The Importance of Base Rates
An interactive Bayesian eikosogram
Posterior probability: c(H | says H)
0.90
Prior: c(H)
0.50
Evidence reliability
0.90
Split into sensitivity & specificity
How to read this diagram: The total square represents all possibilities.
The width of the left column is c(H) — the prior probability of the hypothesis.
Within each column, the height of the top region is the probability of a positive report
(says H) given that column's truth value.
The highlighted regions are where the evidence says H.
The posterior c(H | says H) is the ratio of the
dark blue area (H true and says H)
to the total highlighted area (everything that says H).