Nate Silver on the Oscars

As I'm sure you know, I'm a total politics junkie and die-hard Keith Olbermann fan. So needless to say, I've got a huge geek crush on Nate Silver of FiveThirtyEight.com, the most brilliant statistical mind of our time.

Nate's original claim to fame was developing the PECOTA system for accurately forecasting baseball players' career development and performance. He became a real soothsayer of the baseball world. Last year, he switched his prediction model to politics, and predicted that Barack Obama would be the Democratic nominee long before any other pundit had abandoned Hillary Clintion. He started FiveThirtyEight.com (a geekingly beautiful site that makes wonkitude cool!) and with the same precision, accurately called every US Senate race and 49 of the 50 states in the presidential race (sorry, Indiana!).

So if he can accurately call baseball and politics, can his model work for the Oscars? New York magazine asked Silver to apply his genius in computer-based logical regression to Hollywood, and he makes some very interesting picks. http://nymag.com/movies/features/54335/

So who does this statistics guru like for each category?

Leading Actor
Mickey Rourke: 71.1%
Sean Penn: 19%
Brad Pitt 5.9%
Frank Langella 3.4%
Richard Jenkins 0.5%
Rourke and Penn split the two awards that traditionally predict success (SAG and Golden Globes). As a result, Silver gives the advantage to Rouke, since Penn has already won an Oscar. Silver's statistical model has shown that once an actor wins, his odds go way down, whereas someone who's been nominated without winning sees their odds increase.

Leading Actress
Kate Winslet 67.6%
Meryl Streep 32.4%
Anne Hathaway 0.0%
Melissa Leo 0.0%
Angelina Jolie 0.0%
Although Streep has won most of this year's big awards (including SAG), she's already won an Oscar, whereas Winslet is "sitting on her sixth nod without a win." (Yes, those are actual zeroes that Silver gave Hathaway, Leo, and Jolie. Very sad.)

Best Picture
"Slumdog Millionaire" trounces its competitors with a 99% likelihood.

Supporing Actor
Heath Ledger 88.5%
Josh Brolin 5%
As Harry Shearer said on the Olbermann show the other night when asked if he was upset to lose the Grammy for best comedy album to the late George Carlin, awards show voters like dead people. So even if Ledger's performance wasn't amazing (and it was!), he'd still probably have a lock on it.

Supporting Actress
Here's where it gets tricky...Kate Winslet was categorized as a supporting actress in the other major awards groups, but the Oscars gave her a leading actress nod. As Silver put it, this "is not so nice for our computer."
Taraji P. Henson 51%
Penelope Cruz 24.6%
Viola Davis 11.6%
Silver predicts that Henson will win the Oscar because although Cruz would be the logical defualt after taking Winslet out of the category, because Henson is the only actress in the category in a picture nominated for Best Picture ("Benjamin Button"), the model gives her the nod because her film "is the only best picture nominee with a supporting actress nod, and best picture nominess tend to have an edge in other categories."

So what do you think? Will Silver keep up his streak of big predictions? Personally, I want him to make more predictions. All the more reason to get a big giant geek fest of statistical geekitude on "Countdown with Keith Olbermann."

So who do you think will win the major awards? Why?