Larry J. Sabato‘s estimated results:

For the House of Representatives:

[W]e are raising the total [from +44] to +55 net Republican seats. We consider 47 to be in the ballpark still, but more of a floor than a ceiling.

Which results, if correct, would produce a 233 Republican — 202 Democrat House of Representatives.

For the Senate:

The Crystal Ball has operated within a very narrow range all year. When others were projecting GOP Senate gains of just +3-4, we were already at +6. Depending on the primary results and other circumstances, we’ve landed between +6 and +9 in the last half-year. We have never gotten to +10, the number needed for Republican takeover of the Senate, and we do not do so in this final forecast either. To us, the number of GOP gains looks to be +8. Ten was always a stretch. …

In our pre-Labor Day analysis, however, we noted a historical anomaly: Since World War II, the House has changed parties six times, and in every case, the Senate switched, too. In five of the six cases, most prognosticators did not see the Senate turnover coming. (Only in 2006 did some guess correctly, including the Crystal Ball.) So if we have a big surprise on election night, this could be it, despite the pre-election odds against it.

Their predictions also have the GOP gaining 8 or 9 governorships, and at least a dozen additional state legislative chambers.