> A better explanation for the relative lack of success of disulfiram is that you don't take it from birth without choosing to. You have to want to take it to g

> A better explanation for the relative lack of success of disulfiram is that you don't take it from birth without choosing to. You have to want to take it to gain the benefit of it, but if you need it, you don't really want to take it. You want to drink. It's the big problem of addiction, you want things that contradict each other, and you want the bad things more. I don't see what supports this being a better explanation, I feel at this point all explanation that fits the data are equally likely, otherwise it's just bias. In fact, I would say the proposed model failed the first test, which is that a good model should be good at predicting outcomes. In this case, the model failed to predict the outcome, as causing the condition in people did not reduce their alcohol consumption as we expected from the model. Sure, it could be that only some details of the model are wrong, like that the condition only works if established prior to a history of alcohol consumption, or that it only works if having no choice on the matter of the condition, etc. But at this point it seems just as likely that the condition might not be the cause, but only correlated in possibly another way. For example, we might have it backwards, and it is the lower drive to consume alcohol which caused the gene variant in the fist place.