Good site that analyzes RCP poll averages a bit differently.

Latestpollresults.com ^

[PubliusEXMachina]

Probability analysis by LatestPollResults.com using Real Clear Politics data Date/Time Bush Win Kerry Win 269-269 Tie 10-29-04 87.22% 10.97% 1.81% 10-28-04 76.58% 21.73% 1.69% 10-27-04 80.24% 18.12% 1.64%

[PubliusEXMachina]

Excellent site that takes RCP polls and weights the samples to make the results more accurate. Check is out and see for yourselves.

[cajungirl]

thanks, nice post

[GnL]

Any chance of deciphering your data for those of us who don't read ASCII code?

[Excuse_My_Bellicosity]

Interesting. I've always had a gut suspicion that these "nationwide" polls were actually San Francisco/Greenwich Village polls.

[ConservativeGreek]

That one is on my favorites now!!

[FreeKeys]

Thanks!  I have added it to my list at http://FreedomKeys.com/polls.htm 
Remember the "DUI discovery" worked against Bush by a LOT at the last minute in 2000.   Who knows what the 'rats will pull on Sunday night and Monday?   It ain't over until it's over.    So KEEP ON CAMPAIGNING !    Consider copying and emailing things like THIS again.

[timpad]

Great find, thanks for the post. I'll be very interested to see if their prediction is accurate. Interesting that only PA and NJ (NJ!) are "tied."

[PubliusEXMachina]

Sorry, hit the link and the info will be there for you.

[Joe Bonforte]

Try this instead:

Probability analysis by LatestPollResults.com using Real Clear Politics data
Date/Time	Bush Win	Kerry Win	269-269 Tie
10-29-04 12:01am PDT	87.22%	10.97%	1.81%
10-28-04 12:01am PDT	76.58%	21.73%	1.69%
10-27-04 12:01am PDT	80.24%	18.12%	1.64%

[1stMarylandRegiment]

I am taking a break from leaf-raking, so this is "by eyeball," not scientific.

Bush Electoral Votes

1 standard deviation (65% of probable outcomes) range is about 250 to 330

2 standard deviations (95% or probable outcomes) is about 230 to 350.

Median (most probable outcome) is about 291.

According to this calculation, 78.1% of the probable outcomes are Bush gets more than 270 Electoral Votes.

[SaintDismas]

Everybody is forgetting about that faithless elector in West Virginia. Heard about him lately?

[1stMarylandRegiment]

OK

Median outcome (highest probability) = (291-"1 faithless elector") = 290

[Joe Bonforte]

"According to this calculation, 78.1% of the probable outcomes are Bush gets more than 270 Electoral Votes."

Just for fun, I ran a different check. I took all the states listed on realclearpolitics.com as toss ups. I wrote a short computer program that tried all 128 combinations of the outcomes in those states, and totaled EVs for Bush and Kerry.

The results? 89 cases result in a Bush win, vs. 39 in a Kerry win. That's 69.5% for Bush.

Of course, this assumes all the states are really 50-50, which I don't believe. I think Florida will likely go Bush and PA likely go Kerry. Assuming those two cuts us down to 32 cases. Bush wins 27 of them and Kerry wins only 5. That's 84%+ for Bush.

Just some idle diddling in VB...

[Hermann the Cherusker]

An analysis is only as good as the data going into it. If you take garbage like the Zogby polls, Research2000 polls or ARG polls in the south, and shove it into the analysis, you come out with garbage.