Judge Eric Stansifer’s follow-up post directly engaging Michael Weissman’s “An Inconvenient Probability” (S-64 - Weissman’s ‘An Inconvenient Probability’ Bayesian analysis). Concedes a “very good catch” on Weissman’s point that Stansifer had conflated a Bayesian likelihood with a p-value in places, but defends his own use of Worobey-derived Huanan Seafood Market location data against Weissman’s data-integrity concerns (arguing a full Chinese cover-up/fabrication of the market-clustering pattern — a “Potemkin market” — is itself improbable), and disputes Weissman’s low (~1/50) estimate for the probability WIV would attempt engineered gain-of-function work, defending a higher figure from his own close reading of the DEFUSE proposal. Does not restate a single final number but maintains the disagreement does not move his conclusion past even odds.
relevance_note: a primary methodological/prior-setting piece — a direct, itemized argument between two of the case’s six analysts about which Bayes factors to trust, useful for weighing how much the “23 OOM” spread reflects genuine evidential disagreement vs. correctable methodological error.