Frequentist Persuasion — Posterior Distribution
Frequentist Persuasion Distribution of posteriors given m = guilty
Parameters
μ₀ = Pr[guilty] 0.30
z (lying prob) 0.25
Pr[m = guilty | θ = innocent]
N (sample size) 8
Summary
z^KG
—
μ₀ / (1−μ₀)
Pr[convict] = c_N(z)
—
weight of points with posterior ≥ ½
π(z) = Pr[m = guilty]
—
V_N(z) = π · c_N
—
Each point = one (X,Y) pair with X+Y > 0 · x-axis: posterior X/(X+Y) · y-axis: Pr[X=x, Y=y]