A mainstay of current multiple sclerosis (MS) clinical trial conduct is the comparison of randomized groups of patients using experimental therapy and a placebo. If placebo must be used, it may be possible to limit the number of subjects exposed to placebo by unbalanced randomization, in which fewer subjects are randomized onto placebo than onto experimental therapy. It could, in theory, be possible to replace a placebo arm with a treatment arm in which patients are exposed to a considerably lower dose of active therapy that is not expected to be maximally effective, but is expected to show some benefit. Creating a virtual placebo cohort using extant data from natural history and placebo-controlled studies could reduce the need for, or even replace, placebo groups in future studies. Clinical trials in MS have traditionally been designed with a frequentist approach to statistical inference. An alternative statistical inference approach uses Bayesian procedures.