I don't see why the first hypothesis should necessarily be rejected out of hand. If the supposed mechanism is unconscious then having it react to erotic pictures and not particular casino objects seems perfectly plausible. Obviously the real explanation might be that the data wasn't strong enough to prove the claim, but we shouldn't allow the low status of "psi theories" to distort our judgement.
One good thing about Bayesian reasoning is that assigning a prior belief very close to zero isn't rejecting the hypothesis out of hand. The posterior belief will be updated by evidence (if any can be found). And even if you start with a high prior probability and update it with Bem's evidence for precognition, you would soon have a posterior probability much closer to zero than your prior :)
BTW there is no supposed mechanism for precognition. Just calling it "unconscious" doesn't render it any more plausible that we have a sense that would be sup...
Here's the new thread for posting quotes, with the usual rules: