He even got the Germany versus Brazil World Cup wrong.
Not a soccer fan, but it looks like the data is showing him to be just another prognosticator.
Stats fail when non-quantitative data is at play. Home field advantage, prior record, touches on the ball, hours of travel, can be counted, but those do not account for lack of rest, improper diet, head games, adrenaline, team chemistry, desire, ambition, etc.
That's what makes contests such as matches, elections, exciting. Silver is way out of his league and shows that analytic models have limitations, especially when the analyst builds in bias!