13 It has been suggested that this type of reasoning was disapproved by the Victorian Court of Appeal in Kocis v S E Dickens Pty Limited [1998] 3 VR 408. However, although Philips JA in that case expressed himself as having some difficulty with the reasoning in Rose, in my opinion, that reasoning was substantially endorsed (and in my opinion correctly so) in the judgment of Hayne JA at 430-432:
"In Rose v Abbey Orchard Property Investment Pty Ltd [1987] Aust. Torts Reps. 68925 the New South Wales Court of Appeal said (at 68929): "In many, if not most, 'spillage' cases the plaintiff will fail to prove a causal connection between breach and damage unless he establishes how long the substance had been on the premises. But in some cases it may be possible to establish on the probabilities that a proper system would have eliminated the risk of injury even though it is not possible to determine how long the substance had been present." The central question in the present case is whether there was sufficient evidence from which a jury could conclude, on the probabilities, that the defendant's failure to follow a reasonable system of cleaning caused the plaintiff's damage.
The Court held in Rose's Case that a proper system of inspection of the premises concerned would have required inspection of the floor intervals of not more than 20 minutes. It followed that if the oil upon which the plaintiff slipped was present before the last of those inspections should have taken place, the implementation of that system would have led to the removal of the oil. The Court went on to say that: "The oil upon which the plaintiff slipped may have fallen at any time during the hour before her accident. If it was spilled in the period between 2.50pm and the time of the accident, a proper system of inspection could not have prevented the accident. But, other things being equal, the probabilities are twice as great that the oil was spilled in the 40-minute period between the last inspection and 2.50pm rather than in the 20-minute period after 2.50pm. The evidence contained no suggestion that the use of the parking station increased or changed in the 20-minute period immediately before the accident. Accordingly, we think that as a matter of probability the oil was spilled before 2.50pm and not after that time. To so find is not to engage in speculation but to make a finding in accordance with probability theory.
It was submitted on behalf of the respondent that the "probability theory" referred to in Rose's Case was too mechanistic an approach to be adopted in cases of this kind. It was submitted further that it was an approach that, at least, stood apart from the approach adopted in other cases. In particular, it was submitted that the approach adopted by the Court in Rose's Case was to be contrasted with the different approaches taken by the members of the Court of Appeal in Brady v Girvan Bros Pty Ltd (1986) 7 NSWLR 241.
…
Now it may be accepted that the judgments reflect some differences in approach to the problem presented in that case. But in my view none of the judgments in Brady v Girvan is to be taken as stating any principle more specific than that it is for the plaintiff to show that it is more probable than not that the plaintiff's injury were cause by the defendant's breach of duty. Nor are the judgments in Brady v Girvan (or, for that matter, the judgment in Rose's Case) to be taken as stating some special principle of law that is to be applied in slipping cases. The question of causation is a question of fact. If, as McHugh, JA considered to be the case in Brady v Girvan , the evidence reveals how long the substance concerned had been on the floor, the question becomes whether reasonable care would have led to its removal. But if it is not possible to say when the substance was dropped, it by no means follows that the evidence does not permit a conclusion that more probably than not the defendant's breach of duty caused the plaintiff's damage.
The point is well illustrated by facts that are different in degree (but not kind) from those considered in Rose's Case. Let it be assumed that a reasonable occupier of certain premises would inspect the premises for spillages once each hour. Let it further be assumed that the evidence demonstrates that the occupier made no inspection of the premises at all on the day on which the plaintiff slipped and fell eight hours after the premises opened for business. If that is all that is known, it is of course possible that the substance upon which the plaintiff fell was dropped one minute or 59 minutes before the fall occurred but what are the probabilities? In my view it is open on those facts to conclude that it is more probable that the spillage occurred in the first seven hours of trading than it is that it occurred in the last hour. It would follow that had a proper system of inspection been implemented, it is more probable than not that the spillage would have been detected and removed. Implicit in the example I have given is that there is no basis for concluding that the spillage is more likely to have occurred at one particular time (or at some particular times) rather than others. That is, the example assumes an undifferentiated period of time in which the spillage may have occurred. That is why it would be properly condemned as "speculation" to say that it is more probable than not that the spillage occurred at a particular time in that period; but that is not the chain of reasoning that is suggested. Rather, it is suggested that common sense dictates that where there is nothing pointing to a particular time as the time of occurrence, the longer the time under consideration, the more likely it is that the spillage occurred during that time than in a different, shorter period.
Such an analysis does not depend upon considerations of "increased risk" or "exposure to risk". In my opinion to speak in terms of increased risk or increased exposure to risk may serve only to obscure the fundamental factual question whether the defendant's negligence caused the plaintiff's damage. It may do that because it may distract attention from exactly what is meant by there being an increased risk or increased exposure to risk. The conclusion that there is an increase in risk or exposure to risk in cases of the kind I have described depends upon an earlier (unstated) conclusion that as time passes the probability that something has been spilled on the floor rises. Further, to analyse the matter in terms of increased risk or exposure to risk is apt to mislead. To return to the example I gave earlier, it is clear that the longer the premises are not inspected, that is, the more inspections that should have been, but were not, carried out, the greater the risk that something has been spilled on the floor which should have been cleaned up. The risk of someone slipping at the premises has increased. But it is still necessary to ask whether the breach of duty has caused the damage. Can it be said that the failure to clean caused the plaintiff to slip? If the evidence permitted the conclusion that it was more probable than not that the plaintiff fell upon something that had been dropped no more than one minute before the accident, the defendant's failure to clean the premises would not have caused the plaintiff's loss, however much it could be said that the defendant's failures had increased the risk to persons entering the premises. As I have said earlier, the question of causation is a question of fact requiring close attention to the facts; it is not a question which requires or permits the application of any special or different principle of law in slipping cases".