---
title: "Declared Priors versus Fitted Posteriors"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Declared Priors versus Fitted Posteriors}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(collapse = TRUE, comment = "#>")
```

The bridge uses the backend-independent prior specification retained by
`gp3bayes`; saved backend prior draws are not required.

```{r, eval=FALSE}
bridge <- prior_posterior_bridge(fit)

prior_posterior_summary_table(bridge)
prior_posterior_distance_table(bridge)

plot_prior_posterior_density(bridge)
plot_prior_posterior_intervals(bridge)
plot_prior_posterior_shift(bridge)
plot_prior_posterior_contraction(bridge)
```

Shift, contraction, interval overlap, Kolmogorov-Smirnov distance, and
quantile-based Wasserstein distance are descriptive marginal summaries, not
automatic measures of prior adequacy.