8/4 KERRY WIN PROBABILITY: 96.65% ; VOTE: 52.78% ; EV:322

[TruthIsAll]

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

[DemocratSinceBirth]

1. This Race Is Going To Be Close

I want Kerry to win with every fiber of my being but to sit here and say there is a 96.65% of a Kerry win ninety days before the election with American troops all over the world and terror alerts springing up like weeds is just plain silly...


There are too many opportunities for events to change the dynamic to use a static model....


Peace 04


Brian

[TruthIsAll]

5. You are incorrect. The model is NOT static. I update it almost daily

with new data. The purpose is not to predict, but to show the changing probabilities of a win. This is the best way to eliminate the volatility of small daily changes in the states which can have a big effect on the EV.

For instance, say Kerry is ahead by 51-49 in FL with a 60% chance of winning. If he falls behind by 49-51, it will not effect his average EV by +/-27 votes, since he still has a probability (40%) of winning the state and will win approximately 40% of the trials. His average EV may drop by 2.

Your argument is silly.

[DemocratSinceBirth]

6. You Couldn't Submit Your Model To One Political Science

professor in the country and get him or her to endorse it...

Do it and I'll retract my statement...


You are basically saying Kerry's going to win in November, errrrrrrrr, wait there's a 97% chance he will win...

Lots of things can happen between now and November and even if you are saying based on polling that if the election was held tomorrow Kerry would win that wouldn't be true because there are fluctuations is the various polls and neither Kerry or Bush have large enough leads in the states to give them an Electoral College vote majority...Neither Kerry or Bush have leads outside the margin of error in enough states to give them 270 Electoral College votes...

I realize you like to do these exercises but they are pretty sloppy social science.

As Larry Sabato, professor of Government at the University Of Virginia says " This (election) is close. Anyone that believes this election can not go either way is too partisan to be helped."

http://www.centerforpolitics.org/crystalball/2004/president/electoratemap.php

[Zynx]

7. Sabato is unquestionably pro-Bush.

If you've ever listened to him on Fox, you would know what I'm talking about.

[DemocratSinceBirth]

8. I Don't Care About Sabato's Bias...

I just question the efficacy of the seminal poster's model....


He might as well ask the magic eight ball who will win....

[yardwork]

10. You know, it's ok to disagree

but you're getting a little harsh in your comments here. Lay off the poster. You disagree. Point made. Drop it.

[DemocratSinceBirth]

12. You Need To Read The Whole Thread...

Maybe I was a bit harsh but his model is as useless or useful as a ouija board....


If the poster really believes there is a 96.65% chance of a Kerry win he should sell all his belongings and lay it on a Kerry win at Harrahs or Vegas....

[TruthIsAll]

9. I don't NEED anyone's endorsement, not even yours.

I have been following elections since 1952.

How many have you witnessed?

I have three degrees in mathematics. How many do you have?

I was the first one on DU to warn about the Repuke voting machines

Point out one mathematical error in the model. If you want to get into a theoretical discsussion regarding the statistical methodology, be my guest.

Otherwise, as Teresa would say:

Shove off!

[DemocratSinceBirth]

11. "I Have Three Degrees In Mathematics"

Well I better back off....


I'm not going to get in a pissing match about credentials on an internet chat board....


Again... I issue my challenge.... Find one political science professor in the nation that will endorse a model of voting behavior that says that on this day there is a 96.65% chance that will Kerry will win an election that will be held in November.....

[Tweed]

14. Yeah, I would be interested if you could get ANY Poly-Sci professor

to back this up. I've always been in awe of the amount of mathematically ability and dedication to the numbers, but overall, I think it's a waste of your time. So many different things can happen between now and Nov. 2nd.

[DemocratSinceBirth]

15. I Am Sorry For Some Of The Harsh Things I Said

but these deterministic models are silly....



You can not control all the intervening variables or to use layman's language alot of shit can happen between now and election day....

[Tweed]

16. Truth is all

I don't think the things you said were harsh. I agree with everything you are saying. Like the poster says himself, Truth is all, even if it hurts.

[DemocratSinceBirth]

17. I Wish He Was Right.....

I Wish He Was Right....


I Wish He Was Right...


"......"


Ad Infinitum.....


But none of us know what will happen on Election Day...


This isn't like 72, 84, 92, or 96 when we had a good idea of what will happen ; bad or good....

[Tarheelhombre]

2. Did this predictor exist in 2000?

And how did it do?

[newyawker99]

3. Hi Tarheelhombre!!

Welcome to DU!! :toast:

[TruthIsAll]

4. No, I have been working on the model over the last month

My only forecast assumption is the allocation of the undecided/other vote to Kerry.

Otherwise, the model relies on statistical probabilities, using the latest available state polling data from electoral-vote. com and 15 national polls from PollingReport.com.

Obviuousy, the probability, Vote% and EV forecasts change when I update the input (at least weekly, but maybe more often) as the new data becomes available..

[carpetbagger]

13. Your conclusion is not supported by your methodology.

The most you could say is that if the election were held on the days of the polls, the chances of a Kerry victory that day would be x% (if the assumptions made in sampling error and turnout are approximately correct or if the bad assumptions cancel each other out).

It's a long way to November, and I suspect there will be times when Bush will be ahead in the state and national polls. I wouldn't put any money on Kerry as a 1:19 favorite.