Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Nate Silver of FiveThirtyEight Blog Is to Join ESPN Staff

nytimes.com

Post reply on HN
Print view View original

61–67 of 67 posts

Re: Nate Silver of FiveThirtyEight Blog Is to Join ESPN Staff

#61

Post by blatherer »

markbao wrote: post #15

I think it's a waste of Nate Silver's intelligence to work on something like sports , when there are many other bigger problems to potentially be solved with the help of statistics and prediction. Why might ESPN be a good option for him? Can someone help me see the other side of this? It's totally his decision to make, and I'm sure he's looked at his options rationally, and life circumstances and past experience and…

his major achievements to date have been in predicting player performance in baseball, an area which he has spent a large part of his life analyzing. Baseball analytics rewards attention to detail and persistence over technical brilliance/advanced statistical techniques. his special ability is really very domain specific, so it's not like we should expect him to be much more effective than other qualified people if he applied himself to different topics...

Re: Nate Silver of FiveThirtyEight Blog Is to Join ESPN Staff

#62

Post by blatherer »

crapshoot101 wrote: post #53
mccolin wrote: post #47

Earlier quoted context omitted.

I am also a longtime fan of his from his BP days. Here's to hoping ESPN makes great use of his skills. Buried as a contributor to a late night program seems, on the surface, like a situation ripe for underutilization of his talents.

What's funny is that back in the old BP days, you could still argue most of the interesting work in baseball / statistics was being done in the public domain working of standardized data sets - whether it was DIPS via Voros McCracken, or Tom Tango's work, or that of MGL / some of the early BP guys in the past few years. Now, since we're pretty good at capturing offensive value, the really interesting stuff is being d…

what are the best publicly available sources of baseball data?

Re: Nate Silver of FiveThirtyEight Blog Is to Join ESPN Staff

#63

Post by eitally »

markbao wrote: post #30
lbrandy wrote: post #22

Earlier quoted context omitted.

> Can someone help me see the other side of this? Your profile says you dropped out of college and are currently backpacking the world. Is that a waste of your intelligence? Why or why not?

In the same way that I only see Nate Silver's decision from my own perspective, I'm interested in why he might have picked to go to ESPN, because I'm sure that he has his own internal reason for doing so, like I do for traveling. So I'm interested in what that is and why he might have chosen it. From my perspective, I don't think so. Gaining experience about how the world works and learning about other perspectives a…

Because he loves sports, it sounded fun and they offered him a boatload of money? Those seem like perfectly legitimate reasons to me. I don't, and never have, understood why people bash celebrity intellectuals who pursue their own passions rather than the passions critics believe are more valid or important.

Re: Nate Silver of FiveThirtyEight Blog Is to Join ESPN Staff

#64

Post by alohahacker »

If Nate or anyone could program a statistically accurate sports predictor model there would be huge potential in this as far as sports betting goes.

The guys that make millions in large part use their own software.

I used a simple piece of hardware last year for college football season that was suprisingly accurate. I weighted my own opinion most heavily but would use the software to confirm my picks. If both ways checked out, I would place a bet.

Never had a losing week last year and would usually pocket a couple thousand in profit a week. Of course there was alot of lucky but a large part of my success was due to the statistical software I was using.

Re: Nate Silver of FiveThirtyEight Blog Is to Join ESPN Staff

#66

Post by patrickk »

alohahacker wrote: post #64

If Nate or anyone could program a statistically accurate sports predictor model there would be huge potential in this as far as sports betting goes. The guys that make millions in large part use their own software. I used a simple piece of hardware last year for college football season that was suprisingly accurate. I weighted my own opinion most heavily but would use the software to confirm my picks. If both ways ch…

> If Nate or anyone could program a statistically accurate sports predictor model there would be huge potential in this as far as sports betting goes.

Absolutely.

> The guys that make millions in large part use their own software.

I was fascinated reading about Bob Haralabos' incredible tracking of basketball games: http://espn.go.com/blog/truehoop/post/_/id/4974/following-up...

> I used a simple piece of hardware last year for college football season that was suprisingly accurate. I weighted my own opinion most heavily but would use the software to confirm my picks. If both ways checked out, I would place a bet.

Care to elaborate (without giving away your secret sauce?) I love reading about this stuff.

I try to approach soccer betting in a similar fashion (it's trickier than US sports in some ways, as soccer doesn't lend itself to stats as easily as say baseball and American football). See a comment of mine on reddit here, with tons of resources:

http://www.reddit.com/r/SoccerBetting/comments/15dcao/what_v...

Re: Nate Silver of FiveThirtyEight Blog Is to Join ESPN Staff

#67

Post by T-hawk »

cmsmith wrote: post #37

Possible reasons for this move: 1) People only care about election handicapping every 4 (or 2) years in the US. 2) Nate Silver especially made political predictions look really easy (in comparison to the partisan clowns who populate cable news). The outcome you're trying to predict is how people will vote, and you're free to go and ask them how they will vote. Compare to sports where (1) you have sporting events all…

The outcome you're trying to predict is how people will vote, and you're free to go and ask them how they will vote.

True, but what Nate Silver brings to the table goes far beyond that. Polls are inaccurate. His statistical models adjust for all sorts of information you don't get from polling. Like which way undecideds will swing late (towards the incumbent) or sampling bias (poll respondents lean Republican) or even outright lying ("sure I'd vote for a black president", then doesn't in the privacy of the voting booth.) That's what set Silver's predictions far above all others in accuracy, and how he correctly captured the election swinging decisively for Obama as hardly anyone else foresaw.

Post reply on HN