The outcome you're trying to predict is how people will vote, and you're free to go and ask them how they will vote.
True, but what Nate Silver brings to the table goes far beyond that. Polls are inaccurate. His statistical models adjust for all sorts of information you don't get from polling. Like which way undecideds will swing late (towards the incumbent) or sampling bias (poll respondents lean Republican) or even outright lying ("sure I'd vote for a black president", then doesn't in the privacy of the voting booth.) That's what set Silver's predictions far above all others in accuracy, and how he correctly captured the election swinging decisively for Obama as hardly anyone else foresaw.
True, but what Nate Silver brings to the table goes far beyond that. Polls are inaccurate. His statistical models adjust for all sorts of information you don't get from polling. Like which way undecideds will swing late (towards the incumbent) or sampling bias (poll respondents lean Republican) or even outright lying ("sure I'd vote for a black president", then doesn't in the privacy of the voting booth.) That's what set Silver's predictions far above all others in accuracy, and how he correctly captured the election swinging decisively for Obama as hardly anyone else foresaw.