Nate Silver suggests it's possible that races could break two points for Democrats and they might not lose the House. A week ago I looked at the results of the expert predictions, showing when the pollsters predict a toss-up the races split and don't go one way or the other.
Since Monday independent pollsters have released 17 House races polled within 4 points. These races could go either way. In 2006 and 2008 Republicans won 31 races by 5 points or less. Democrats won 34. It's certainly possible that where that center is could move a point or two in one direction or another, but it's likely that the two sides will split the toss-ups.
A simple way to look at the forecasts would be to say Republicans win all the races that the forecasters have leaning GOP, lose all the races that are leaning Democratic and split the toss-ups. If that were to happen, Larry Sabato, Stu Rothenberg, Charlie Cook, and CQ Politics would all have 41-43 Republican gains. Real Clear Politics would have 62. Of course you should expect the Republicans will win a few Lean Democratic seats and the Democrats will win a few Lean Republican. That'd make Republican gains bigger. Right now I see a Republican 233 Democratic 202 split based on my formula. Nate Silver has 231-204. While I really think it'll be 225-227 I'm probably being cautious.