Simulations

Checking what you believe before the data has a chance to flatter you.

A prior is a claim about what data is plausible. That claim can be simulated, looked at, and rejected — before any real observation is involved. These posts do that, and also take the wider case: run a method against assumptions it does not hold and see precisely where it bends.

A Normal(0, 10) prior on a logistic coefficient is not weak — it is a confident bet on certainty

prior-sims

Priors are chosen on the log-odds scale and interpreted as vague. Pushed through the inverse logit, a wide Normal prior puts most of its mass on probabilities of essentially 0 or essentially 1.

---
title: "Simulations"
subtitle: "Checking what you believe before the data has a chance to flatter you."
body-classes: hub-prior-sims
listing:
  contents: posts
  include:
    categories: "prior-sims"
  sort: "date desc"
  type: default
  date-format: "YYYY-MM-DD"   # match the post pages; see blog.qmd
  # partial, never true — a full feed embeds each post's rendered body,
  # figure JSON included. See CLAUDE.md § Listings, feeds, and page weight.
  feed:
    type: partial
  fields: [date, title, description, categories, reading-time]
toc: false
---

A prior is a claim about what data is plausible. That claim can be simulated,
looked at, and rejected — before any real observation is involved. These posts
do that, and also take the wider case: run a method against assumptions it does
not hold and see precisely where it bends.