Rereading John Platt's "strong inference" paper I'm once again struck by how relevant it is half a century later. Every scientist should read it.