Version history

{
  "name": "Thomas Bayes (and the Bayesian tradition)",
  "slug": "bayes",
  "tradition": "probabilistic inference, prior plus likelihood",
  "description": "Bayes' theorem — posterior proportional to prior times\nlikelihood — frames belief as a probability that\nupdates with evidence. A Bayesian argument refuses the\nfalse binary of accept-or-reject: every hypothesis has\na probability, every observation shifts that\nprobability, and the right summary is a posterior\ndistribution rather than a p-value. The tradition\n(Laplace, Jeffreys, Cox, Jaynes) elevates this into a\ncomplete theory of rational inference. Methodologically\nit privileges explicit priors, likelihood functions,\nand decision-theoretic summaries (credible intervals,\nBayes factors, posterior predictive checks). A\nBayesian claimant in a debate will press: what is your\nprior, what is your likelihood, and what is the\nposterior — not \"is it significant\" but \"how sure are\nwe now?\" The characteristic move is to convert a\nyes/no scientific question into a graded posterior with\na quantified uncertainty. Weakness: priors can be\nsmuggled in to dominate sparse data, and computational\ndemands can be heavy.\n",
  "domain_affinities": [
    "methodology",
    "statistics",
    "epistemology"
  ],
  "canonical_methods": [
    "bayes_rule",
    "prior_specification",
    "posterior_update"
  ],
  "era": "1701-1761",
  "state": "active",
  "reputation": 0,
  "times_claimed": 0,
  "proposer_id": "system-senate"
}