Baseball statistician tries politics and bats almost one thousand | Hacker News Reader

Baseball statistician tries politics and bats almost one thousand

nytimes.com

67 pointsjacobscott24 comments

This is hilariously awesome:

When we took him to preschool one time, we dropped him off, and he announced, ‘Today, I’m a numbers machine,’ and started counting, Brian Silver said. When we picked him up two and a half hours later, he was ‘Two thousand one hundred and twenty-two, two thousand one hundred and twenty-three...’

Sounds like an early sign of autism...

My second thought (after the laughter) - but then I think it shows how much we want to ascribe "problems" to kids. It's obvious he turned out okay and it's sad to think that under other parents he may have been diagnosed rather than encouraged.

Yeah. It surprises me, when I hear stories about parents overreacting about kids doing weird things. They're kids. They love experimenting with weird stuff like that. Hell, some people keep doing it when they get older - they're the fun sorts of people.

Can anybody here honestly say they didn't do anything weird as a kid? I remember reciting numbers from the Guinness Book of World Records into a tape recorder, in the belief that any number I said was the money I'd make. (I also think I did that with computer pixels on a low-res screen, for that matter.) I'm sure almost everybody here did strange stuff as a kid.

Let's give people who actually make austim and Asperger's diagnoses more credit.

Being obsessed with and good at maniuplating numbers and rule bases systems, and consequently being better at it than peers, is a sign of autism and Asperger's. But one sign is not enough for a diagnoses. If he exhibited no communication and socialization problems, he would not fit the description.

It does but he seems a little too personable and maybe even jock-ish.

Silver predicts White Sox lose 90 games. Is ridiculed in baseball circles. Vindicated when they end up sucking.

Silver predicts Rays win 88 games. Is ridiculed further. Rays win 97 games and go to the World Series. Absolute vindication.

That said, I hate his player projection systems. You can't predict baseball (though he comes about as close as you can).

The GM of the Arizona Diamondbacks is a stats whiz as well:

http://www.azcentral.com/sports/diamondbacks/articles/2008/0...

http://www.usatoday.com/sports/baseball/nl/diamondbacks/2007...

This year wasn't an accident.

Why don't you like player projection systems? While it's impossible to accurately predict almost anything as complex as baseball consistently, players projections are an awesome tool.

I can understand why people like them. I don't say they suck and they're of no use. I just don't like them personally. Not really sure why. I think it's because the line "you can't predict baseball" keeps running through my head. PECOTA is as accurate as they come, though.

Thanks for posting this. A very enjoyable article that demonstrates how the broader application of entrepreneurial behaviors to a problem, that Nate clearly has a passion for, can blow the incumbents clean out of the water.

I doubt we've heard the last of Nate Silver.

By the end of [election] night, Mr. Silver had predicted the popular vote within one percentage point, predicted 49 of 50 states’ results correctly, and predicted all of the resolved Senate races correctly.

From his posted predictions back in March? W-o-w.

No, his site maintained an updated assessment of the current probability. I'm not really sure what the article means when it says he 'called it' in March. He was probably just showing Obama with a greater chance of winning.

He showed Obama with a higher percentage except for immediately after Palin, and then he explained that it was a bump that promised to diminish.

What impressed me about his site isn't just his predictions, but how good he was at predicting how his model would change in the future. He could predict how his model would change every day for a week in advance, and he was almost always right as to how the "market" would react.

Uh, no. The final result was not predicted from March's data. For one thing, remember how the polls changed in October when the Economy hit the fan?

Doh. Yes that would make more sense. The article was probably referencing his day-before predictions. Thank you for the correction.

Um, I (and the whole world) called it for Obama before 10pm that night. It seemed clear to me that the "incumbents" must have been legally required (or something) to wait till the polls closed on the west coast before they could make it official.

And lots of people called the election months before too. I followed http://electoral-vote.com/ just because I followed them the last two elections. The job of prediction they did was comparable. Though FiveThirtyEight.com did seem to have more/better features.

Yeah, but he had it predicted back in March, and by two weeks before the election he had nearly every state accurately predicted.

I was mostly commenting on the fluff at the start of the article about calling the election at 9:48 EST.

Agreed. That was silly.

Agreed, if this guy had called the election in 04, or a more difficult call in '00 I would be impressed. As a democrat who saw 1 election stolen and 1 botched by Kerry, even I was confident in falling asleep early knowing we'd pull this one out.

He didn't call the election in '04, but one of his co-workers did call Obama a future president back then, after his convention speech.

Does Nate Silver invest on Intrade (or another prediction market) and make money off of his predictions? If not, why not?

There's not really a whole lot of liquidity on the majority of the state markets. Difficult to make money when the bid-ask spreads are so large.

Reply on news.ycombinator.com