OT: Nate Silver breaks down the remaining teams

[jdubs30]

http://fivethirtyeight.blogs.nytime...ulf-coast-and-the-rest-of-the-final-16-teams/

IU's chances to win the title have decreased by 8%, ours have gone up by 3%. He seems to like us to win on Thursday

 

[moqui]

" if the Orange are back in the early-season form that saw them start the season 18-1 and beat Louisville, they will make a very tough opponent for Indiana."


in other words: if Syracuse is good then they might win. that is a gem of statistical analysis right there.

 

[bballbeadle]

I love Nate Silver. He is dispassionate and extremely intelligent. He reminds me a little of Spock. This article soothed me, and that's what I need at the moment.

 

[dasher]

I love Nate Silver. He is dispassionate and extremely intelligent. He reminds me a little of Spock. This article soothed me, and that's what I need at the moment.

me too

 

[im4cuse]

Not a fan. I think his 15 minutes are about up and especially not impressed by this "analysis", as he's obviously never actually watched the teams play. This isn't a cut an dry choice as an election. Drawing any conclusion about our 1st round game is laughable.

 

[bnoro]

Not a fan. I think his 15 minutes are about up and especially not impressed by this "analysis", as he's obviously never actually watched the teams play. This isn't a cut an dry choice as an election. Drawing any conclusion about our 1st round game is laughable.

apparently you hadn't heard of him before the election. He's damn good at this.

and his conclusions are drawn from statistics, i don't know why you care so much about "disproving" his analysis. If you don't like it go watch Dickie V ramble about nothing and pick teams based on the amount of PTPers they have.

 

[Toga]

Good stuff

Not a fan. I think his 15 minutes are about up and especially not impressed by this "analysis", as he's obviously never actually watched the teams play. This isn't a cut an dry choice as an election. Drawing any conclusion about our 1st round game is laughable.

As a sports nut I suspect he caught quite a few of the games. But his thing is statistics and he's good at it.

 

[AccuRater]

http://fivethirtyeight.blogs.nytime...ulf-coast-and-the-rest-of-the-final-16-teams/

IU's chances to win the title have decreased by 8%, ours have gone up by 3%. He seems to like us to win on Thursday

His recent book: "The Signal and the Noise" is remarkable. He reviews the history and methodology of prediction of weather, baseball performance, gambling, elections, climate among other topics. A tough, but good read. He is thorough, data driven and logical. Reminds me of our own SWC. I was impressed.

As far as basketball, not so good. At least for this tournament. Of the 48 games so far, he has correctly predicted 33. That's the same success rate as chalk, RPI and SOS.

 

[44cuse]

" if the Orange are back in the early-season form that saw them start the season 18-1 and beat Louisville, they will make a very tough opponent for Indiana."


in other words: if Syracuse is good then they might win. that is a gem of statistical analysis right there.

I think we both know its about 50/50.

44cuse

 

[Flacusian]

Not a fan. I think his 15 minutes are about up and especially not impressed by this "analysis", as he's obviously never actually watched the teams play. This isn't a cut an dry choice as an election. Drawing any conclusion about our 1st round game is laughable.

If you do a little research on his methodologies and partition your opinion from your political ideology there can be no mistaking that Silver is himself a statistical aberration. Certainly, his delving into the NCAA tournament, while still based on statistical analysis, is outside of his particular area of expertise but in matters political he has clearly demonstrated that he has a mastery of prediction based on his own models which others can only marvel at.

He is a unique talent and for myself I wouldn't place as much credence in his picks for the tourney as I would for an election, but his analysis is interesting nonetheless and if you compare his results with those of say... Joe Lunardi, he's doing pretty dang good especially in light of the fact that he concedes he doesn't really know a heck of a lot about basketball whereas it's Lunardi's raison d'etre.

 

[SWC75]

It's all about how close or one-sided the games are. What about match-ups? Akron's guard couldn't handle VCU's press but Michigan's guards sure could.

 

[Knicks411]

Nate's obviously an incredibly sharp dude, but there's so much noise in predicting a single elimination college basketball tournament. (Which of course he's well aware of)

 

[CusefanATL]

" if the Orange are back in the early-season form that saw them start the season 18-1 and beat Louisville, they will make a very tough opponent for Indiana."


in other words: if Syracuse is good then they might win. that is a gem of statistical analysis right there.

i went to the 96 final 4 as a 10 yr old... i was interviewed by paula garrel? from local syracuse tv.

she asked me what syracuse's keys to victory were vs kentucky

cusefanatl: "i think we really need to start in the 1st half and then finish in the 2nd half"

 

[OrlandoCuse]

The Signal and the Noise is recommended reading for anyone that's interested in the art of predicting, and why so many people are really bad at making predictions. While he doesn't really go into the fine details of how to make accurate predictions (the Signal), he does go into a lot of details as to why people get swayed by the Noise (typically because of their biases and lack of proper methodologies).

 

[Fly Rodder]

I like Silver's work and I like that SU's probability of winning is bumped up, but I don't think basketball statistics offer as much predictive power as baseball statistics. Basketball statistics are flawed and skewed and most of the time they don't measure what they purport to measure. There's progress being made, but since so much depends on which way a ball may randomly bounce and human referees (inconsistent foul calls), it's to draw very many specific conclusions. Dumping a bunch of garbage stats (not that I know which ones Nate is using) into a model doesn't magically make the result representative or accurate.

 

[Col. Bleep]

I like Silver's work and I like that SU's probability of winning is bumped up, but I don't think basketball statistics offer as much predictive power as baseball statistics. Basketball statistics are flawed and skewed and most of the time they don't measure what they purport to measure. There's progress being made, but since so much depends on which way a ball may randomly bounce and human referees (inconsistent foul calls), it's to draw very many specific conclusions. Dumping a bunch of garbage stats (not that I know which ones Nate is using) into a model doesn't magically make the result representative or accurate.

Nate ought to stick to politics.
He credits our increase in likelihood of winning to the huge margin of victory against Montana.
I'm sure there are correlations between margin of victory and later round tournament success.
But if he really thinks the big win over Montana has ANYTHING to do with the outcome of our matchup against Indiana tomorrow...then he'd better revise his model.

And may I suggest that on offense he look at free throw shooting, wasted possessions and percentage of possessions in which guards got into the lane with the ball? Those factors will go a long way to determine the outcome.

 

[BuffOrange]

Nate ought to stick to politics.
He credits our increase in likelihood of winning to the huge margin of victory against Montana.
I'm sure there are correlations between margin of victory and later round tournament success.
But if he really thinks the big win over Montana has ANYTHING to do with the outcome of our matchup against Indiana tomorrow...then he'd better revise his model.

How does that make any sense? Every good model (kenpom, sagarin) uses margin of victory. Bad ones like the rpi do not.

 

[Cusefannotindc]

Nate Silver is a witch.

 

[PhatOrange]

" if the Orange are back in the early-season form that saw them start the season 18-1 and beat Louisville, they will make a very tough opponent for Indiana."


in other words: if Syracuse is good then they might win. that is a gem of statistical analysis right there.

Of course we can say - no sheet we have to play well to win, stating he obvious, but If we play like we did against Cal, which was real bad, we have no shot at winning at all. We'll be run off he court