FYI, the graph in the original posting (and the "prediction") is from InTrade. The 90% probability is based on how people are willing to bet their money, not any kind of statistical analysis.
The idea is that bettors have the same information and are making informed bets. But, emotion always figures into betting, and drags the actual "percentage" up or down.
A better statistical analysis is the Monte Carlo simulations that Nate Silver does for the NY Times:
http://fivethirtyeight.blogs.nytimes.com/
After 100,000 simulations, it predicts that there's only a 20% chance that the Democrats will control the House in 2011.
I can't link in the graph for the Senate, but the prediction is an 81% chance the Democrats will control at least 50 seats in the Senate (with Biden, that's enough).
If you are a Democrat Senator and want to be a committee chairman on a responsible and/or important committee and it's 50/50 (with /Biden controlling in tie votes) you can make a deal with the Republicans the Democrats may not want to make with you. Remember, with Hairy Reed out it's Schmucky Chucky Schumer you have to deal with and he is totally unprincipled. ANY Republican majority leader is better than Chucky.
This is where RINOs come from ~ Democrats of value to Republican control in a legislative body ~ not to be confused with tired old Republicans no longer of value to the Republicans.
Problem is, if InTrade bettors base their bets on the same polls we see, they’re just mirroring the polls.