An easy way to predict the election

[President von Cat]

Just use the "Presidential approval/disapproval" numbers.

For example, the current RCP average is 47.2% approve, 48.4% disapprove.  If the election were held today, Obama won't get a single one of those 48.4%.

Not true. There are times when I disagreed with the president's performance, but still intended to vote for him.

The latest of the polls -- by Gallup (all voters -- scratch that, it includes even unregistered people) and Rasmussen (likely voters) are 47-46 and 46-52, respectively.  http://www.realclearpolitics.com/epolls/other/president_obama_job_approval-1044.html?utm_source=rcpwidget&utm_medium=widget&utm_campaign=chart_1044

Then you can take the "undecideds" which normally break 2:1 against the incumbent.  If the 47.2% to 48.4% are accurate, that leaves 4.4% undecided.  If you break it 2:1 for the challenger, that would give Obama 48.7% (47.2% plus 1.5%) and the challenger 51.3% (48.4% plus 2.9%).

This is wrong for the point I listed above, and where did you get that little factoid about how undecideds break? Dick Morris? Sounds suspiciously crackpot to me, so I'm gonna go with Dick Morris..