Discussion: Paper 8

Mr V. C. Marshall (Director of Safety services, University of Bradford)

Mr Farmer has exemplified the fact that prediction is a very tricky business, particularly when it is concerned with the future! The authorities he quoted represent the more sober end of the business: some of the figures have ranged as high as 8000 fatalities for 1 ton of chlorine spilt.

The question of mean results is perhaps more important than he has suggested. In any statistical distribution there is a mean and some measure of variance. In this field the curve is always heavily skewed because while there is a mean, the minimum is zero, because whatever the curves might say, there is no probability of actually creating life as a result of one of these accidents! Therefore there is a sharp boundary at zero at the low end, but at the high end there is a long tail off, with lower and lower probabilities attached to it.