http://news.yahoo.com/nate-silver-joe-scarborough-wanna-bet-...Political polling guru Nate Silver is so confident in his statistical models that he just offered to bet MSNBC's Joe Scarborough $1,000 that Barack Obama will win re-election. Scarborough, you may recall, criticized Silver's math earlier this week, saying that "Anybody that thinks that this race is anything but a tossup right now is such an ideologue ... they're jokes." He was specifically talking about Silver's FiveThirtyEight website, which shows Mitt Romney with just a 1-in-4 chance of becoming president.Silver has spent the week firing back, criticizing political pundits for not understanding how odds and probability work and aggressively defending his method against critics. As the week has progressed, his model has only shown Obama's chances of winning increasing, which has not coincidentally increased Silver's confidence in the outcome. (As of this morning, Five Thirty Eight gives Obama a 79 percent chance of winning, with a final Electoral College total over 300.)The bolded portion is the key question in my mind. Is Silver's prediction based on probability math or is it an effort keep conservative voters home?I guess we find out Tuesday.
My only comment is that models continue to work until they don't. I used to think my cash flow models good and even better as I made them more complex. As time went on they started giving bad projections caused by the model and/or bad input by the users. While we haven't abandoned the model, I've taken out some of the complexity.This is why I take results from "accurate" models with a grain of salt. Remember the old saying: "Past performance is no guarantee of future results."
"My only comment is that models continue to work until they don't. I used to think my cash flow models good and even better as I made them more complex. As time went on they started giving bad projections caused by the model and/or bad input by the users. While we haven't abandoned the model, I've taken out some of the complexity. This is why I take results from "accurate" models with a grain of salt. Remember the old saying: "Past performance is no guarantee of future results." - akck ------------You got to know how to skin the rabbit which means you have to know how to make do with what you got. Get by in life. It helps to be frugal to start with. If all you got is a bag of rice and a bag of beans you have to know how to cook rice and beans. If the only kind of food you know how to fix is processed food and use the microwave you might be hurting in hard times but if you know how to make do with just a little bit you can get by in some pretty hard times. I made Pho for brunch this morning using all leftovers and it was awesome. I picked a couple of Cholula peppers off a plant in my living room. I brought it in when it started to get cold outside. Art
slver has spent the week firing back, criticizing political pundits for not understanding how odds and probability work..The bolded portion is the key question in my mind. Is Silver's prediction based on probability math or is it an effort keep conservative voters home?I guess we find out Tuesday.How will we find out? We already know what the right wing is going to say. If Obama wins then obviously Silver's voter suppression efforts worked. If Romney wins then it is obvious Silver was sandbagging. They've set up a condition where they can claim victory no matter what. However there is a way to find out, but no right-winger will ever use it. More on that below. However the bold part is correct, if a bit generous. I've made a series of posts related to this topic over the last month or so. I've given many examples of top right-wing pundits who simply don't understand (or pretend not to understand) math at all. Basically, the right wing is in full-on conspiracy/melt down mode whenever someone mentions polling data. Who are you going to believe? Your lying eyes or Rush?Thing is, you don't need Nate Silver to come up with a conclusion that is similar his. As a first approximation, you can simply add up the electoral votes for states where Obama is likely to win, and if you do that it comes up to be more than 270. That has been true all summer and fall. The pros use Monte Carlo simulations and such and calculate all possible outcomes, which is certainly a refinement. I can think of about five or six sites who do the same thing, with slightly different methodologies, and they all come up with close to the same result. The right wing response is as predictable as the sun rising in the east. Back to the question you raised, how do you find out? If this really was a conspiracy as the right-wing pundits are claiming, then it would be dead simple to unmask it. Simply get the topline polling data (easily obtainable) and run the Monte Carlo simulation yourself. You can do this in Excel. If it turns out the probabilities are in favor of Romney that would be pretty damning. Surely there has got to be a right winger somewhere who has actually done the obvious thing and used this method, right? Guess what? Not one of the people making the attacks on the data have tried this, and they don't know anyone who has tried it, and they don't know anybody who knows how to try it. Funny that. Instead conservatives react as they always do when confronted by facts and numbers. Instead of examining the evidence for themselves they: 1. Attack the messenger2. Sow seeds of doubt3. Impune the character of knowledgable people4. Invoke conspiracies5. Advance implausible alternate theoriesThis creates the scenario in my first paragraph. By ignoring the obvious question if the data are true or not, any result that happens can be used as evidence the right wingers were correct.
My only comment is that models continue to work until they don't. You may find this poece interesting.http://www.washingtonpost.com/opinions/robert-j-samuelson-po...Among pollsters, there’s fear that changing technology (mainly cellphones) and growing public unwillingness to do interviews are undermining telephone surveys — and that there’s no accurate replacement in sight. A recent study by the Pew Research Center reported its response rate at 9 percent, down from 36 percent in 1997. Put differently: in 1997, Pew made about three residential calls to get one response; now it makes 10.Cellphones pose problems because people who use them exclusively — people who don’t have landline phones — are younger, poorer and more Democratic than the general population. By late 2011, 32 percent of Americans 18 and over had only a cellphone, up from 16 percent in early 2008. Among those 25 to 29, the share was 60 percent. Under-surveying these people could distort polls. Many pollsters, though not all, now canvass cellphones. But this is increasingly expensive. By present trends, half of Americans could be exclusive cellphone users by the 2016 election.All this threatens the largest upheaval in polling since the 1930s.
Aren't you the one who gave us http://votamatic.org/ ?I am re posting it here because I couldn't exactly remember the name but finally found it.Lindap.s. President Obama is stealing the election. He is the most powerful man ever because he and his henchmen are able to sway millions and millions of people who aren't paying attention. But President Obama is incompetent and not a leader. Except when he is showing the government working more efficiently after Hurricane Sandy. And that traitor Chris Christie. He's to blame, too. The fix is in.
Instead conservatives react as they always do when confronted by facts and numbers. Instead of examining the evidence for themselves they: 1. Attack the messenger2. Sow seeds of doubt3. Impune the character of knowledgable people4. Invoke conspiracies5. Advance implausible alternate theoriesFunny I could probably find this sentiment over @ campfire discussion group with the exception that the word liberals would be substituted for the word conservatives. IMHO they would be both true. Both political parties are leading our country to bankruptcy thru their prolific spending. The vote between the parties [lesser of 2 evils] is merely the speed at which we reach that bankruptcy.I view elections now as entertainment. Similar to the circuses that Roman Emperors put on to distract the populus.Its gonna be quite entertaining once medicare is bankrupt due to the increased # of beneficiaries & rampant health care inflation & doctors refusal to accept any new medicare patients*. There will be much finger pointing from our duly elected officials casting blame. But no acceptance of blame as the issue was ignored by them prior to the bankruptcy of the program. *4 years ago my 95 year old uncle had a 6 month wait before one of best [imho] primary care docs would take him on. I fully expect the trend to worsen as time goes on.
Aren't you the one who gave us http://votamatic.org/ ?Nope. I appreciate the link though. I don't have time to look at it now. Off to yogi class & the treadmill.
Nate Silver's reputation and living is based on the accuracy of his forecasts. Ideologues aren't as accurate as those who stick with the algorithm.In 2008 he got just 2 states wrong by a hair -- he said Indiana would go McCain and Obama won. The reverse happened in Missouri. It's over barring "the live boy or a dead hooker", as former Louisiana Gov. Edwin Edwards used to say.intercst
You may find this poece interesting.http://www.washingtonpost.com/opinions/robert-j-samuelson-po......Count me in as one of the people who refuse to answer polls. That plus cell phones make polling models about useless. For all we know, Romney could win big or Obama could have a blowout. As you said, Tuesday will tell.
Is Silver's prediction based on probability math or is it an effort keep conservative voters home?Do conservatives--the kind that would stay home because they thought their candidate might lose--even read liberal statistical bloggers in the NYTimes?And who stays home coz their candidate might lose? That would motivate me more. Hey, I'm in South Carolina, and I think my candidate is gonna lose here--so I voted this morning.
Best Of |
Favorites & Replies |
Start a New Board |
My Fool |