This non-partisan prediction is based on current statistics from RealClearPolitics.com [RCP] -- which averages multiple polls -- plus a percentage of the Undecided/Other Candidate vote [the percentage not attributable to McCain or Obama].
Chuck Todd of MSNBC predicts McCain will win 70% of the Undecided vote [I'll post a source when I find one]. I use this as a guideline. Current polling statistics for Obama and McCain are readily available for each of the battleground states. But the percentage that cannot be attributed to McCain or Obama is a mixture of Undecided and Other Candidate preferences. 3rd party candidate polling statistics are not as readily available, so I understate Todd's 70% prediction before applying it to this non-Obama/McCain percentage: I award McCain 60% of this vote, and 40% for Obama. My rationale for these percentages:
Suppose a state is polling 8% as Undecided/Other Candidate [about average in the current polls]. If half of this is Other Candidate [about what I've seen in various polls], McCain's numbers increase by 2.8% [70% of the 4% who are Undecided], and Obama's by 1.2% [30% of 4%]. This gives McCain a net gain of 1.6%.
60% of the combined Undecided/Other Candidate percentage yields a 4.8% increase for McCain, and 40% for Obama yields 3.2%, which also nets a 1.6% gain for McCain.
This method results in a slightly inflated adjusted percentage for each candidate, but the difference between the two adjusted numbers should approximate the real world difference -- even when the Undecided percentage varies somewhat from this example -- and an accurate difference is all that matters in this game, because all we really need to know is who wins each state.
After adjusting the polled percentage for each candidate within a state, I award all of that state's electoral votes [EVs] to the candidate with the larger adjusted percentage. I perform these calculations only for Toss Up states as classified by RCP. Those are the only states where the Undecided vote can flip the results [see Minnesota in the current prediction]. I then add the Toss Up EVs to RCP's Leaning and Solid EVs for each candidate in each of the 50 states to determine a predicted winner.
Check back often, as this gets updated with the latest RCP numbers every day. And note how close the results are. If your candidate of choice isn't the predicted winner, just wait a few minutes.
Subscribe to:
Post Comments (Atom)
3 comments:
test
Bill,
From one New Englander to another, I've gotta admit, a considerable amount of due admiration for your ability to put all these numbers together. But don't let my compliment get to your head. It took me three times to get out of Math 101 in college and the priest who taught the final "last hope" class for math challenged guys like me was more than happy to pass me just so he'd never have to deal with me.
If only we can do the same with Obama--but back to his senate seat!
What's terribly irritating about all this "battleground" states stuff is how the rest of us gets ignored until all of a sudden one of the candidates pulls an Al Gore and forgets what one of our late and fellow New Englander told all younger pols: All politics is local. Had he not lost TN, it might've been Gore standing in the wreckage of the WTC. But maybe not. He might've formed an advisory cmte. to advise him on whether or not to heed his national security advisor, the Pentagon brass and nation that wanted ObL's and Sheik Omar's heads.
But Bush hasn't gotten either of those yet, either. Maybe that'll be our "October Surprise," after the "September we've gone socialist in two weeks surprise."
If only it wasn't just a bad dream.
Great post. Keep it up!
Always a pleasure, Steve.
Post a Comment