Each observation changes the confidence assigned to three hypotheses. The leading belief gains visual resolution. Alternatives remain present while their probability is still material.