Ezra makes an immaculate free-market argument here, pointing out that Silver’s influence is based on his excellent track record and that if someone came along with a better formula for predicting election outcomes that person would supplant Silver as the go-to numbers geek. If you don’t like Silver, then build a better mousetrap.
Didn't some prof at the University of Colorado (with just as good of a track record) predict a Romney win?
The Lefties have clung to him like he was Salvation itself (witness shilol), but seem incapable of doing any independent analysis on their own.
Mark O said...
I know Nate. He's from Reno by way of Dodge City.
He ended up as a wrangler on the Ponderosa, before that he was one of Matt Dillon's deputies and went out to the Coast as an outrider on Seth Adams' wagon train.
Bob Boyd said...
Did I ever tell you the one about Nate the Talking Snake who guards the lever that stops the world?
Silver has a good track record. But there is a lot of similarity to people who have a good track record picking stocks. Just because you are good for a short time doesn't mean you will be good for a long time. Predicting a national election is a matter of predicting the aggregate of millions of personal decisions based on local conditions. You are basing your decision on polls with 9% response rates. Good luck. I'm going to wait until next Wednesday morning.
People are poking holes in Silver's method by raising interesting points. Also it is odd to call Silver a predictor because keeps changing his prediction up to the last minute. His track record in 2010 was not great. He picked most of the swing states in 2008, only missing on Indiana. The day before Election Day, the only states in doubt were NC, MO, and IN. So Silver got 2 of 3 right. If Romney wins decisively, then Silver will have been way off, just like he was in 2010 when he gave the Republicans a 30 percent chance of winning 50 House seats. They won well over 50, so not only were his odds wrong on 50, he substantially underestimated the number. So his prognostication skills aren't that great.
I was told by some MLB experts that the Tigers would win the AL Division playoffs because they had the best pitching rotation for a short series.
Lots of stats, & these guys have more expertise than I, but I just didn't believe.
Then when the Tigers beat the Yankees, I was again told by some experts that the Tigers would win the World Series because they had the best pitching rotation for a short series & moreover, the staff was well rested.
Lots of stats & now I believed.
OOPS, SF wins the WS.
Could someone help me here?
(One version: The Tigers pitching staff was OK with the Yankees' HR crew but not with SF's punch & Judy crew. What Sandavol & Three HRs? Nevermind!)
My feeling is that it's 100% certain that Nate's conclusions are based on faulty polls with D+whatever & GIGO.
But that as with the Tiger's pitching staff, he has a good chance of being right & a good chance of being wrong.
Nevertheless, because I can't say that Nate has, say, a 76.98756% chance of being wrong, some may conclude that I'm not as scientific as Nate.
Actually I feel, repeat feel that this "a splendid example of sham science and spurious specificity”, along with “a penchant for pompous mystification”.
Hey, if one ignores the margin of error or D+5-9 in Silver's "swing states" & assumes that most or all of them will go to Obama, of course Obama wins. And shouldn't he win by over 300 Electoral votes?
Finally, I'm 66.666% certain that my feelings are less simplistic than Nate's.
" If Romney wins decisively, then Silver will have been way off, just like he was in 2010 when he gave the Republicans a 30 percent chance of winning 50 House seats. They won well over 50, so not only were his odds wrong on 50, he substantially underestimated the number"
If I tell you the coin has a 50% chance of landing tails... And it lands tails...I did not substantially underestimate that probability.
So, basically, what Silver does is run hundreds of models and then tells us what most of his models said? So... he does what every other poll aggregation site does, but with extra snark?
This is why I'm not impressed with Silver; he doesn't -do- anything unique or risky. He just does the same work everyone else does, but then disguises it with some voodoo and claims loads of credit when he is somewhat right.
He who has the gold determines the model. So on the list of things that determine the way a model works, long term results proven accuracy is well below things like money, cocktail party invites and sex.
Sadly for Nate, his tweeting and his other obnoxious responses, are setting him up for a serious hubristic beating on Nov 7 if he is wrong. The left may have no history but they never forget people who make fools out of them.
The coin analogy doesn't work. If Romney wins decisively, It's more like Silver says the 2005 White Sox have only a 30 percent chance of beating the Red Sox in the ALDS, and then the White Sox sweep the Red Sox take the ALCS 4-1, and then sweep the World Series. It's one thing to be just off. It's another thing to be dead wrong. Silver was dead wrong in 2010 (and 2005).
Nate Silver had the best predictive model in 2008 I understand - I've read that claim numerous times and believe it to be true. He was about average in 2010. So this experienced and competent statistician is either one of the best and proved it in 2008 and just had an off year in 2010 and will once again be one of the best this year - OR - he is just a leftist (is with the NYT after all) and was right in 2008 by making it look like he hoped it would be, was off in 2010 because the election went against him and will be further unmasked this year by maintaining his outlier predictions of a fairly easy Obama win. So this time next week, we'll all know which one he is.
Meanwhile I notice that he uses the 2008 breakdown of party lines and very little of 2010 thrown in - something that is true of any aggregator btw. So for his model to be true this time rather than partisan, he is saying that 2010 has only the slightest influence on the 2008 voting patterns and that the tea party has faded away. I think that is wishful thinking rather than a real statistical approach. I can hardly wait to see how it actually plays out.
Silver made his bones in '08. Turns out he was getting inside polling info(i.e. gold standard polling) from the Obama campaign against which to check his assumptions. Remember the great polling guru John Zogby who was almost dead-nuts on in 2000? Another one-hit wonder...
I am a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for me to earn fees by linking to Amazon.com and affiliated sites.
Encourage Althouse by making a donation:
Make a 1-time donation or set up a monthly donation of any amount you choose:
३६ टिप्पण्या:
I know Nate. He's from Reno by way of Dodge City.
The only thing anyone needs to know about Nate Silver is that he is a former daily kos blogger. Enough said.
"Nate Silver" is clearly a porn pseudonym.
Nate Silver saved my life in 'Nam!
Ezra makes an immaculate free-market argument here, pointing out that Silver’s influence is based on his excellent track record and that if someone came along with a better formula for predicting election outcomes that person would supplant Silver as the go-to numbers geek. If you don’t like Silver, then build a better mousetrap.
Didn't some prof at the University of Colorado (with just as good of a track record) predict a Romney win?
He is a thin and effeminate man...and Im certain he is guilty of witchery.
Did I ever tell you the one about Nate the Talking Snake who guards the lever that stops the world?
"He is a thin and effeminate man...and Im certain he is guilty of witchery."
So he is a metrosexual liberal who is like totally gay.
Blogspot must have cranked up their AT server again this morning!
There's no way to thoroughly discredit someone's math ... Like pointing out their appearance.
The Lefties have clung to him like he was Salvation itself (witness shilol), but seem incapable of doing any independent analysis on their own.
Mark O said...
I know Nate. He's from Reno by way of Dodge City.
He ended up as a wrangler on the Ponderosa, before that he was one of Matt Dillon's deputies and went out to the Coast as an outrider on Seth Adams' wagon train.
Bob Boyd said...
Did I ever tell you the one about Nate the Talking Snake who guards the lever that stops the world?
You mean better Nate than lever?
"Ezra Klein explains what's going on".
And with that assurance of quality, I promptly closed my browser window.
Silver has a good track record. But there is a lot of similarity to people who have a good track record picking stocks. Just because you are good for a short time doesn't mean you will be good for a long time. Predicting a national election is a matter of predicting the aggregate of millions of personal decisions based on local conditions. You are basing your decision on polls with 9% response rates. Good luck. I'm going to wait until next Wednesday morning.
People are poking holes in Silver's method by raising interesting points. Also it is odd to call Silver a predictor because keeps changing his prediction up to the last minute. His track record in 2010 was not great. He picked most of the swing states in 2008, only missing on Indiana. The day before Election Day, the only states in doubt were NC, MO, and IN. So Silver got 2 of 3 right. If Romney wins decisively, then Silver will have been way off, just like he was in 2010 when he gave the Republicans a 30 percent chance of winning 50 House seats. They won well over 50, so not only were his odds
wrong on 50, he substantially underestimated the number. So his prognostication skills aren't that great.
hey, whatever happened to shiloh and crack emcee? I miss scrolling over their comments.
I was told by some MLB experts that the Tigers would win the AL Division playoffs because they had the best pitching rotation for a short series.
Lots of stats, & these guys have more expertise than I, but I just didn't believe.
Then when the Tigers beat the Yankees, I was again told by some experts that the Tigers would win the World Series because they had the best pitching rotation for a short series & moreover, the staff was well rested.
Lots of stats & now I believed.
OOPS, SF wins the WS.
Could someone help me here?
(One version: The Tigers pitching staff was OK with the Yankees' HR crew but not with SF's punch & Judy crew. What Sandavol & Three HRs? Nevermind!)
My feeling is that it's 100% certain that Nate's conclusions are based on faulty polls with D+whatever & GIGO.
But that as with the Tiger's pitching staff, he has a good chance of being right & a good chance of being wrong.
Nevertheless, because I can't say that Nate has, say, a 76.98756% chance of being wrong, some may conclude that I'm not as scientific as Nate.
Actually I feel, repeat feel that this "a splendid example of sham science and spurious specificity”, along with “a penchant for pompous mystification”.
Hey, if one ignores the margin of error or D+5-9 in Silver's "swing states" & assumes that most or all of them will go to Obama, of course Obama wins. And shouldn't he win by over 300 Electoral votes?
Finally, I'm 66.666% certain that my feelings are less simplistic than Nate's.
" If Romney wins decisively, then Silver will have been way off, just like he was in 2010 when he gave the Republicans a 30 percent chance of winning 50 House seats. They won well over 50, so not only were his odds
wrong on 50, he substantially underestimated the number"
If I tell you the coin has a 50% chance of landing tails... And it lands tails...I did not substantially underestimate that probability.
So, basically, what Silver does is run hundreds of models and then tells us what most of his models said? So... he does what every other poll aggregation site does, but with extra snark?
This is why I'm not impressed with Silver; he doesn't -do- anything unique or risky. He just does the same work everyone else does, but then disguises it with some voodoo and claims loads of credit when he is somewhat right.
"I am Nate Silver."
He who has the gold determines the model. So on the list of things that determine the way a model works, long term results proven accuracy is well below things like money, cocktail party invites and sex.
I am not familiar with Nate but I remember his Mom...Hi Ho.
She was a TV star in the 1950's.
She was sort of like the presidents Mom.
Ridden by guys in masks and Indians.
Just sayn'
Sadly for Nate, his tweeting and his other obnoxious responses, are setting him up for a serious hubristic beating on Nov 7 if he is wrong. The left may have no history but they never forget people who make fools out of them.
Kchiker,
The coin analogy doesn't work. If Romney wins decisively, It's more like Silver says the 2005 White Sox have only a 30 percent chance of beating the Red Sox in the ALDS, and then the White Sox sweep the Red Sox take the ALCS 4-1, and then sweep the World Series. It's one thing to be just off. It's another thing to be dead wrong. Silver was dead wrong in 2010 (and 2005).
On Myth Busters Halloween special, they hosted an experienced odor scientist who sniffed out the scent of fear on collected sweat samples.
...wonder what she'd smell if she sniffed a scent of silver.
Yeah, Nate Silver sounds like a made up name, however, did you know that Neil Diamond's real name is Neil Diamond?
Yeah, Nate Silver sounds like a made up name, however, did you know that Neil Diamond's real name is Neil Diamond?
Nate Silver had the best predictive model in 2008 I understand - I've read that claim numerous times and believe it to be true. He was about average in 2010. So this experienced and competent statistician is either one of the best and proved it in 2008 and just had an off year in 2010 and will once again be one of the best this year - OR - he is just a leftist (is with the NYT after all) and was right in 2008 by making it look like he hoped it would be, was off in 2010 because the election went against him and will be further unmasked this year by maintaining his outlier predictions of a fairly easy Obama win. So this time next week, we'll all know which one he is.
Meanwhile I notice that he uses the 2008 breakdown of party lines and very little of 2010 thrown in - something that is true of any aggregator btw. So for his model to be true this time rather than partisan, he is saying that 2010 has only the slightest influence on the 2008 voting patterns and that the tea party has faded away. I think that is wishful thinking rather than a real statistical approach. I can hardly wait to see how it actually plays out.
Too bad Nate Silver's statistics can't smell the coffee.
How does calling one election constitute an excellent record?
How does calling one election constitute an excellent record?
Wow, it's Dan Quayle in reverse.
(Lloyd Bentsen, technically.)
Silver made his bones in '08. Turns out he was getting inside polling info(i.e. gold standard polling) from the Obama campaign against which to check his assumptions. Remember the great polling guru John Zogby who was almost dead-nuts on in 2000? Another one-hit wonder...
टिप्पणी पोस्ट करा