Swing state polling still makes electoral college arithmetic hard for Romney

In national polls, Obama and Romney are virtually tied. But checking numbers in battleground states, the picture changes

There are 11 days left in the presidential election, and the picture is static. Mitt Romney and President Obama remain knotted in the national vote. Obama, however, maintains an edge in the all-important electoral college. The lead is small, but seemingly solid.

I have written about three paths for Romney to win this election. These paths all assumed that Romney wins Colorado, Florida, and Virginia.

Colorado is a toss-up at this point. The Real Clear average has the race dead even. So, Romney can't really count on that state. Florida is a state where I feel fairly confident that Romney will win. He holds a near 2-point lead in the Real Clear Politics average.

Virginia is a tougher state. The Real Clear average has Romney ahead by about a point, and the conventional wisdom is that Romney is doing well.

That leaves us with an electoral vote score of 248 for Romney and 238 for Obama.

Even if Romney were to carry Colorado, I'm fairly convinced that he's likely done in Nevada. There has only been one Silver State poll this entire year with him in the lead. Nevada early voting numbers look fantastic for Obama. He leads by 2.5 in the Real Clear average, and in every poll in the average.

The fact that Romney can't win in even a single poll in Nevada is a key point. There have been five presidential state contests where all the averages (Pollster.com, Real Clear, etc) were wrong since 2000. That translates to a 97% success rate; the equivalent conversion rate in 2004 and 2008 was 98%. You can say, though, that averages do occasionally get the result wrong.

The chance of every single poll getting the result incorrect in a general election is far less likely. In all presidential state contests except one since 2000, there was at least one poll where the eventual winning candidate led even as the average had him behind. That was New Mexico in 2000 where two of the polls showed a tie and the other one had a small Bush lead. Bush lost that contest by less than 400 votes.

Thus, in only one in 150 of the state contests in the last three presidential elections was there an instance of a candidate winning a state where he didn't hold the edge in at least a single poll in the closing weeks of the campaign.

This is a major problem that's plaguing Romney in multiple spots.

There are a few more examples of a candidate winning out of nowhere in Senate and gubernatorial contests, but these include either third-party candidates (such as the Illinois gubernatorial election in 2010) or large Spanish-speaking populations (such as the Nevada senatorial in 2010). Only the latter factor applies in one of the states Romney needs (Nevada), and effective exclusion of Spanish-speaking voters in polls is likely to benefit Obama, not Romney.

If Romney doesn't win in Nevada, then, he must win either Ohio, or Wisconsin and New Hampshire, or Iowa.

Obama's advantage in Ohio is only 2.3 points in the Real Clear average. Again, Romney doesn't lead in a single poll in the average. The best he can do is a tie in 11 polls conducted over the last week. Candidates in this position almost never win.

Wisconsin is the same exact story. Romney hasn't been ahead in a single poll since August, even though the current average has him down only 2.3 points.

Taking these states together and too-close-to-call races in Colorado, Iowa, and New Hampshire, my current map, if the election were held today, looks like this:

On that showing, Obama wins with at least 271 electoral votes.

That's not to say the map won't look different in a few days. I believe Romney can win in Ohio and/or Wisconsin on election day. He doesn't even need to be ahead in the averages at the end of the campaign. History suggests, however, that he's got to be leading in at least one of the polls in these two contests. That's why I'm waiting for a poll, any poll, not sponsored by a political action group that has Romney with an edge in either the Buckeye or the Badger State.

Now, there's always the chance that this year is different than every other year; I don't doubt that possibility. That's why you won't hear me say this is a shoo-in for Obama even if his state leads stay steady right until election eve.

But it's very difficult to look at the success rate of polling over the last few years and conclude that it's anything but a long shot for Romney – if the polls hold as they are.

© Guardian News and Media 2012