph 39).

In paragraph 40, the Judge was said to have drawn the wrong inference, and to have failed to recognise that in November 2015, Nicholas Stannard had described A as falling over and then getting up, that is, as not yet standing. The Judge was said to have ‘entirely ignored’ the significance of this evidence. It was the only witness evidence that was consistent with the ‘hard’ evidence about body position, it was given by Nicholas Stannard in November 2015, very soon after the accident and long before the medical evidence, ‘which so clearly corroborated Nicholas’s evidence’ emerged.

Instead, the Judge wrongly found that the disparity between the facts and Nicholas Stannard’s evidence was due to the factors which degrade the quality of recall (per Gestmin). The Judge was said to have ignored the November 2015 account. ‘His recall was consistent from, on his evidence, the time of the accident, and, at the least, from November 2015. There was no basis for the suggestion that the quality of his recall had degraded, quite the opposite’ (paragraph 42).

This finding was ‘key to the Judge persuading himself he was to reject Nicholas’s evidence…’. Nicholas Stannard is said to have given consistent evidence at trial, in the face of skilful cross-examination ‘again, a point not acknowledged or taken into account by the Judge’ (paragraph 42). Another ‘signpost to the one-sided nature of the Judge’s analysis’ was said to have been an inference that if A had slipped and was getting to his feet, Nicholas would have warned him about the approaching traffic. This is said to be ‘a remarkably unfair inference’ to draw about an 11-year old. ‘It is to be contrasted with the high degree of benevolence shown to [R], an adult’ (paragraph 43).

The Judge’s approach to the road conditions was said to be one-sided. A took issue with the difference between ‘slightly damp’ and ‘fairly damp’ (the Judge’s record of the evidence and the actual evidence), and a failure to allow for the drying of the road between the time of the accident and PC Wheeler’s arrival (paragraph 44).

The Judge is criticised for choosing not to set out Alexander Gent’s evidence that A bounced off the bonnet, or that he ‘folded over the bonnet and rolled off’. Alexander Gent accepted at trial that if the physical evidence did not support his account, his account must be wrong (paragraph 45). The Judge should have tested this evidence against the ‘hard’ evidence. Had he done so, he would have realised that the ‘hard’ evidence contradicted Alexander Gent’s evidence (paragraph 46). The Judge’s approach to ‘this part of Mr Gent’s evidence is highly revealing and regrettably indicative of his approach overall’. A then quoted paragraph 106 of the judgment (see paragraph 55, above). A argued that the evidence that A went over the bonnet of the car meant that the evidence of Alexander Gent was ‘holed beneath the waterline’. The ‘hard’ evidence showed that that did not happen. Mr Gent’s evidence did not describe the foot trap which occurred.

It was then said that this Court is as well placed as was the Judge because there was no evidence for the ‘plausibility theories’ adopted by the Judge. This Court should not be deterred by the Judge’s finding that Alexander Gent was ‘impressive’ as that is exactly the sort of danger a court should guard against. The important point is that Alexander Gent had said from the beginning that A bounced off the bonnet. That was wrong and shone ‘the brightest light on the reliability of his evidence’ (paragraph 49).

The Judge also failed to take into account evidence that Alexander Gent was driving behind another southbound car, which would have blocked his view. Nicholas Stannard’s evidence at trial was that A had started to cross after one southbound car had passed and when two other cars were approaching. This was consistent with Alexander Gent’s evidence that he was driving behind a grey car.

A criticised the Judge for not referring to the unchallenged evidence of Mr Hardaker that the grey car flashed its headlights as a warning to northbound cars because A was crossing the road and had slipped in the middle. This was consistent with evidence that Alexander Gent was behind the grey car when A started crossing the road and with the grey car passing A just before the collision, although Nicholas Stannard did not notice this.

The Judge’s approach to R’s evidence is also criticised. A suggested that R had changed her statement to the police from evidence that A was running to a description of A being crouched, so that she assumed he was running, but had only seen him the moment before the collision. Her evidence was consistent with A’s case that he was crouched at the moment of impact, yet the Judge found that her evidence was inconsistent with it. The Judge was taken to task for his premise that it was unlikely that R would not have seen A if he had been fallen in the road and was trying to get up at the point of the collision. A complained that it is ‘transparently unfair’ to assume that drivers are not careless. It was suggested that the Judge did not decide the case on the evidence. This ‘approach’, A then submitted ‘tainted’ the Judge’s analysis of R’s evidence. The Judge ‘persuaded himself’ that Nicholas Stannard must be wrong because Alexander Gent did not see A in the road, either. The Judge applied a ‘numbers approach’ the evidence rather than analysing its contents. The evidence of Alexander Gent was ‘clearly flawed’ because he ‘promoted a version of events which categorically did not happen’.

Properly analysed, R’s failure to see A was explained by the fact that she was concentrating on Nicholas Stannard. She accepted in cross-examination that she thought he might come out into the road. It was ‘not so surprising’ that she failed to keep a look out ahead’, failing to see A until the last moment and failing to see a car coming towards her until the last moment. The Judge did not refer to this evidence in the judgment.

The Judge, it was said, failed to recognise that A’s crouched position was consistent only with A’s case. This Court, it was said, was as well placed as the Judge to rely on its own experience and to decide whether a person could get into a crouched position by falling forward. The Judge ‘respectfully erred’ in rejecting the submission that the ‘hard’ evidence of A’s body position was only consistent with the evidence of Nicholas Stannard. A suggested that there was ‘a strong sense that the Judge, having attached himself to the theory [that ‘drivers are not generally careless’]’ then ‘subconsciously fitted his analysis of the evidence to fit this theory’. He was said repeatedly to have drawn unsound inferences which favoured R, and to have failed to credit the consistency of the evidence of Nicholas Stannard with the ‘hard’ evidence of body position. ‘Each aspect of his justification for his findings can be shown to be flawed’.

In his helpful submissions, Mr O’Sullivan pointed out that neither party at trial attributed to the ‘hard’ evidence now relied on by Mr Weir the significance which Mr Weir says it now has. His broad point, which he supported with cogent detailed written and oral argument, was that it was open to the Judge to make the findings which he did on the key issues, and that A has not identified any finding or evaluation made by the Judge which is wrong. This Court, therefore, should respect the decision of the Judge.

At the start of his oral submissions Mr Weir acknowledged the statement of Lewison LJ in paragraph 114 of Fage UK Limited v Chobani UK Limited [2014] EWCA Civ 5; [2014] FSR 29. This Court should not interfere with findings of fact, inferences from findings of fact, or a judge’s evaluation of facts, unless compelled to do so; that is, unless those findings are ‘wrong’. Lewison LJ listed six well known reasons for that approach. They include that ‘In making his decisions the trial judge will have regard to the whole sea of evidence, whereas an appellate court will only be island hopping’. He added, in paragraph 115, that the judge is not obliged to deal with every single piece of evidence in his judgment. ‘The primary function of a first instance judge is to find facts and identify the crucial legal points and to advance reasons for deciding them in a particular way’. His reasons need not be elaborate. He does not need to deal with every argument or to spell everything out.

I have summarised the Judge’s careful and thorough judgment, and Mr Weir’s detailed criticisms, at some length. At the outset, I reject Mr Weir’s submission that the Judge approached the case in an unfair way. He weighed the evidence conscientiously. His function was to decide the case for one side and against the other. There is no evidence of unfairness in his conclusion, or in the detailed and transparent reasoning which supports it. I also reject the contentions I summarise at paragraph 70, above. The grant of permission to appeal on the grounds that the appeal is ‘just arguable’ does not permit this Court to consider and re-make findings of fact unless this Court decides, first, that those findings were not reasonably open to the Judge, and were, therefore, wrong. It seemed to me that most, if not all, of Mr Weir’s submissions were designed to show that the Judge could have made different findings on the evidence, rather than to show that the findings which he did make were wrong.

The Judge has already trawled the sea of evidence, including the ‘confounding factors’, in search of ‘a reasoned view of the most probable explanation’. At the risk of ‘island hopping’, I will only therefore consider what seem to me to be the three main issues on this appeal.

At the start of his oral submissions, Mr Weir contended that the Judge had, in paragraph 3, framed the core issues incorrectly. He was asked to explain in what way the Judge had erred, since the context for those factual questions posed by the Judge was set by the parties’ answers to those two questions, which the Judge summarised in paragraphs 5 and 6. The key issue in the case was whether A was in the road for a short, or for a long time, as that, the parties agreed, determined whether R would have had time to see him, and avoid him, and thus, whether she had been negligent or not. In the light of way the Judge posed the questions, described the parties’ answers, and carefully analysed the evidence, I do not consider that the Judge misunderstood what the case was about. That is supported by the Judge’s description of the issues in, for example, paragraph 42 (see paragraph 23, above). The Judge’s understanding of the issues is only wrong if Mr Weir is right, and the evidence about body position did unlock the case. If, instead, it was open to the Judge to decide that the evidence about body position did not unlock the case, there is nothing in this argument.

The first point is that the Judge recognised that the ‘hard’ evidence might unlock the case. He analysed the evidence with that point in mind, and decided that the ‘hard’ evidence was not the key (see, for instance, paragraph 99, paragraphs 48 and 49, above, and the cogent analysis in paragraphs 112 and 113, see paragraphs 61-62, above). That is a classic example of an assessment of the primary evidence by a first instance judge. Any attack on it runs up against Lewison LJ’s strictures in Fage. In any event, there is no merit in this attack on the Judge’s approach. Mr Weir’s argument was that the ‘hard’ evidence showed that A was in a ‘crouched’ position when he was hit, and that the Judge should have used that ‘hard’ evidence to help him to decide which evidence he should accept from the eye witnesses. There are three main difficulties with that argument.

The first is that its premise is that the evidence of a ‘crouched’ position necessarily showed that the second account given by Nicholas Stannard was true; that is that A slipped backwards onto his bottom, and was then trying to get up by crouching forwards (‘crabbing forwards’, in Mr O’Sullivan’s phrase) when he was hit by R’s car. The problem with that submission is that there was no ‘hard’ evidence about the biomechanics of falling (or indeed, of getting up from a fall). Mr Sorton was sceptical that figure 4 could show what a fall forward from a run would look like, but accepted that he had no expertise on this issue and that whether it did was for the Judge to decide, as a matter of common sense (paragraph 45, above). Nor was there any expert evidence about how common it was for a pedestrian to be in a crouched position when hit by a car (cf paragraphs 19 and 31 of A’s skeleton argument). A’s argument requires this Court to disagree with the Judge’s conclusion on this point, about which there was no expert evidence. The Judge’s resolution of this issue is a textbook example of the sort of question of fact which it is for the judge to decide, and with which this Court will not interfere on an appeal.

The second difficulty with this argument is that it is not entirely supported by the evidence Mr Weir relied on. Nicholas Stannard did not use the word ‘crouched’, as Mr Weir accepted. He used the word ‘slip’ (either when running or walking) and in his second and subsequent accounts described A trying to get up after slipping. R did use the word ‘crouching’, but explained that what she meant by it was ‘leaning forwards’. The submission requires one word to summarise the pictures at figure 4, and it puts a significance on the word which it cannot bear.

The third related difficulty with the argument is that it uses the word ‘crouched’ as a proxy for a contested sequence of events. It is clear from the evidence of the experts which the Judge summarised, and to which Mr O’Sullivan took us in his oral submissions, that neither Dr Hulse nor Mr Sorton suggested that any inference about what A was doing immediately before he was in the crouched position could be drawn from their limited evidence about A’s body position when he was hit by R’s car.

The starting point is that the terms of the judgment clearly show that the Judge was alive to the problems with the recollection of eye witnesses which are highlighted in Gestmin. In the light of that evident understanding, a submission that the Judge misunderstood the weaknesses of such evidence is ambitious. Indeed the Judge relied on one of the points made in Gestmin (that an account given nearer to the event in question is more likely to be reliable than a later account or later accounts): paragraphs 36 and 38, see paragraph 20 and 21, and paragraph 104, see paragraph 55, above. The Judge was entitled to adopt that approach. He also well understood that it was necessary to analyse the expert evidence to see to what extent it helped him to decide the issues in the case. Moreover, he was right to think that Gestmin, which was a commercial dispute, in which there was a significant ‘digital footprint’, did not require him to reject the evidence of the eye witnesses in a road traffic case (paragraphs 34-35, see paragraph 19, above). He also recorded that the parties accepted that Nicholas Stannard’s evidence was central to the dispute (paragraph 42, see paragraph 23, above). Finally, the Judge’s ‘plausibility theories’, to which Mr Weir took exception, are tests of the evidence by reference to its inherent probability, which, Gestmin recognises, are not only unobjectionable, but useful.

The Judge was entitled to accept the evidence of PC Giles that the account he recorded at the scene was accurate. PC Giles was cross-examined at length. The Judge summarised the cross-examination in paragraph 59 (see paragraph 31, above). This is another textbook example of a decision for the first instance judge. There are two telling points about this account, as the Judge recognised. First, Nicholas Stannard described A crossing the road with him, realising he had forgotten something, going back home, being warned by him, and running across the road and slipping. That description was inconsistent with Nicholas Stannard’s later account, and with A’s case. I reject Mr Weir’s attempt (which was forensically necessary) to suggest that it was consistent, albeit laconic. That was not Ms Rodway QC’s assessment of the significance of that evidence (judgment, paragraph 58, see paragraph 31, above). I also reject Mr Weir’s submission (see paragraph 75, above) that the Judge’s approach to this omission showed he was partisan: Ms Rodway, like the Judge, appreciated that this gap in the account was an important omission. Second, although Nicholas Stannard and his mother disavowed that account, Nicholas Stannard accepted in cross-examination that he had warned A and that other aspects of the account were also accurate. That account could only have come from Nicholas Stannard or from him, via his mother. To suggest that the Judge fell into ‘the Gestmin trap’ and simply accepted PC Giles’s evidence because of his robust response to cross-examination is not an accurate reflection of the Judge’s careful approach to that evidence.

Having decided, as he was entitled to, that the evidence of PC Giles was accurate, the Judge was also entitled to decide that Nicholas Stannard’s first account of the accident was more likely to be reliable than his second and subsequent accounts. He was also entitled to make the points which he made about the intrinsic improbability of the second and subsequent accounts. The Judge cannot be criticised for nevertheless accepting that Nicholas Stannard was doing his best to help the court. I reject the suggestion that the Judge’s reference to Nicholas Stannard’s five accounts was ‘pregnant with criticism’. It was factually accurate and a necessary foundation to the point, derived from Gestmin, to which I referred in the last paragraph but one.

Mr O’Sullivan pointed out in his skeleton argument that if A had run across the road from his drive, it would have taken him less than a second to reach the middle of the road where he was hit by the car, whereas if he had walked across the road, slipped over backwards and then tried to get up, that would have taken more than ten seconds. If Alexander Gent’s evidence that A ran out of his drive was correct, he would have had a view of A for a fraction of a second. The Judge expressly acknowledged that the evidence of Alexander Gent was not, in every respect, accurate or comprehensive. This was scarcely surprising, given the short time Mr Gent had to observe events (if he was right that A had run out of his drive). A premise of Mr Weir’s attack on the Judge’s approach to the evidence of Alexander Gent is that, having rejected, expressly or by implication, his evidence about A’s body position immediately after the collision, the Judge was obliged to reject all of his evidence. First, that is a non sequitur, as part of his evidence could nevertheless have been accurate. Second, a decision to reject part of the evidence of a witness but to accept some of it is another example of an evaluative assessment which it is for the first instance judge to make. The suggestion that the Judge accepted Alexander Gent’s evidence only because he was ‘impressive’, without more, is unfair to the Judge. First, the Judge explained, in paragraph 105 (see paragraph 55, above) why he found Mr Gent an impressive witness. Second, the Judge was well aware of the statements in Gestmin that it is dangerous to rely only on the impression made by a witness at trial. Third, the Judge did look critically at the evidence of Alexander Gent, and recognised that some of it was not accurate.

Mr Weir suggested that the Judge’s unfair starting point was that it was unlikely that R had been negligent. That was not the Judge’s starting point, for at least two reasons. First, what the Judge considered was unlikely was that R would have failed to have seen A if he had been in the road for the length of time which Nicholas Stannard’s evidence at trial indicated he must have been. Second, as Bean LJ observed in oral argument, the Judge came very close, in paragraph 104 (see paragraph 54, above) to saying that, were it not for the evidence of Alexander Gent, he might well have found for A. To consider whether or not R’s evidence was supported by other evidence is not to adopt a crude or wrong ‘numbers approach’ to the evidence, but rather to test it with a traditional forensic tool. This is yet a further example of a misguided attack on an entirely conventional approach to weighing the evidence. Moreover the Judge was not merely asking whether R’s evidence was ‘corroborated’. He was also asking whether it was probable that two unrelated witnesses, driving towards A from opposite directions, would both have failed to see him lying in the road and trying to get up over several seconds (paragraph 104, see paragraph 54, above).

The hearing of the trial in this case lasted five days. In deciding what evidence would help him to decide the issues, the Judge had, necessarily, to focus on the evidence which was likely to be important, and, in deciding what was important, he had also to decide the relative importance of different parts of the evidence. That process of selection is necessarily reflected in the judgment, which, without such editing, would be as long as the witness statements, the experts’ reports and the transcripts of the evidence at trial. I consider that Mr Weir’s arguments that the Judge erred in not mentioning all the evidence on which Mr Weir relies, and/or in not giving reasons for not dealing with it, or (by inference, perhaps, for not accepting it) are makeweights. They do not persuade me that, in an otherwise impeccable judgment, the Judge erred in any way. I have taken a similar approach to other points made by Mr Weir, such as his objection to the Judge’s description of the evidence about the precise condition of the road surface in paragraph 110 (see paragraph 59, above).

I reject Mr Weir’s submission that the Judge erred in law in his description of the relevant law. This case did not depend on any point of law. The law which applied was well known, and uncontroversial. The case was all about the factual questions which the Judge identified in paragraph 3, and the parties’ agreed positions about the implications of a decision on those questions (paragraphs 4-6), as the Judge observed (see paragraph 17, above). Even if the Judge had erred in his brief summary of the law, any such error would, self-evidently, have been immaterial.

For those reasons, I would dismiss this appeal. Mr Weir has given many reasons why he disagrees with the Judge’s findings, but he has not persuaded me that any one of those findings was not open to the Judge on the evidence, and was therefore ‘wrong’.

In 1978, after five years of deliberation, the Royal Commission on Civil Liability and Compensation for Personal Injury chaired by Lord Pearson recommended a no-fault compensation scheme for injuries caused in road traffic accidents. But Parliament has not acted on the recommendation. It therefore remains the position that where a child such as Felix is injured by a car whose driver is not shown on the balance of probabilities to have been negligent, the driver’s insurers are not liable to pay any compensation.

The deputy judge who tried this very sad case did so carefully and conscientiously, and I wish to pay tribute to the high quality of his judgment. Despite Mr Weir’s best efforts I am not persuaded that the judge made any error or that there is any ground on which this court could properly reach a different conclusion. For the reasons given by Elisabeth Laing LJ, I too would dismiss the appeal.

I agree with both judgments.