Skip to main content

As has been the case for three weeks now, once again there is no change in the Senate Snapshot. The most likely outcome remains 52 Democratic seats (that counts Senators Lieberman and Sanders). The chance of a Republican takeover, if the election were held today, is 1.5%. That is up slightly over the past three weeks, but not by much.

As I discussed on Monday, stability is truly the hallmark of the Senate Snapshot this year.

Senate competitive campaigns chart

Senate seat outcome odds chart

I’ve put a series of question marks next to the Alaska campaign, because I simply have no idea how to accurately poll a write-in campaign. It is an unknown.

Nate Silver published an article about election forecasting yesterday that really caught my eye. The following paragraphs suggested the possibility of an Ultimate Election Forecast for statewide campaigns that would eliminate the need for any other election forecasts (at least forecasts Presidential, Senatorial, and Gubenatorial campaigns). Here are the three relevant paragraphs from the article:

Senate candidates who have a lead of 6 to 9 points in the simple polling average, with 30 days to go until the election — about where Mr. Toomey’s lead stands now — are undefeated since 1998. This isn’t quite as impressive as it sounds, since there are only seven such candidates in the database. But if we expand the scope of our study just a bit, it proves to be the norm rather than the exception. Senate candidates with a slightly larger lead in the polling average — 9 to 12 points — are also undefeated. Candidates with a slightly smaller lead in the polling average — 3 to 6 points — have a pretty good track record, with nine wins against three defeats.

Indeed, no Senate candidate with a lead of more than 5.5 points in the polling average, with 30 days to go in the race, has lost since 1998: these candidates are 68-0. (Martha Coakley in Massachusetts would have been an exception, but special elections, where the polling can be much more erratic because of lower turnout, are outside the scope of our study.)

Candidates for governor with a lead of 6 to 9 points in the polling average, meanwhile, have a 9-to-2 record. If we combine their numbers with those of the Senate candidates, we find that candidates with a lead comparable to Mr. Toomey’s (6 to 9 points, with 30 days to go) have 16 wins against two defeats, which corresponds to an 89 percent winning percentage.

This caught my eye because it is very similar to the methodology I use for the Senate Snapshot. In the Snapshot, I take the simple 25-day polling average (leaving out Zogby Interactive and Columbus Dispatch polls), because I found that number of days to have the least error of any simple average in projecting the final vote margin from the final polling average. Then, I calculate the “current win %” based upon the percentage of instances the 25-day simple polling average has been incorrect by more than the current margin. (Actually, I take half of the percentage of times the 25-day average has been incorrect by more than the current projected margin, since I assume there is an equal chance for error to break in favor of the Democratic or Republican candidate.)

The Ultimate Election Forecast becomes possible by combining what Nate is discussing with what I am doing.  If you had an enormous amount of time on your hands, you could calculate the error between the simple polling average and the final vote margin at any point in the campaign for all Senatorial and Gubernatorial campaigns since 1998 (or earlier, if you have the data). For example, you could determine that simple polling averages taken 100 days before an election erred by 5.0% or more on the final margin in 30% of Senate and Governor elections. You could do this for any number of days before an election--50, 150, 7, 213, however many you wanted. All of them, really. If you did do that, then with you would be able to project the odds of victory or defeat in any statewide campaign at any point in the campaign. And you could do so with real certainty.

Now, I don’t have that much time on my hands. It took me a few weeks just to determine the odds of victory the day before an election. Determining the odds of victory for each of the 500 days before an election would be an utterly galactic undertaking. But, if anyone were to do it, then s/he would actually have ”The World’s Greatest Election Forecast, In the World,” and there wouldn’t be much need for any others.

If you want to be the ultimate election guru for 2012, and beyond, that is how you would do it. Any takers?

Notes
--Polling is not a conspiracy designed to manipulate your emotions.

--Polling averages don’t suck. They are quite accurate.

Actual Notes
--This is a snapshot, not a forecast. All of the odds presented here are based on if the election were held today. It is not a prediction of future trends.

--Only campaigns within 14% are listed. If a campaign isn't listed here, then it is not currently as close as any of the campaigns listed here.

--Lisa Murkowski’s odds of victory are included in the overall GOP total. Charlie Crist’s odds are split evenly between Democrats and Republicans.

--With few exceptions, all polls used in the averages are taken from Pollster.com.

--A complete description of the methodology behind this snapshot, along with all the research and a FAQ, can be found here.

Originally posted to Daily Kos on Fri Oct 01, 2010 at 04:46 PM PDT.

EMAIL TO A FRIEND X
Your Email has been sent.
You must add at least one tag to this diary before publishing it.

Add keywords that describe this diary. Separate multiple keywords with commas.
Tagging tips - Search For Tags - Browse For Tags

?

More Tagging tips:

A tag is a way to search for this diary. If someone is searching for "Barack Obama," is this a diary they'd be trying to find?

Use a person's full name, without any title. Senator Obama may become President Obama, and Michelle Obama might run for office.

If your diary covers an election or elected official, use election tags, which are generally the state abbreviation followed by the office. CA-01 is the first district House seat. CA-Sen covers both senate races. NY-GOV covers the New York governor's race.

Tags do not compound: that is, "education reform" is a completely different tag from "education". A tag like "reform" alone is probably not meaningful.

Consider if one or more of these tags fits your diary: Civil Rights, Community, Congress, Culture, Economy, Education, Elections, Energy, Environment, Health Care, International, Labor, Law, Media, Meta, National Security, Science, Transportation, or White House. If your diary is specific to a state, consider adding the state (California, Texas, etc). Keep in mind, though, that there are many wonderful and important diaries that don't fit in any of these tags. Don't worry if yours doesn't.

You can add a private note to this diary when hotlisting it:
Are you sure you want to remove this diary from your hotlist?
Are you sure you want to remove your recommendation? You can only recommend a diary once, so you will not be able to re-recommend it afterwards.
Rescue this diary, and add a note:
Are you sure you want to remove this diary from Rescue?
Choose where to republish this diary. The diary will be added to the queue for that group. Publish it from the queue to make it appear.

You must be a member of a group to use this feature.

Add a quick update to your diary without changing the diary itself:
Are you sure you want to remove this diary?
(The diary will be removed from the site and returned to your drafts for further editing.)
(The diary will be removed.)
Are you sure you want to save these changes to the published diary?

Comment Preferences

  •  If... (5+ / 0-)
    "this is a snapshot, not a forecast", why does the headline call this the "ultimate election forecast"?

    Does Christine O'Donnell weigh as much as a duck?

    by osterizer on Fri Oct 01, 2010 at 04:49:50 PM PDT

  •  Doesn't Liebermann Caucus with the GOP (1+ / 0-)
    Recommended by:
    divineorder
    this next Congress?

    What "puts him over the edge"?

    Notice: This Comment © 2010 ROGNM

    by ROGNM on Fri Oct 01, 2010 at 04:49:52 PM PDT

  •  Woah (5+ / 0-)
    Recommended by:
    askew, elwior, divineorder, Losty, scamperdo

    The chance of a Republican takeover, if the election were held today, is 1.5%

    At intrade the wingnuts have pushed the chances of Democratic control down to just over 50%.  http://www.intrade.com/...

    I'm pretty sure that's not so much a good predictor as the possibility that a lot of the people betting there are FOX drones.  

    An 'attack on the left' is the GOP in office destroying lives.

    by Sun dog on Fri Oct 01, 2010 at 04:53:16 PM PDT

  •  Is there a reason you keep leaving Leahy VT out? (1+ / 0-)
    Recommended by:
    fladem

    he's never in your senate projections and he's up for re=election this election.

  •  Don't you think the PA polls will change (5+ / 0-)
    Recommended by:
    Odysseus, joanneleon, alba, elwior, divineorder

    in Sestak's favor once Toomey's opinion of Social Security gets widely circulated?

    Doesn't PA's population skew a little older and retired? Seems those older Americans would be a bit afraid to gamble on losing what might be their only means of support by electing a guy like Toomey who wants to kill Social Security.

    "The difference between the right word and the almost-right word is like the difference between lightning and the lightning bug." -- Mark Twain

    by Brooke In Seattle on Fri Oct 01, 2010 at 05:01:27 PM PDT

  •  Doesn't account for malfeasance (1+ / 0-)
    Recommended by:
    elwior

    I think you migh be better looking at a group of five or so pollsters with a demonstrable history of accuracy.  There is a lot of money floating around this election and a lot of unknown or suspevt polling companies throwing numbers into the pot.  At this point we are seeing same state, same time period, polls differing by several times their stated margin of error.  Keeping to "peer reviewed" polls may give a more accurate result.  Then again it could just be I'm bummed at seeing so many deems getting crushed.

    •  The polling used in these averages (0+ / 0-)

      has been around for a couple of cycles.  Most of the newer pollsters I have seen are polling in the House races.

      It is clear that taking the average (I actually prefer the median) is empirically the best model.

      The bitter truth of deep inequality has been disguised by an era of cheap imported goods and the anyone-can-make-it celebrity myth - Polly Toynbee

      by fladem on Fri Oct 01, 2010 at 05:16:42 PM PDT

      [ Parent ]

  •  its all about Sestak and Feingold (7+ / 0-)

    how two Blue States can go wingnut is beyond reason.

    "The way to see by faith is to shut the eye of reason." - Thomas Paine

    by shrike on Fri Oct 01, 2010 at 05:02:14 PM PDT

  •  Anything for Congress? (1+ / 0-)
    Recommended by:
    We Won
  •  why so positive on WV (1+ / 0-)
    Recommended by:
    CA Berkeley WV

    The polling seems to be a dead heat at best.

    •  'Cuz Raese bought flowers for his young daughter (1+ / 0-)
      Recommended by:
      elwior

      in Sisterville, and had them sent to Palm Beach, Florida. It may only be a town of 1500, but that gets around. He tries to claim he still lives in Morgantown, but the mansion in Trump's neighborhood means he is claiming residence in a state with no income tax, and no personal property tax on those limos he bragged about have two of that day.

      Jay is stinking rich, inherited like Raese, but he came to W. Va. as a VISTA volunteer and stayed. And if he had not lost that gov race in 1972 to felon Arch Moore, father of WV-02 Shelly Cap-pee-toe, I bet he would have stayed a little more skeptical of MTR.

      One-third of the population is on Social Security and he is scaring them. Chamber and NRA endorsed Joe. The polls have been erratic. Scratch the rooster on the one ballot. What more do you need to know?

  •  I have been trying to figure out (7+ / 0-)

    why the odds of a takeover in this analysis and in similar analysis by others seems so off.  The odds of a GOP takeover I think are a hell of a lot higher than 1.5%.

    The answer is that it treats these elections as independent events.  BUT THEY AREN'T. The Senate race in Illinois is connected to the Senate race in California.  Both are dominated by the same issue (unemployment), and both are liable to move in the same direction.  So the cumulative odds are actually incorrectly computed.  

    Let's take a step back.  Let's assume there are two factors in a given Senate race:
    *National factors (Beta)
    *Local factors (Alpha)

    This would mean that the probability of an outcome in a Senate race would equal the following:
    Pr = Beta + Alpha

    So Alpha would equal the local polling, and Beta would equal national polling (say generic ballot in off year elections).  To get the data for this model, you would have to explore how signficant changes in national factors are on local polling.  You might find that national factors account for 30% of a Senate Race, and local factors 70%, for example.

    Those familiar with stock market caluations will recognize alpha and beta.

    In grad school I actually did research on the relationship of stock prices to external events (pollution, in particular).  It strikes me that quantitative analysis is far more nuanced in finance than it is in politics.  

    But is only a matter of time before our models get significantly better.

    Make no mistake, Chris's model is the best I have ever seen.  But I can see a day when you get some calculation like this:

    Outcome = national factors + local candidate + local economy + racial makeup + political consultant.

    What would be interesting is you could start to ask whether different political consultants or strategies (what is the effect of a particular advertizing strategy) had on the outcome of an election.

    In 10 years we will have all of that data - and it will be as interesting as hell.

    Sorry for the long comment.

    The bitter truth of deep inequality has been disguised by an era of cheap imported goods and the anyone-can-make-it celebrity myth - Polly Toynbee

    by fladem on Fri Oct 01, 2010 at 05:07:03 PM PDT

  •  Incumbents (3+ / 0-)

    Here is why I think these statistics might not hold up so well this year.  Up until now, it was conventional wisdom (with strong evidence) that incumbent Senators were extremely difficult to unseat.  

    Nate's data for the past 12 years IIRC would include a lot of years in which the incumbent had a great advantage.  This year, the incumbent advantage is no longer a fact.  I think this may affect the predictive ability of the model.

    Disclosure: I'm working as an unpaid citizen journalist covering the Sestak campaign/ PA Sen. race for Huffington Post's "Eyes and Ears 2010" project

    by joanneleon on Fri Oct 01, 2010 at 05:09:01 PM PDT

    •  You have just identified a real weakness (4+ / 0-)
      Recommended by:
      Odysseus, joanneleon, Scientician, elwior
      in these models: it only goes back 12 years.

      There is no detailed data on an election taking place during a significant recession.

      These probabilities all have enourmous selection bias.

      But as Chris said, getting the data is hell.  It took me months to put together the data on the relationship between Iowa and New Hampshire.

      The bitter truth of deep inequality has been disguised by an era of cheap imported goods and the anyone-can-make-it celebrity myth - Polly Toynbee

      by fladem on Fri Oct 01, 2010 at 05:15:17 PM PDT

      [ Parent ]

      •  Sample size (2+ / 0-)
        Recommended by:
        Odysseus, elwior

        Yes, I read Nate's article when it came out and the sample size for the 6-9 pt range (a sample size of 7) just seems absurdly low.  The fact that we're looking at a 12-yr period, which doesn't allow for a lot of variation in business cycles, and other things, adds to the absurdity.  The fact that Nate now has a very large audience of likely voters (NY Times readers) makes what he is doing rather irresponsible too, IMHO, especially when he keeps using one race, PA Senate, as his example.  It seems like he desperately wants to be right on this one, and has not a care in the world about what effect he might be having on the race at a time when voter turnout is going to be especially important.

        I'm not a statistics expert, but I have a little background, having taken two courses in it, and having had a lot of math courses in college (engineering degree.)

        Also, the margins of error and the lack of consistency among the polls are a problem.

        Now, I'm very biased on the Sestak race, as I've been a supporter of his for years.  So I'm also thoroughly pissed off that Nate keeps using the Sestak/Toomey race as an example, which in itself, since he is now published in the NY Times, could skew the results of the race or at least depress voters who favor Sestak, IMHO.  I'm pretty sure the NY Times is widely read in Pennsylvania, particularly in the region where Sestak is most likely to draw votes.  

        From my perspective, Nate is like a short-seller right now.  And I'm beginning to question the value of these predictive models.  It's not like the weather, which isn't going to change its behavior after it reads Nate's (and/or Chris') predictions.  It's more like the stock market.  I'm inclined to tell both Nate and Chris to knock it off, and let the voters decide based on the merits of the candidates, and knock it off with the largely ego-based modeling on this election, and quit talking about their "Ultimate Election Forecasts" and "Uncanny Accuracy of Polling Averages" and "The World's Greatest Forecast Election Forecast, In the World."  (For anyone who might be wondering, no, I didn't make those up, those are actually phrases used by Nate and Chris.)

        Uncanny, and World's Greatest, et al, are not claims that should be made by guys with 12 years worth of data, and Nate, in particular, should not be predicting with such Grand Poobahness with a sample size of freaking 7.

        Disclosure: I'm working as an unpaid citizen journalist covering the Sestak campaign/ PA Sen. race for Huffington Post's "Eyes and Ears 2010" project

        by joanneleon on Fri Oct 01, 2010 at 07:26:45 PM PDT

        [ Parent ]

        •  I agree with you joanne, and I hope you do (0+ / 0-)

          tell these guys to knock it off, particularly Nate's use of the Pa. race.

          "We the People of the United States...." -U.S. Constitution

          by elwior on Fri Oct 01, 2010 at 08:58:15 PM PDT

          [ Parent ]

        •  In finance (1+ / 0-)
          Recommended by:
          joanneleon

          the quants (I am one) built models of stock price movements.  Nassim Taleb criticised them for the same reason you and I have concerns about the numbers Nate generates.  Taleb's nickname is the "Black Swan".  The name comes from his observation that quantitative models systematically underestimate the chance of a "black swan event".

          The theory is based on the knowledge that there are more black swans than genetic theory would suggest should exist.  Taleb's theory was designed to warn hedge funds that their models would fail because of:

          1. the disproportionate role of high-impact, hard to predict, and rare events that are beyond the realm of normal expectations in history, science, finance and technology, 2) the non-computability of the probability of the consequential rare events using scientific methods (owing to their very nature of small probabilities) and 3) the psychological biases that make people individually and collectively blind to uncertainty and unaware of the massive role of the rare event in historical affairs

          Nate's model is far less built on solid statistical theory than the quantitative models that came unraveled during the financial crisis were.

          In particular, if you look at national polling, there is about a 33% of a 5 point national swing in the last 5 days of a national election since 1968.  But Nate's model ignores this possibility.

          In any given election Nate's model will probably work.

          Until, like the proverbial black swans, it doesn't.

          The bitter truth of deep inequality has been disguised by an era of cheap imported goods and the anyone-can-make-it celebrity myth - Polly Toynbee

          by fladem on Sat Oct 02, 2010 at 03:00:52 AM PDT

          [ Parent ]

      •  Uh, 2008? (0+ / 0-)

        That was right in the middle of a significant recession.

  •  Kirk has $5.2M yet to spend... (2+ / 0-)
    Recommended by:
    LeftOfYou, Losty

    Republican Mark Kirk has reserved almost $5.2 million worth of advertisements for the final weeks of the competitive Illinois Senate race, POLITICO has learned.

    The ad blitz comes as most public polls show the race between Kirk and Democratic nominee Alexi Giannoulias in a dead heat. But Kirk has held a fundraising edge over Giannoulias throughout what has become one of the tightest – and most negative – races of the election cycle.

    Read more: http://www.politico.com/...

    Grassroot passion comes and goes - Corporate interest in consistent. Message being that if you want to stay elected then sell-out, it's safer.

    by Jonze on Fri Oct 01, 2010 at 05:19:35 PM PDT

    •  He'll need every nickel to win. (0+ / 0-)

      Kirk the Jerk, a self-promoting flack of no genuine accomplishment except in the department of resume embellishment, is pretty well known to a heck of a lot of people in the Prairie State. He is a formidable nominee for the R's, all things considered. This final push by Kirk and his myriad corporate sock puppets will be relentlessly negative against Alexi.

      Alexi's support here is pretty strong. I dropped a C note on him this week and will do more. It will surprise me if Kirk takes Blue Illinois.  Kirk's widely known reputation for dishonesty, a common trait in Illinois politicians, will sap much of the effect of the negative onslaught.  

      "If you are going to tell people the truth, be funny or they will kill you." Billy Wilder 1906 - 2002

      by LeftOfYou on Fri Oct 01, 2010 at 05:33:30 PM PDT

      [ Parent ]

  •  I think it sucks that the DSCC has shown (2+ / 0-)
    Recommended by:
    Scientician, Losty

    little interest in NH. We have a chance to actually pick up a seat: Ayotte will make a HORRIBLE Senator.

    She's incompetent, she's got extremist religious nutball views: there's a whole bunch of back stories that eat away at potential support for her.

    Yet the DSCC refuses to play ball.

    They do love that Blanche Lincoln, though.

    Encourage others to register! Deadlines coming up fast! http://tinyurl.com/2ck77p http://tinyurl.com/28sykel

    by shpilk on Fri Oct 01, 2010 at 05:21:15 PM PDT

  •  OK, Lieberman and Nelson go to their party.. Rep. (0+ / 0-)

    Who would be the other one to go over??

  •  Nevada and Illinois should be in the light Blue (0+ / 0-)

    We are up 1.5% in both. There is also a lot of movement likely to come out of October surprises, 10.2.10, 10.10.10, 10.30.10.

    Judt possibly one Octo9ber surprise may include some mention of the fact that Karzai has been negotiating a new government brokered by Petraeus with 70 some odd moderate Taliban in it and with new elections bringing us a good chance of just walking away in 2011 with the Taliban in charge.

    Live Free or Die --- Investigate, Incarcerate

    by rktect on Fri Oct 01, 2010 at 05:34:49 PM PDT

  •  Partisan Polling - Treated as Neutral & Isn't (4+ / 0-)

    I doubt this relevant because I don't remember an election with so much partisan polling. Nate hasn't factored bad data into his model.

    You've had a few charts with and without Rasmussen. These have been very telliing. Rasmussen is not the only one skewing the samples. I'd like to see a model showing the full effect of all the bad polling.

    Kos had an piece not long ago noting the tell tale signs of Rasumssen being used for "narrative setting" and not prediction. I think these are now being used for "narrative re-enforcement".

    Just as Conway is pulling even, Rasmussen shows an 11% favor for Paul. Just as Sestak is pulling even, Ras comes in with a 9% for Toomey.  This adds to the piles of other Rassmussen polls that weight the prediction model.

    Now that CA is being heavily polled, Ras. has eased a bit, and Boxer's numbers are, probably, what they always were... ahead of Fiorina.

    There's a term for using bad data that you don't hear much anymore "garbage in - garbage out". I'm interested in the models and follow them. My brain adds 3% for the Dems, and then the margin or error.

     

    •  A Very Real Concern. (1+ / 0-)
      Recommended by:
      inHI

      I've commented separately on this thread that models like Chris and Silver are using rely too much on past performance by pollsters to predict the accuracy of the pollsters current projections against actual outcomes.  

      You raise a point that is different and, possibly a more important reason to distrust these analytical models: their over-inclusiveness. And they have to be over-inclusive because their models are statistical constructs that require it. Nevertheless, it stains and weakens my faith in the results.

      "If you are going to tell people the truth, be funny or they will kill you." Billy Wilder 1906 - 2002

      by LeftOfYou on Fri Oct 01, 2010 at 05:47:41 PM PDT

      [ Parent ]

      •  I have just researched .. (2+ / 0-)
        Recommended by:
        Odysseus, LeftOfYou

        A couple of polls that their numbers were being manipulated to give the proper narration mainly by how they interpreted what is or is not a likely voter. I think it was an AP poll that showed in headlines Republicans lead Democrats(in the generic congessional poll) 47-37% amongst likely voters but if you read that poll you find out that amongst RV's thew DEMS lead 49-37%, all in all way too big of a swing from the RV to LV models. That poll also showed 36% of the respondents were from the South which was 13% more then any other region.
         In another poll CNN I think it showed over all the GOP leading the DEMS by 9 points again amongst LV's and trumpeted the enthusiasm being lead by the GOP by a 46-34 margin in the very enthused group but it convieniently left out the fact that in the somewhat enthused group the DEMS lead by a 49-34 margin. I really fail to much of a deficits in enthusiasm except that it's the INDYs that are most affected by un-enthusiasm.

        •  AP Reporting of thier Polls has been awful. (2+ / 0-)
          Recommended by:
          Odysseus, inHI

          I've commented on this before.  Actually, I diaried on your very point, last month. Hence, I agree that your are 100% correct. Polls have simply become props in a media narrative that pays little heed to the actual significance of the numbers, even the numbers of the less well constructed polls.  

          "If you are going to tell people the truth, be funny or they will kill you." Billy Wilder 1906 - 2002

          by LeftOfYou on Fri Oct 01, 2010 at 07:32:00 PM PDT

          [ Parent ]

          •  rasmussen.. (1+ / 0-)
            Recommended by:
            zapus

            I really like how he will post a poll showing like Reid loosing,Paul pulling ahead by 11-12 points and on and on after another poll shows say Conway or Sestak getting within a couple of points or Boxer/Murray pulling ahead. Pure narrative if you ask me.

      •  Not certain they "have" to be over - inclusive. (0+ / 0-)

        Nate notes the "house factor".
        Chris has made some Rasmussen-free charts.

        Nate could take averages of those polling companies with house factors and re-balance his model.

        I can't name pollster that consistently produces outliers slanted to D's.  In fact, I can't even remember a single outlier poll with a D in the lead.

  •  OH-Sen, Note that number. Zero.. (0+ / 0-)

    Zip, Nada, No Chance..

    From 3 to Zero.. How was it ever three..

    •  From tied to zero (1+ / 0-)
      Recommended by:
      Losty

      Absolutely appalling. I should have given more credence to what I was hearing form long-timers about Lee Fisher. I did not honestly believe he could personally blow a campaign like this — and then in the homestretch, blame it all on his primary opponent.

      De-orangify Congress: Justin Coussoule for Oh-08 http://www.coussouleforcongress.com/

      by anastasia p on Fri Oct 01, 2010 at 08:35:47 PM PDT

      [ Parent ]

      •  And you posting this allows me to vote.. (0+ / 0-)

        I knew the first time I saw him speak he was "dead man walking"..

        Note to some people on here... FIGURATIVELY.. Not actually, just politically..
        (Not you Anastasia)

        This may be a good message race, though I would like to see non-Ras polls on the races all the way down ticket.. and you have to vote for Justin, even if that race may be done too..

        I'd love to see how Owens getting the Tea nod is affecting the others, and if more endorsements from them lead to more close races..

        When I walked by a Union hall and saw Strickland/Cordray/O'Shaunnasy(sp) signs, and not one Lee, he's done..

  •  My Problem With Nate Silver (3+ / 0-)
    Recommended by:
    Odysseus, Houses in Motion, zapus

    And with you too, Chris, is the importance in your analytical predictions of the idea that past pollster performance in prediction of outcomes is, itself, a predictor of future pollster performance in predicting outcomes.

    It's a tempting idea. It seems intuitive. That's why Federal law requires that investors be warned that they should not make such assumptions with regard to the performance of investments. The past is no doubt instructive, but not necessarily predictive.  

    At no time is this more true than when transitional forces are in play that have not occurred widely before.  This may be one of those times.

    I am buoyed by the optimism of these forecasts in predicting continued Democratic power to block the worst excesses that might take place if the Republicans make significant gains during the mid-terms. Holding the Senate means everything in that sense. But I actually think there are a large number of reasons to believe, based upon evidence of boots on the ground, that the results will be much more favorable to our side than these projections suggest.  

    "If you are going to tell people the truth, be funny or they will kill you." Billy Wilder 1906 - 2002

    by LeftOfYou on Fri Oct 01, 2010 at 05:41:26 PM PDT

    •  For better or worse.... (0+ / 0-)

      ...this is the very essence of Statistics.  Statistics is an interesting subject to me, but "Lies, Damned Lies and Statistics" rings very true.  I like seeing what they do, because I would be very interested to see how this matches up with reality in the end.  Also, Nate and Chris have to change their projections every so often as new variables occur in certain races, or across the board.  But, it will be very interesting how it comes out in the end.  

      If you listen to fools, the Mob Rules

      by CO Democrat on Fri Oct 01, 2010 at 08:39:23 PM PDT

      [ Parent ]

  •  Neek? (0+ / 0-)

    It's about time I changed my signature.

    by Khun David on Fri Oct 01, 2010 at 05:45:38 PM PDT

  •  Seeing that two-term incumbent Russ Feingold (1+ / 0-)
    Recommended by:
    Odysseus

    has only a 2% chance of being re-elected, I have to doubt the underlying methodology of this snapshot. I'll give you that Feingold is in a jam but his chances of being re-elected have gotta be better than 2%.  

    •  This "snapshot" is based on if the vote was today (1+ / 0-)
      Recommended by:
      Garrett

      It doesn't predict where anybody believes the races will be Nov 2nd.  

      However looking at the snapshop, I think we could actually be worse in Nov, with Manchin potentially losing.

      I was feeling bullish on Alexi, but not if Kirk has $5M to spend over the next month.  

      Grassroot passion comes and goes - Corporate interest in consistent. Message being that if you want to stay elected then sell-out, it's safer.

      by Jonze on Fri Oct 01, 2010 at 06:00:51 PM PDT

      [ Parent ]

      •  The "if the election were held today" stuff (0+ / 0-)

        doesn't seem right to me.

        Probabilistic methods that account for elections being held on election day certainly exist.

        Farther off from the election, the forecasts should be wider and less precise. As the election approaches, the forecasts would become tighter and more precise and more confident.

  •  I struggled with Calculus. This is harder. (0+ / 0-)
  •  Any Mountain Party people from W. Va. read this?? (0+ / 0-)

    Notes

    --Polling is not a conspiracy designed to manipulate your emotions.

    polling without minor party (4% last gov race) candidates in a federal election is NOT propaganda. (as diary at wvablue claimed)

    Sestak's numbers inPenn. surprise me. Bummer.

  •  huh? Manchin is 3:1 over Raese? (0+ / 0-)

    Is that what that 75% thing means?

    PPP just put Raese ahead.

    "Those are my principles, and if you don't like them... well, I have others." - G. Marx

    by Skeptical Bastard on Fri Oct 01, 2010 at 06:20:18 PM PDT

  •  Hey Wisconsin - Feingold deserves better !!! (3+ / 0-)
    Recommended by:
    Odysseus, Jonze, zapus
  •  OK, so you and Nate are telling us these races (1+ / 0-)
    Recommended by:
    Yosef 52

    are lost causes. I won't bother sending any money then...

    or is that NOT what you're saying.  So why even write these articles?

    I guess I just don't get it -- it's not a prediction, but it is?

  •  Somebody tell me... SOMEBODY tell me (0+ / 0-)

    how the F*CK Pennsylvania, our most heavily unionized state, could elect an utterly fanatical Club for Growth asshole like Toomey?

    Somebody tell me how the F*CK Wisconsin might elect a lying, pedophile defending asshole like Johnson over a man like Russ?

    WTF is going on??

    Dammit, it's time for some poetry! And some news!

    by Yosef 52 on Fri Oct 01, 2010 at 09:24:26 PM PDT

  •  So, in the ones where we're down... (0+ / 0-)

    we should just quit, right? If not, then what exactly is the purpose of these posts? To discourage us?

    Dammit, it's time for some poetry! And some news!

    by Yosef 52 on Fri Oct 01, 2010 at 09:29:19 PM PDT

  •  Sen Feingold (0+ / 0-)

    Does anyone on here really believe that Sen. Feingold has only a 2% chance of winning his race?

    •  Feingold is done. (0+ / 0-)

      My Wisconsin friends have told me he is perceived as a national candidate that neglected the folks at home for far too long.

      He hasn't really done much for the state and forgot where he his bread was buttered.

      "I'm not bad. I'm just drawn that way."-Jessica Rabbbit

      by Common Cents on Sat Oct 02, 2010 at 08:53:29 AM PDT

      [ Parent ]

  •  I am really excited you are still doing this! (0+ / 0-)

    It's such a common sense method, I hope it's more accurate than what Silver is doing.

    The Raptor of Spain: An Alternate Histoyr Blog
    From Muslim Prince to Christian King (Updated Sept. 8!)

    by MNPundit on Fri Oct 01, 2010 at 10:23:44 PM PDT

  •  I wish we could have taken Obama's victory margin (0+ / 0-)

    and saved some of it for this year.  That would have been quite helpful.

  •  Estimating the standard error of a proportion (0+ / 0-)

    Warning: This is geeky.

    If you are using past performance to estimate future performance, then, as you note, you need to take account of the sample size of the past performances in your database.  You can't rely too heavily on 7-0.  But how much can you rely on it?

    The standard formula for estimating the standard error of a proportion is

    (pq/n)^.5

    where p is the proportion saying one thing (e.g. "the leader wins") q is 1-p (e.g. "the trailer makes a comeback") and n is sample size.  But when one side is 0, this breaks down.  (1*0/n)^.5 = 0, and the standard error is estimated at 0, which is silly.

    Oddly enough, it doesn't work very well when p or q is close to 0.

    Agresti and Coull wrote a seminal article in American Statistician.

    A simple adjustment that they recommend is to add two to the numerator and denominator, so, instead of 7-0 we would calculate the standard error on 9-2.  Then p = 9/11 = .82, q = 2/11 = .18, n = 11 and the estimate of the standard error is 0.11

    So, the 95% confidence interval for a 7-0 is

    .82 +- 1.96*/.11 = .60 to 1.00

    We all differ in ways that matter. But we're all the same in the ways that matter most.

    by plf515 on Sat Oct 02, 2010 at 03:31:54 AM PDT

  •  I thought Sestak was supposed to do better? (0+ / 0-)

    What a dud he turned out to be.

    "I'm not bad. I'm just drawn that way."-Jessica Rabbbit

    by Common Cents on Sat Oct 02, 2010 at 08:52:10 AM PDT

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site