Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
Yep. Romney down to a 22.6% chance of winning. Latest update lowered Romney's chances due to polling in OH favoring Obama and fact that Silver's model weights OH more so than other States. He's fallen steadily for the last 19 days. I suspect as we get closer to election day, Nate Silver's forecast will merge with his "nowcast" and Romney will have a 15% chance of winning. This is all on cruise control now, no debates, no big news, nothing to move the needle.
Location: On the "Left Coast", somewhere in "the Land of Fruits & Nuts"
8,852 posts, read 10,461,442 times
Reputation: 6670
BTW, Nate Silver happened to be on PBS this evening interviewed by Charlie Rose. And unless he's a heckuva good actor, its pretty clear that Silver is basically just your typical math geek, who truly loves statistics and data of all kinds, including everything from historical and economic info, to sports stats and polling data. And as the title of his book suggests, "The Signal and the Noise", its all about being able to digest all that and recognizing the larger (and more consistent) trends, without getting mislead by the "data du jour" or dramatic changes (aka, the "noise"). And consistent with that theme, he also pointed out that while it is a close race, and there have been fluctuations, the overall positions between Obama and Romney have still stayed remarkably stable since June.
Rose also bluntly asked him about the impact of his own biases, which he doesn't deny, admitting he's probably closer to a liberal-libertarian, falling somewhere closer to Gary Johnson, although he also said he doesn't vote since he started working for the NYTimes. But to Silver it's still basically about "doing the math", and just as he predicted the GOP would have a much larger win in 2010 than the polls were indicating, this time he says all the data continues to favor Obama... not unlike a team that's heading into the last quarter, albeit ahead by 3 points in a very close game.
I follow 538...been following it for some time from back before Silver affliated with the NYT.
I also like the "13 Keys to the White House" method (which, tho it looks sort of subjective, has some mathemetics called "Operations Research" behind it). The 13 Keys method predicted a Kerry defeat back in 2004, too.
The 13 Keys, for me, does the intial prediction, and I follow 538 to see how that prediction plays out. Silver does a good job aggregating polls, the way "Polling Report" used to, but with neat graphics. Ive been liking his state by state analyses, too. Nice job.
As for this anti-Silver rap....to be honest the Right right now reminds me of the Dems/Liberals in 2004, grasping at straws and incredulous that W could possible be re-elected, and predicting ruination if he did (you saw this a lot at the partisan site Democratic Underground). So I've seen this all before and sort of shrug it off.
Just saw an update to Obama 77.4% vs. Romney at 22.6%. A five point jump. I wonder what poll came out that caused that. He mentioned nothing in today's blog post. Seems like RCP is holding steady or minimally moving Romney.
I do agree that the probabilities are a good way to describe it. If he says Obama is up 75% and he loses, well in one out of four cases that would happen. Hardly a slam dunk forecast. If he said 99% and Obama lost big time, people would rightly wonder what he missed. Personally, I think it will go just about as forecast. Obama has a minor structural lead in the swing states, and he has a largely occult hispanic advantage in NV, CO, and possibly FL that will be seen on election day. Both will be hard for Romney to erase this late in the game, with so few people undecided. I could well be wrong, but that is my read of the patterns, data I see at this point. At some point, unless we see a major change, the run to Romney will be too little too late.
Yesterday's OH polling.
Also, the fact that as we approach the election, there is less time for things to change. A glance at Intrade shows Obama climbing over 65% this morning, too. Unless Romney can take poll leads in states totalling at least 269 votes, we're probably going to see Obama's %s continue to increase at the number-crunching, betting and market sites between now and election day.
Also, the fact that as we approach the election, there is less time for things to change. A glance at Intrade shows Obama climbing over 65% this morning, too. Unless Romney can take poll leads in states totalling at least 269 votes, we're probably going to see Obama's %s continue to increase at the number-crunching, betting and market sites between now and election day.
Thanks. I know it had to be something going toward Obama, but there was no accompanying blog post.
BTW, Nate Silver happened to be on PBS this evening interviewed by Charlie Rose. And unless he's a heckuva good actor, its pretty clear that Silver is basically just your typical math geek, who truly loves statistics and data of all kinds, including everything from historical and economic info, to sports stats and polling data. And as the title of his book suggests, "The Signal and the Noise", its all about being able to digest all that and recognizing the larger (and more consistent) trends, without getting mislead by the "data du jour" or dramatic changes (aka, the "noise"). And consistent with that theme, he also pointed out that while it is a close race, and there have been fluctuations, the overall positions between Obama and Romney have still stayed remarkably stable since June.
Rose also bluntly asked him about the impact of his own biases, which he doesn't deny, admitting he's probably closer to a liberal-libertarian, falling somewhere closer to Gary Johnson, although he also said he doesn't vote since he started working for the NYTimes. But to Silver it's still basically about "doing the math", and just as he predicted the GOP would have a much larger win in 2010 than the polls were indicating, this time he says all the data continues to favor Obama... not unlike a team that's heading into the last quarter, albeit ahead by 3 points in a very close game.
Now about 2/3rds through the book. VERY interesting stuff. He seems to love chess, baseball, backetball, economics, meteorology, seismology, and politics, but especially poker. It is in large part because poker involves an ability to weight different possible outcomes. He seems to be pretty dismissive of the sort of braying, argumentative, my cause is right just because approaches to political punditry.
If you think about it, most pundits are poly sci., english, or law students, who have learned to carve out narratives and argue persuasively for their points. They are often wrong, and not held to account by their partisan audiences. He knows his predictions are going to be better than most of theirs, and they know it too. They hate Silver because he threatens them. The guy is a super math nerd, but actually quite articulate as well. I'd rather listen to someone like that than to hear Pat Buchanan blathering on.
One thing to remember about Silver is that he doesn't dictate odds. He created a model months ago, feeds data into it, and percentages result. He does adjust polls based on established bias and deviation from the norm.
Example:
On Tuesday, two of the three OH pollsters have shown a consistent Democratic bias. Silver thus downgrades the influence these polls have on the model.
But what he is not doing is saying "Today, I think Obama's chances are 68.9%!" and thus posting that. It doesn't work that way.
i am sure a lot on this board would consider me a classic rwnj. I wear the moniker with pride.
I have never said a bad thing about Sliver so maybe im losing cred?
I think the race is closer than Silver says but i still think Obama is the odds on favorite. it is just very hard to unseat a president.
I like vertually nothing about Obama but I do think he has a better chance of winning than not.
You are being rational. Be careful. That would disqualify you from the "nj" part of the rwnj.
In any case, the odds are 3 to 1 on Nate Silver and 2 to 1 on Intrade, both favoring Obama. I think the biggest advantage to Obama is that the structural advantage in the swing states has proven pretty solid through the season, at least by the polls. We'll see if that is indeed the cast next week. Romney still has a very real chance.
Ultimately all these statistical models are hampered by the paucity of data. There are only so many elections where the candidates and electorate parallel today, so statistical relationships will tend to be weaker. It is not a slam dunk either way.
i am sure a lot on this board would consider me a classic rwnj. I wear the moniker with pride.
I have never said a bad thing about Sliver so maybe im losing cred?
I think the race is closer than Silver says but i still think Obama is the odds on favorite. it is just very hard to unseat a president.
I like vertually nothing about Obama but I do think he has a better chance of winning than not.
Silver isn't saying the race isn't close. In all of his interviews he compares it to a 3 point basketball game in the final seconds. Odds are, 75% of the time, the team with the 3 point lead wins.
There can be a statistical anamoly, the "any given Sunday" effect, but the odds are, Obama will win.
I'm giving it 60/40 odds personally for the President.
The Quinipiac polls in Ohio don't look good for Romney.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
