Any way you slice it, 2012 looks like a tough year for Senate Democrats. They have 21 Democratic and 2 independents who caucus with them up for re-election. Even excluding the 2 independents, Democrats have a -11 margin of seats to defend. That's never been easy. Here are all the times since World War II one party has had to defend 8 or more seats than the other party has had to defend.
At first glance this doesn't look so bad. While there have been times when a party defending a significant number of seats has lost a lot, it doesn't happen all the time and the average is only 5.9. Six of these elections are mid-terms. Comparing a mid-term to a Presidential election is apples and oranges because the election and electorate are different.
The six Presidential years are actually better for the party with more seats than the mid-terms. Excluding the mid-terms favors the Democrats. Of course there are two Presidential elections which have circumstances that won't be repeated. In 1963, a popular Democratic President was assassinated. Strong sympathy for him prevented his party's losses. In 1976 the Republicans were coming off Watergate and couldn't take advantage of the differential. The other four Presidential elections have one thing in common, the Presidency changed parties. In 2008 and 1968 the Presidents were far more unpopular than Obama is today. The other two years, 1960 and 1980, the opposition candidate proved to be far more popular than he was thought to be even a year before. His strength led to his party's strong showing.
So let's approach this from a different perspective. How did the President's party do when the President faced re-election. Here I've excluded 1964 and 1976 because of circumstances and because that President had no initial election baseline.
This contains an interesting mix of Presidents. Eisenhower, Reagan, and Nixon were re-elected in landslides. Bush 43 and Clinton won closer elections. Carter and Bush 41 lost. It's difficult to say that Obama, Nixon, and Reagan underperformed in their election campaigns. While they didn't take more than the differential, they all made significant gains. If this were a new election, such a big differential wouldn't be as important.
As expected, almost all the Presidents underperformed two years later. It's tough to regard Clinton's +2 very highly because he lost 8 seats.
Reagan, Nixon, and Eisenhower didn't have the huge differential Obama has in 2012. Even their landslides were unable to prevent some losses due to a negative differential. Jimmy Carter and Bush 41 lost re-election and that hurt their party in the senate. Obama's coattails in 2008 only resulted in 20 Democratic wins vs. 15 Republican wins. If re-elections are at all telling, even a strong performance in re-election likely won't prevent a loss of 7-11 seats.
If you're looking for a more pleasant answer for Democrats, vulnerability is a good way to look. States were grouped based on how Republican or Democratic the state is. In 2010 Democrats only lost 75% of the most vulnerable states and half of the moderately vulnerable states. Repeating this they'd lose 8.2 seats. The Republicans, on the other hand, had 4 moderately vulnerable and 14 low vulnerability seats in 2010. In 2012, Maine and Massachusetts will be much better opportunities for Democrats than anything this year. Picking them up would limit the losses to 5 or 6.
My current analysis has the Democrats at a net loss of 5 seats. That'd be lower than these other historical indicators but still enough to cause them to lose the Senate.
Wait a second, a Democrat might say. If you count Lieberman and Sanders as Democrats the Democratic party went 24-9 in 2006. If these guys were good enough to win in 2006 why couldn't they be good enough again? Circumstances are far different. In 2006 you had an unpopular Republican President, an unpopular scandal ridden Republican congress, a disillusioned Republican base, and a war that energized Democrats. Obviously an unpopular President won't be a problem for Republicans. They only control one house of congress. At worst they're not likely to be that unpopular again. The Republican base seems unlikely to be disillusioned this time around. Democrats could once again be energized, but without the other factors winning anything close to 24 races is extremely unlikely.