Rasmussen reports that Bush's confidence numbers are declining for security/Iraq (no surprise), but are actually going up for the economy, probably because of the jobs reports.
Ray Fair's mathematical model which is based largely on the economy, predicts a Bush win. It has successfully predicted 20 of the past 22 elections.
The CW until now has always been: if this is a security election, Repubs win, if an economic election, Dems win. Now the exact opposite may be true.