Basically the same sort of reasoned assumptions as Opta then, but without the computers.
They assume that past numbers will dictate the future, with no allowance for the variable performance of humans under pressure or the luck of the bounce. So no matter how many times they run the model they...