Could it be that, realistically, 4 or 5 horses in a race only have a combined 1 or 2 percent chance of winning? If this is the case then the book would still 'approximate' 100%.