Using data from June, Wonkblog's model predicted Obama had a 67 percent chance of winning the election. At the time, I thought that was way, way too high. But the model was right and, going by the latest data, I was wrong.
The conventional wisdom on this election is that if the economy is front and center, Obama will lose. But that's not what the data tells us.
Overall, more models have Obama winning than not. But the two that don't should give Romney backers hope.