Interpreting a survival curve I