The Democrats currently have a 78-seat margin in the House of Representatives. That means they could lose a net of 38 seats (half of 78, less one) and still control a majority of the chamber.
CQ Politics, which does terrific work, has identified 59 competitive races involving Democratic Representatives. Conversely, there are 41 competitive races involving Republican Representatives.
At first, this math looks pretty decent for the Democrats. If each side won one-half of the other's competitive seats, the Democrats would lose a net of 9 seats, and their majority would be reduced from 78 seats to 60.
There are two things, though, to bear in mind. One, in "wave" elections, it's quite common for the "waving" party to defend nearly all of its competitive seats, allowing it to concentrate solely on offense. In 2006, the Democrats didn't lose any of their seats. In 1994, the Republicans lost only four of theirs, and all were cases in which the incumbent was retiring or running for higher office.
Secondly, in "wave" elections, at least a handful of seats that nobody is expecting to become competitive usually up that way. In a post a couple of weeks ago, I mentioned the example of Jim Leach's seat in IA-2, which he'd held for 30 years and which nobody was really monitoring -- except for poker players like me, who were watching the seat because of Leach's role in pushing forward anti-gambling legislation in the House. But that seat was won in 2006 by a virtually unknown political science professor, a Democrat named Dave Loebsack.
Of course, in both 1994 and 2006, the opposition party was tremendously well organized. We've seen nothing as smart as the Contract with America that the Republicans put together before 1994, and nothing as impressive as the 50-state strategy that the Democrats had working for them in 2006. On the contrary, the Republicans have something of a power vacuum and weren't done any favors by the McCain campaign, which put little emphasis on ground game and did not help to develop the party's voter lists. Also, the Democrats have pretty significantly outfundraised Republicans in House races so far, which is a pretty good leading indicator.
Nevertheless, in wave elections, things have a way of just happening. We don't know whether 2010 will resemble 1994 or 2006. But if health care falls apart and the economy remains sour, I wouldn't put too much limitation on the Republicans' upside.