"The trial judge, although not depending in any respect on the credibility of any witness, may have preferred one possible view of the primary facts to another as being in his opinion the more probable. Such a finding may, in my opinion, be disturbed by an appellate court but this should only be done if other probabilities so outweigh that chosen by the primary judge that it can be said that his conclusion was wrong. Again, the trial judge, having found the primary facts, may decide that a particular inference should be drawn from them. Here no doubt the appellate court has more room for setting aside that conclusion. But, even in that case, the fact of the trial judge's decision must be displaced. It is not enough that the appellate court would itself, if trying the matter initially, have drawn a different inference. It must be shown that the trial judge was wrong. This may be achieved by showing that material facts have been overlooked, or given undue or too little weight in deciding the inference to be drawn: or the available inference in the opposite sense to that chosen by the trial judge is so preponderant in the opinion of the appellate court that the trial judge's decision is wrong."