[johno]

This is my rather limited understanding!

Baye's theorem states that the probability of a hypothesis (innocence) given a piece of new evidence (the shoe print) is dependent on -

1) pre-evidential probability (innocence without knowledge of the shoe print).

2) ratio of the probability of the evidence 'ocurring' given the hypothesis is true to the probability of the evidence occuring for all hypotheses (guilty or innocent in this case).

In theory the effect of the new evidence can be quantified - the applicability of this result depends on the accuracy in variable quantification-

1) The accuracy of determining the probability of innocence without the shoe print evidence - this may be very difficult to assess, if for example only people with that shoe size were considered suspects.

2) The accuracy in assessing the probability of the evidence 'occuring' if the defendent is guilty. If you assume that the murderer definitely made the footprint then this is easier - otherwise you need to attach some kind of weighting or probability to the possibility that it is a random footprint.



In addition, you need to believe that Baye's theorem is applicable to the degree of a person's belief in a proposition, and have a view on the degree of subjectivity required to apply the theorem in this situation.


An aside - if the prisoner is left to rot in jail, is he serving 'sklansky years'?