- I do not think that there is any hard and fast rule. The general position established by Warren v Coombes  is that on an appeal by way of rehearing from a judge sitting without a jury, an appellate court is in general in as good a position as the trial judge to decide on the proper inference to be drawn from facts which are undisputed or which, having been disputed, are established by the findings of the trial judge. The appellate court should give appropriate weight to the conclusion of the trial judge but, if of a different opinion, must give effect to that opinion  .
- The decision in Warren was considered and applied in Fox v Percy  . Gleeson CJ, Gummow and Kirby JJ referred  to Warren and, having done so, reiterated  “the need for appellate respect for the advantages of trial judges, and especially where their decisions might be affected by their impression about the credibility of witnesses whom the trial judge sees but the appellate court does not”. Their Honours said  that the requirement to afford appropriate respect to the reasons of trial judges does not “derogate from the obligation of courts of appeal… to perform the appellate function as established by Parliament”.
- Gleeson CJ, Gummow and Kirby JJ then said (omitting footnotes)  :
 Over more than a century, this Court, and courts like it, have given instruction on how to resolve the dichotomy between the foregoing appellate obligations and appellate restraint. From time to time, by reference to considerations particular to each case, different emphasis appears in such reasons. However, the mere fact that a trial judge necessarily reached a conclusion favouring the witnesses of one party over those of another does not, and cannot, prevent the performance by a court of appeal of the functions imposed on it by statute. In particular cases incontrovertible facts or uncontested testimony will demonstrate that the trial judge’s conclusions are erroneous, even when they appear to be, or are stated to be, based on credibility findings.
 That this is so is demonstrated in several recent decisions of this Court. In some, quite rare, cases, although the facts fall short of being “incontrovertible”, an appellate conclusion may be reached that the decision at trial is “glaringly improbable” or “contrary to compelling inferences” in the case. In such circumstances, the appellate court is not relieved of its statutory functions by the fact that the trial judge has, expressly or implicitly, reached a conclusion influenced by an opinion concerning the credibility of witnesses. In such a case, making all due allowances for the advantages available to the trial judge, the appellate court must “not shrink from giving effect to” its own conclusion. Finality in litigation is highly desirable. Litigation beyond a trial is costly and usually upsetting. But in every appeal by way of rehearing, a judgment of the appellate court is required both on the facts and the law. It is not forbidden (nor in the face of the statutory requirement could it be) by ritual incantation about witness credibility, nor by judicial reference to the desirability of finality in litigation or reminders of the general advantages of the trial over the appellate process.
- Their Honours observed  , further, that both trial judges and appellate courts should “limit their reliance on the appearances of witnesses and… reason to their conclusions, as far as possible, on the basis of contemporary materials, objectively established facts and the apparent logic of events”. That approach, their Honours said, “does not eliminate the established principles about witness credibility; but it tends to reduce the occasions where those principles are seen as critical”.
58 It was against that statutory and factual background that Heydon J said, in the particular passages of the paragraph on which Mr Donaldson relied (omitting footnotes)  :
Dr Harvey gave an account of her mental processes in an affidavit. The respondents’ searching cross-examination of her is recorded over seventy pages of the trial transcript. The record of her re-examination extends over three pages of that transcript. The assessment of a witness’s mental processes is an assessment of that witness’s state of mind. It is pre-eminently a matter in which a trial judge has a considerable advantage over an appellate court. In the course of his great speech in Nocton v Lord Ashburton, (  AC 932) Viscount Haldane LC said:
“it is only in exceptional circumstances that judges of appeal, who have not seen the witness in the box, ought to differ from the finding of fact of the judge who tried the case as to the state of mind of the witness.”
59 I do not think that there is anything in what Heydon J said in Barclay  , to suggest that that the principles relating to appellate review, by way of rehearing, established in the cases to which I have referred, should be revisited. On the contrary, I think, Heydon J intended to stress the significance of the advantage of the primary judge in Barclay, and thus the significance of the commensurate disadvantage of the appellate court.