Topic: Thompson Sampling for Trials

Posterior Probabilities That Treatments Are Best

Posterior Sampling With Nonconjugate Response Models

Minimum Allocation Floors and Renormalization

Monte Carlo Error in Estimated Allocation Probabilities

Implementing Posterior Sampling for Trial Allocation

Handling Ties in Posterior Best-Arm Selection

Evaluating Best-Arm Probabilities for Two Beta Posteriors

Expected Arm Counts Over a Fixed Trial Horizon

Efficient Computation Across Many Treatment Arms

Eliciting Beta Priors From Clinical Information