Obama 348, Romney 190: UAH astrophysicist, who got it right twice before, predicts electoral results

 There is little suspense left in our presidential horse race, even in most of the so-called swing states, says astrophysicist and statistician Wes Colley.
The way he measures the poll data and his deceptively simple median method, which have correctly predicted the last two presidential elections, show President Obama has locked up Ohio and Virginia and cemented a small, but stable lead in Florida.
Bottom line: Obama is well on his way to re-election.
"I think the operative word is disaster if you're a Republican," said Colley on Monday. "I think the operative word is confidence if you're a Democrat."
"These things can swing a lot. But it would have to swing a lot a lot."
Colley, who also runs one of the six computer rankings that help the Bowl Championship Series decide the top teams in college football, said things could change with an "October surprise." Barring that, his method has proven particularly good at accounting for momentary campaign bumps, minimizing the effect of biased polling and dialing down the margin of error to less than 1 percent in closely contested states.
Colley, who teaches statistics and physics-based modeling at the University of Alabama in Huntsville, said he spent 10 hours setting up code to capture all polls reported on realclearpolitics.com. He and his colleague, Dr. Richard Gott at Princeton, haven't been very wrong yet.
"In '04, we only missed Hawaii."
This week, their method places Obama with 348 electoral votes. Governor Mitt Romney trails with 190. The only place Colley sees room for Romney to gain electoral votes without great difficulty would be North Carolina.
"You just feel like (Romney's) got to climb a cliff to win Ohio."
For his Electoral Scoreboard, Colley looks at polls conducted within the last month. He then lines them up in order, from most in favor of Obama to most in favor of Romney. He circles the poll in the middle, or the median, and awards the state to the victor of that middle poll.
In Ohio, Obama won the last 14 polls. That's 14-0, so Obama wins the state. In Florida, Obama won 12 polls to Romney's 4. So Colley awards Florida to Obama. He compared it to scoring for Olympic divers, who forfeit their lowest and highest scores.
"The beauty is it kicks out the outliers," he said.
"We would say the numbers don't dance around," said Colley, contending pundits incorrectly place too much credence in the latest poll, despite the large margin of error attached to a single poll.
The more data, the more precise the measure. So while there may only be a couple of percentage points dividing Romney and Obama in Florida, the 16 polls shrink the margin of error by more than four times, said Colley, making it smaller than the margin between the candidates.
In 2008, Colley and Gott came within two electoral votes of predicting totals. However, the median method led them to miss their predictions for Missouri and Indiana. The two swapped places. And they also missed North Dakota.
If there is a weakness in the system, it's that at times the middle poll is a tie. Colley said that was the case in Missouri and Indiana four years ago. To force a tiebreaker, they went back to the most recent poll before the one-month window. The tiebreaker hasn't proved as precise.
In 2004, when they missed Hawaii, only one poll had been conducted there. And although they doubted the validity of that poll, both believing John Kerry would win the Rainbow State, they went ahead and stuck to their mathematical model and called the state for President George W. Bush.
But they're not seeing an abundance of purple states and tied polls this year. Currently, the state most on the fence appears to be North Carolina, where five polls favored Obama, one ended in a tie and three went to Romney.
Colley said Virginia and Florida have drifted from leaning toward Romney when he began modeling on Aug. 11. He said Romney saw a bump after the Republican National Convention, but that has faded. And come November, Romney will need an even bigger bump than he received from the convention, said Colley.
For 11 years, Colley has also run one of the six computer polls that help the Bowl Championship Series rank college football teams. His poll, which adjusts for strength of schedule, currently ranks Oregon State first, with Alabama fourth.
But even he doesn't see that holding up for long. He said Oregon State, given the workings of the ColleyMatrix, benefited greatly from beating the likes of UCLA and Arizona early. Asked who he thinks should be number one based on his personal opinion: "Oh, Alabama."
"But if people wanted to know what I think they'd put me in the Harris poll," he said.
He predicts Alabama will rise quickly as the team gets into conference play. "We're still in the shakeout period."
As for the presidential race, he said he'd like to see more polls in Missouri, where all three have ended in a tie. His tiebreaker method currently awards the state to Romney, but he's less certain about that one.
Of his past success with the Electoral Scoreboard, Colley said: "I think what we learned in '04 and '08 is the polls must be pretty good."