It sounds to me as if he was arguing in favour of basically looking at a balance of the probabilities, whereas the standard currently being applied would be sort of a reversed version of beyond a reasonable doubt.
Does that sound like a rough way of describing the two?