Here I will list the evidence in this case that I believe is probative, and assign probability values to them. And then calculate the probability of guilt.
Evidence pointing to innocence:
1. Evidence of a break in. There is a cut screen and a bloody finger print that could not be matched to anyone in the house. The evidence of breakin is kind of weak, but there nonetheless. I estimate this evidence as having a probative value of 40% --> Pointing to Routier's innocence for 60% or 40% guilt.
2. Routier's wounds do not appear to be self-inflicted. I am no expert but from the descriptions and pictures I've seen, I think it is unlikely she caused them herself. Guilt value: 35%
3. Routier's sons were alive when she called 911---one was still alive when the police arrived. Why would a murderer call 911 before her victims were dead? She changed her mind? She thought they were dead? or was she not the murderer? guilt value: 25%
4. The bloody sock. It is unclear how Routier could have planted this. An intruder may have used it and dropped it while he was running away. Guilt value: 40%
Evidence pointing to guilt:
5. Darlie Routier could not give a straight story about what happened. While this points to guilt, I also believe someone sufferring a traumatic experience may possibly not remember things well. So, it is not absolute proof of guilt--not proof she was lieing. Guilt value: 60%
6. An intruder stabbing the boys to death but only non-fatally injuring an adult also there. This seems unlikely---though if Darlie fought back, it may have scared the intruder off, maybe why he didn't kill her. Guilt value: 55%
7. Darlie Routier was concerned that she touched the knife and telling the operator over the phone she might have got her fingerprints on it. This could be an indication that she was making up an allibi to explain her fingerprints. Guilt value: 60%
The probability of guilt, given a list of evidence e1, e2, ..., en each with an associated probability of guilt of p(ei) can be calculated as:
p(e1)*p(e2)*...p(en)/(p(e1)*p(e2)*...p(en)+(1-P(e1))*(1-p(e2))*...(1-p(en)))
.4*.35*.25*.4*.6*.55*.6/(.4*.35*.25*.4*.5*.55*.6+.6*.65*.75*.6*.4*.45*.4)
= 0.17 17% probability of being guilty or 83% probability of being innocent.
So I calculate Darlie Routier is innocent with probability 83%
PS: yes, I like numbers!
David