A reanalysis of Lord's statistical treatment of football numbers