The anti-Silver theory you have doesn’t match up with the facts.
Silver never said anything about his expectations for turnout levels and they were irrelevant for his model. The model did predict what he thought the percentages of the total vote would be for both candidates and he was basically correct about that.
Leaving that out of it, if nothing else comes out of the work Silver does, the basic argument that should come out is that if one candidate leads in virtually all reliable polls (and Silver has a clear, non–partisan methodology for determining what he views as reliable), as Obama was in Ohio, that candidate is most likely leading even if all of those polls are within the margin of error. Statistically this statement should not be controversial, but for some reason it is.
It’s also worth noting this is the third election he’s worked on (he also made predictions in 2010 on the Senate and governor races) and so far he’s only been wrong 7 times on the state level (2008: Indiana Presidential, 2010 Alaska, Colorado and Nevada Senate and Illinois Governor, and 2012 North Dakota and Montana Senate). Every one of those times he’s picked a Republican who lost.