>> Has anyone figured out how to reconcile this huge Swing State lead for Romney <<
Sure. Just plug in some extremely high lead numbers for the five states that seem to lean toward Romney (CO, FL, VA, NH & NC), and some much closer nummbers for the other six states.
For example, let’s assume that a unweighted average for the five R-leaning states gives a R advantage of 58-to-42. Then assume that the six other states give Ø an unweighted average advantage of 51-to-49. The result is an R advantage of almost exactly six points.
[Arithmetic Note: You don’t even need Excel to run the numbers! Just compute (5*58+6*49)/11 and then compare with (5*42+6*31)/11. Rounded to one decimal place, the grand average lead for Romney is 6.2%.]
Now if you don’t like the simplifying assumptions above, you could try many variations to produce similar outcomes. You could weight the state-by-state numbers by state populations, by state electoral votes and/or by state vote turnouts in 2008 or 2010. Then if the lead for Romney is too high or too low relative to six percent, you could fiddle with different state percentages until you generate a grand average lead where the mean comes in right at Rasmussen’s number. All in a day’s work!
But that's the problem. Romney doesn't have leads anything close to 58-42 in the states he is winning. If he did I wouldn't be posing the question.
Your analysis is OK. But the required Romney margins are jut unrealistic.