Former Daily Kos blogger Nate Silver turned heads with his Obama-friendly election predictions in the New York Times, but CNN's Soledad O'Brien thinks his conclusions show no bias. Of course, the liberal CNN anchor just might have a blind spot for poll numbers favoring Obama.

"Nate Silver is very careful about focusing on the numbers. And he doesn't have a liberal bias in his calculations, which I think is why a lot of people follow what he has to say," O'Brien declared on Tuesday's Starting Point.

New York Times star poll analyst Nate Silver continues giving hope to Democrats, and he's getting more confident in an Obama victory as the election draws closer, pegging Obama's odds of victory at around 75%. After a heated debate on MSNBC's Morning Joe, the normally mild-mannered Silver offered via Twitter on Thursday to bet host Joe Scarborough $2,000 that Obama would win, which drew some criticism from the paper's outspoken new Public Editor, Margaret Sullivan. Meanwhile, columnist Paul krugman termed conservative criticism of Silver's methodology "scary."

Silver, a former poster at the left-wing Daily Kos, who usually mans the Five-Thirty-Eight blog at nytimes.com, again made the paper on Thursday with "When State Polls Differ From National Polls," which asserted that Barack Obama will probably win both the Electoral College and popular vote:

One thing that you have to admire about Nate Silver is that he isn’t afraid to go out on a limb.  As an example of that, the New York Times political soothsayer currently projects Barack Obama winning the popular vote by 1.7%. 

That would place him well on the left side of most current polling.  The below histogram shows the distribution of the spread between the two candidates in each of the ten polls that make up the RCP Average on the day of his prediction, October 29.

Tarnished Silver? The New York Times's young star pollster Nate Silver got some guff last week for dismissing Mitt Romney's large leads in the Gallup tracking poll.

In an October 18 post on his FiveThirtyEight blog at nytimes.com, "Gallup vs. the World" (it also appeared, heavily edited, in print) Silver claimed the Gallup poll was overrated and "its results turn out badly" when it's an outlier, noting that in 2008 it "had a four-point miss," predicting an 11-point win by Obama that turned out to be a seven-point margin.

Guess what other big-time poll had Obama pegged as an 11-point winner in 2008? The New York Times-CBS News poll. Though to be fair, in 2008 Silver was not with the Times but writing for his own blog after cutting his political teeth at the left-wing blog Daily Kos (Silver calls himself a "rational progressive.")

The Washington Post ran a story slamming pollster Scott Rasmussen on Thursday on the front page of the Style section. Political reporter Jason Horowitz earnestly channeled the Democratic spin from the story's beginning:

ASBURY PARK, N.J. -- Here is a fun fact for those in the political polling orthodoxy who liken Scott Rasmussen to a conjurer of Republican-friendly numbers: He works above a paranormal bookstore crowded with Ouija boards and psychics on the Jersey Shore.

Here's the fact they find less amusing: From his unlikely outpost, Rasmussen has become a driving force in American politics.

Democrats surely dislike how Rasmussen's polls (like this week's showing Harry Reid losing by 11 points) affect the optimism of their donors and activists. But are his numbers accurate?