It was no use relating reports to probability when there was no quantum of knowledge from which to start.