Kerry got an 11.5% bounce! Win prob (5 polls) went from 85.9% to 97.4%

[TruthIsAll]

No bounce, eh?

After Convention: 
Kerry 50.0%, Bush 44.2% (5 poll average)
Spread:5.8%

Probability of Kerry win:97.4%	

Poll	Kerry	Bush	K/K+B MoE	
CBS	.49	.43	0.533	0.030	
ABC	.52	.45	0.536	0.040	
NWK	.52	.44	0.542	0.040	
ZOG	.48	.43	0.527	0.030	
ARG	.49	.46	0.516	0.030	
						
Mean	0.500	0.442	0.531	0.034	
Std	0.500	0.497	0.499		
StdMean   0.016	

	


.................................................
Before convention:

Kerry 49.0%, Bush 45.8% (5 poll average)
Spread: 3.2% 

Probability of Kerry win:85.9%		

Poll	Kerry	Bush	K/K+B	MoE	
CBS	.49	.44	0.527	0.030	
ABC	.48	.49	0.495	0.040	
NWK	.51	.45	0.531	0.040	
ZOG	.48	.46	0.511	0.030	
ARG	.49	.45	0.521	0.030	
						
Mean	0.490	0.458	0.517	0.034		
Std	0.500	0.498	0.500			
StdM	0.016			
						

[cally]

1. Truth...I appreciate all your posts on polls

I read them all and I learn so much. I rarely respond but I do value your posts. Thanks.

[chiburb]

17. And that's the TRUTH!!! n/t

[zulchzulu]

2. Paging Tom Ridge! Paging Ashcroft! Release Terror Pack immediately!

We need to stomp on this story. Send out the Terror Pack Rumsfeld talked about at this morning's meeting!

Hume is ready for the handoff.

[whistle]

3. I can follow the poll summaries, but not quite sure where...

...you got the win probability from. If Kerry had a win probability of 85.9% before the convention that was actually pretty strong. By bouncing to 97.4%, that's almost a sure thing. How can you make such a claim 85 days before the election based solely on the polls? Just curious, not doubting you.

[TruthIsAll]

5. The probability is a function of the spread in the polls.

Without getting too technical:

1) calculate the average (mean) of the poll numbers.
2) calculate the standard error of the mean (volatility).
3) plug the results into the normal distribution function

The NDF returns the probability that Kerry will exceed 50% of the vote, based on 1) and 2).

I am merely trying to show that the bounce is significant, if one considers the % change in probability that Kerry will win.

[southernleftylady]

4. where do you get all these #'s from? i find it so interesting ! n/t

[TruthIsAll]

6. Pollingreport.com has the latest polling numbers

www.pollingreport.com

[The_Casual_Observer]

7. Best Regards!

[Downtown Hound]

8. If the election is clean

Bush is a goner.

[kath]

9. The problem is, of course, that that's a MIGHTY big "if".

They've got so many ways to steal it, the fuckers.

[Downtown Hound]

11. Yep

I'm not so worried about Ohio because they decided not to use Diebold. Thank God. Florida, however, that's another story. The combination of Jeb and Diebold could make for some pretty nasty voter fraud. Hopefully, Kerry is on this. They say he has an army of lawyers waiting. Also, Michael Moore will be there with his camera. Who knows if any of this is really going to help, but at least people are watching.

[nicktom]

10. I also appreciate your poll analysis,

but it seems to me the biggest thing we need to be concerned about is the state polls. The electorate numbers are far more important. Especially the so called "swing states". Ohio, Florida, Nevada, etc. Do you chart these polls?.

The electoral votes are the one's that count.

[TruthIsAll]

12. This is my National/State EV Simulation Model, updated daily

I use pollingreport.com for national polls and electoral-vote.com for state polls and calculate probabilities for both.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=104x2153657

[newyawker99]

15. Hi nicktom!!

Welcome to DU!! :toast:

[cthrumatrix]

13. how unamerican to point out facts at a time like this

[TruthIsAll]

14. Someone has to do it.

tia

[Stephanie]

16. I don't understand the math

How do you figure 11.5?


before
49.0 Kerry/45.8 Bush

after
Kerry 50.0%/Bush 44.2%

Isn't that more like a 1% change?

I plead ignorance of statistics, percentages, and math. How do you get 11.5?

[BillZBubb]

22. The 11.5% is not a difference in the polls

It represents the difference in the probability that Kerry will win the election based on a statistical model. What it is saying is that after the convention if the election occurred, Kerry was 11.5 percent more likely to win.

To put it into coin flips: For Bush* to win before the Dem convention, he would have to flip heads eight times in ten tries. Since most times it would be five, that's hard to do. After the convention, Bush* would need nine heads in ten tries--that doesn't happen too often.

Does this really mean anything? It depends on whether the polls are reflecting what's really happening. And it is only valid right now for those polls.

[Stephanie]

23. Okay thanks

Still a little confused - that's why I don't go to Vegas. But thanks for the explanation. So Kerry's odds are very, very good. Do they take bets on this in Vegas?

[TexasSissy]

18. I like optimism, but this is overzealousness in an effort to put

a better face on things.

ALL polls should be used, even the CNN one that we don't like. Not just the ones whose results are the best.

Other than that, I don't understand probability analysis. But on the face of things, Kerry did not get an 11% bounce. But he did get a bounce in the range of what was expected by most, given that the country is so divided and heels are dug in.

It DOES look like Kerry is on a winning trajectory.

[Tweed]

19. And in a week the probability could change back again, which means...

All of this 'probability' talk is meaningless. I apperciate the time you put into the posts, but the numbers mean nothing in the end. Truth is all.

[PATRICK]

20. So what bounces ever stayed solid anyway?

Some real losers got real big bounces, probably out of pure hope and getting together at the convention, only to let the hangover of candidate reality slide them back down. No, the real relentless progress and reliability of the Kerry candidacy and the hapless intransigence of the Bush fiasco is the story. Bush's numbers are under siege.

This is the time of slow, determined buildup to November. No one ever freeze framed their poll numbers, even weeks before. the press tried to do that to Gore and he won the popular vote anyway. Our landslide will be just as irrepressible. The GOP will have to gamble on surprises. That gamble could cripple their power base even more.

Reality TV will be on election day, the prelude is unintentional comedy sitcom fare.

[CaTeacher]

21. I think this will be the

biggest Democratic landslide in history!!!

[cthrumatrix]

24. if you get people out to VOTE