Monday, October 18, 2004

Electoral College Scenarios

Things are getting a little dicier for President Bush. Though it seems increasingly likely that Bush will win the popular vote, it also seems increasingly likely that he will lose some "cushion" states in the electoral college. Obviously, the latter losses are all that matter, so it's time to get a little concerned if you're a Bush fan. Let's work through some situations, starting with some reasonably pro-Kerry assumptions:

Let's assume Kerry wins Delaware, Connecticut, Washington, California, and Michigan. That gives him 181 electoral votes (I won't list all the certain states). He needs 270 to win (a 269 tie kicks it to the House, where Bush will almost surely win). So Kerry needs to pick up 89 more votes. I think Kerry is going to win in New Hampshire, New Jersey, and Maine, despite all being within the margin of error on some recent polls, giving him 23 more votes.

Now Kerry needs to pick up 66 votes from: Oregon (7), Minnesota (10), Pennsylvania (21), New Mexico (5), Iowa (7), Ohio (20), Wisconsin (10), West Virginia (5), Florida (27), North Carolina (15), Nevada (5), Colorado (9), and Arkansas (6). I consider all of these states to still be in play. They add up to 147 votes total.

If Kerry wins all of the above states, except Ohio, Pennsylvania, and Florida (the big ones), he'll have 79 votes, more than enough to win. This scenario is unlikely (I'd say less than 5% chance), but it goes to show that you don't absolutely need to win some combination of Florida, Pennsylvania, and Ohio to win the election.

Now let's say that Bush wins North Carolina in the above scenario. A recent Survey USA poll, largely unfavorable to Bush in other states, still gave him a 3 point lead here. Many other polls show it at about 7 points. Now we're at 64 votes for Kerry, and Bush wins.

I'm going to keep making educated guesses. Now let's say Kerry wins Pennsylvania, which the same Survey USA poll gave him at 6 point lead in. Kerry now has 85 votes, and Bush loses, even if he wins Ohio, Florida, and North Carolina. This is a tad more likely scenario, and again goes to show how Kerry's efforts in smaller states, combined with a win in one big one, gives him some chance of winning.

What kind of chance? Still not a good one. In the scenario above, Bush needs 20 more votes to pull of a victory. This is very likely to happen. Just winning Nevada, Colorado, and Arkansas will do it, and Bush is winning in all of those states, as he has been for a while.

So Kerry can't settle for just winning Pennsylvania in my scenario, even if he has a chance of winning with it alone. In reality he needs to aim for either supplementing Pennsylvania with any one of Nevada, Colorado, Arkansas, Ohio, or Florida. How likely is that? I give him better than 50% odds, if just slightly better. Call me a pessimist.

If Kerry picks off Nevada, Colorado, or Arkansas, Bush could compensate by winning a state like Iowa or Oregon. Iowa's the best bet there - that race is exceedingly close, equally likely to go to either side. But if Bush can't compensate, he loses.

Now let's say Bush keeps Nevada, Colorado, and Arkansas, but loses Ohio. This makes things tougher for the President, obviously. He needs to win West Virginia and New Mexico, which now lean slightly for him, plus add Iowa and a more unlikely small state, like Oregon. I give him less than a 25% chance of doing this. Or, Bush can win Wisconsin and any two of the states above. This is more likely to happen.

So Bush can lose Pennsylvania and Ohio, but still win the election by taking Nevada, Colorado, Arkansas, Wisconsin, and any two of West Virginia, New Mexico, Iowa, and Oregon. Overall, I'll give Bush a 33% chance of winning if he loses Pennsylvania and Ohio, an event that could very well happen.

But all of this analysis assumes one thing: Bush wins Florida. The same unfavorable Survey USA report has Kerry winning by 1 point now. If Bush loses Florida, Pennsylvania, and Ohio, he's sunk. If he loses Florida and wins Pennsylvania or Ohio, he needs to win Nevada, Colorado, and Arkansas, plus gain an additional 22 votes somewhere. If Bush wins Iowa, then he must win either of Minnesota or Wisconsin, and then win Oregon, or New Mexico and West Virginia together. Obviously, winning Minnesota and Wisconsin will also do the trick. Still, the odds are sinking pretty low at this point - less than 25% chance of Bush winning with a Florida and a Pennsylvania or Ohio loss, I'd say.

Of course there's many more scenarios to throw out there, and my percentage predictions are about as reliable as Mariano Rivera was tonight. But hopefully this gives some idea of where the candidates are at - assuming things keep going Kerry's way, Bush's chances are diminishing quickly. Keep watching those Florida, Ohio, and Pennsylvania polls people, and keep ignoring those deceptive national polls showing Bush with a lead. They may be accurate, but they aren't helpful.

Oh, and thanks to Gerry at Daly Thoughts for providing a great way to keep up with all of this. Check out his Electoral College Breakdown. It's well done.

No comments: