1.False 2. False 3. True 4. True

Why am I giving (most of) these in boolean terms rather than probabilities? Bayesian probabilities aren't useful in cases where the most probable scenario for (AK guilty) is something like "Two of the perpetrators were secretly ninjas". There really is no rational way to convict someone for leaving no forensic evidence in a room whatsoever.

I have to admit here though that I peeked at your article before posting this. And incidentally, predicted what it would say pretty damn well. (AK not guilty with a probability that reduces to 0, with the other two probabilities also expressible in boolean terms, and on the whole contradictory of the opinion of the jury)

I also have to admit I skipped straight to Wikipedia after reading your article, and found mostly that the facts you gave were correct and thus your argument was sound. My prior probability for any of them being guilty was very low however; around 10%ish. Jury decisions are pretty worthless before an appeal.