2016 Postmortem
Related: About this forumI Have Been Out. What New Polls Pushed Professor Wang's Meta Margin To 2.30
Probability of Obama re-election: Random Drift 93%, Bayesian Prediction 98%
http://election.princeton.edu/
aaaaaa5a
(4,667 posts)But it may not be any one poll. As the election nears, Obama's lead (even if small) looks more insurmountable, thus his odds increase.
For instance,
Say you are playing a football game. If you lead by 3 points in the first quarter, your odds of winning are still small. However if you lead by 3 points with only 2 seconds left, your odds of victory are great. Even if your actually margin is very small.
I hope that makes sense.
DemocratSinceBirth
(99,710 posts)TroyD
(4,551 posts)I think that's what Nate Silver said too.
Even a lead which is "only 2 points" in OHIO can be very relevant if it is the week before an election as opposed to 3 months away.
DemocratSinceBirth
(99,710 posts)Silver does essentially the same thing but his model has a lot more variables.
aaaaaa5a
(4,667 posts)Wang was more accurate in 2008. He beat Silver "head to head."
And he was also dead on accurate in 2004.
Phx_Dem
(11,198 posts)which Obama won by 1 percent. He correctly predicted all 35 senate races. He's got a stellar record.
regnaD kciN
(26,044 posts)...plus, he called the much-closer 2004 race perfectly. (That's one caution I have about Nate -- he's never dealt with a too-close-to-call nationwide contest before.)
At any rate, isn't it ironic that all of the conservatives are attacking Nate Silver, when every one of the other predictors (except unskewedpolls, of course) has their boy as an even bigger longshot?
DemocratSinceBirth
(99,710 posts)One of Wang's criticisms is by weighting polls Silver opens himself up to scrutiny and introduces noise into his model.
What's ironic is the right hate Silver when there are much more bullish Obama models on the net. But those models aren't featured in the New York Times.
aaaaaa5a
(4,667 posts)The GOP rips Silver. But of all of the scientific models out there, its actually Silver's model that is most favorable to Romney. In many of these other models, this election (with regard to who is going to hit 270) isn't even close.
aaaaaa5a
(4,667 posts)That no President has ever lost a state with this much data, that was in a similar position to what Obama is in, with regard to the state of OH.
aaaaaa5a
(4,667 posts)If Obama has a 2-3 point lead in Ohio on Monday, the election is over.
Contrary to popular opinion, polls are not usually wrong. Especially aggregate polls and poll forecasting. This is especially true in a place like Ohio where as Nate Silver has mentioned, there is a ton of data, making predictions and forecasts far more easy to make.
Go to Sam Wang's site. Look at his Presidential track record in 2008 and 2004. He was dead on accurate in both elections. What his site says on Monday will tell us a lot about what will happen Tuesday evening.
The GOP hates math and science. We saw that front and center with Morning Joe this morning.
gcomeau
(5,764 posts)...but not meta margin changes. Meta margin changes are poll driven.
DemocratSinceBirth
(99,710 posts)And isn't the meta margin more sensitive to polls from tipping point or swing states.
jcgoldie
(11,631 posts)...but I'm ashamed to say I'm clueless, WTF is a meta-margin?
DemocratSinceBirth
(99,710 posts)You should go to his site. The man is brilliant and a Democrat.
MSMITH33156
(879 posts)What is the Popular Meta-Margin?
The Popular Meta-Margin is the amount of opinion swing that is needed to bring the Median Electoral Vote Estimator to a tie. It helps you think about how far ahead one candidate really is. For example, if you think support for your candidate is understated by 1%, this can overcome an unfavorable Meta-Margin of less than 1%. If you think that between now and Election Day, 1% of voters will switch from the other candidate to your dude, this is a swing of 2% and can compensate for a Meta-Margin of 2%.
Cali_Democrat
(30,439 posts)Obama's odds will actually increase as the election approaches because he's already ahead.