The awesome Sean Trende applies the type of electorate-makeup shift we saw in Virginia, New Jersey, and Massachusetts to this November’s Senate races, and finds some fascinating results.

If I’m counting correctly, he has the GOP picking up . . . 12 seats in the Senate.

Now, breathe into a brown paper bag if necessary, and recognize that you’re not going to see an even, across-the-board shift like the one this exercise supposes.

I think GOP wins in Washington, Wisconsin and perhaps even Oregon are possible, but I don’t think you’ll see the numbers come out the way they do under Trende’s formula. Some Democrats are going to do much better than average among independent and Republican voters; West Virginia governor and Senate candidate Joe Manchin is a classic example of this. And note that under this formula, Carly Fiorina wins California by a hair, Linda McMahon wins Connecticut by a hair, and Florida is tough to compare to past cycles because of the quirks of Charlie Crist’s independent bid.

Also note that Trende’s formula has the GOP losing Senate races in Delaware and Illinois. I think Illinois is a toss-up. I don’t think Mike Castle will lose, but we’ll have to wait and see if something shakes up this race.