Heh... I saw the "credits" you listed as evidence against Silver. A Corporate Attourney and Jonah Goldberg, a well-known conservative hack. Both of them with as much knowledge in statistical analysis as I have in brain surgery
Look: Just to explain the reasoning in a placative example:
I start "ObamaRoxPolling (ORP)". No matter what the poll, Obama ends up with 100% of the votes.
Now, would Nate Silver have to include me in his analysis? He would be a fool if he did, and it would damage the accuracy of his model. It's not that Nate Silver and his model would "have a problem with ORP", it's just that he doesn't deem my numbers to be credible. And it's entirely up to him to make this decision.
So, if he believes that his model works better without Rasmussen, that's entirely within his prerogative to omit him. Whether or not he's "more correct" remains to be seen. Rasmussen has Romney+1 nationally. Silver has Obama+2.6