Let’s run through the basics. As of now, the Republicans control the House (233-199) and the Democrats control the Senate (55-45). There is little doubt the GOP will hang onto the House (most people think they will pick up some seats there), but the Senate could go either direction.

There are 36 Senate seats being contested, and of those 11 are competitive. The Republicans need a net gain of six seats to take control of the chamber and it looks like they have already locked down wins in South Dakota, Montana, and West Virginia, which gets them half-way there unless the Democrats are able to pick off a Republican incumbent in Kentucky or Kansas. In order to secure the Senate, the GOP needs to win at least three of the following contests. If the Republicans lose either Kentucky or Kansas (of the two, Kansas is far more likely to flip), that number goes up to four or five.

• Alaska
  • Begich (D) vs. Sullivan (R). As of now, the race leans Republican, but polling in Alaska is complex. Leaning Republican.
• Arkansas
  • Pryor (D) vs. Cotton (R). Cotton’s lead has grown against the incumbent Pryor making this a very likely Republican win.
• Colorado
  • Udall (D) vs. Gardner (R). Gardner has pulled into the lead recently, putting incumbent Udall on the ropes, but Colorado has been trending Democrat otherwise. Leaning Republican.
• Georgia
  • Perdue (R) vs. Nunn (D). The open seat is very tight with the lead flip-flopping during the past several weeks. If neither candidate wins a clear majority of the vote, the contest goes to a run-off. Toss-up.
• Iowa
  • Braley (D) vs. Ernst (R). The open seat has shifted around, but Republican Ernst has taken a significant lead during the past week. This leans Republican at the moment.
• Louisiana
  • Landrieu (D) vs. The Field. Louisiana runs an open primary for election day and if any candidate secures a majority, the race is over. If no candidate is able to do so, a run-off happens. As of now, it seems very unlikely Landrieu will win a majority, and polling has her losing a run-off to Cassidy, her most likely Republican challenger. This race is a toss-up.
• New Hampshire
  • Shaheen (D) vs. Brown (R). The Democrat incumbent is a sliver ahead of Scott Brown, the former Massachusetts Senator, though her lead is thinning as we get closer. Unless the Republicans enjoy a very strong night, expect Shaheen to hang on. Leans Democrat.
• North Carolina
  • Hagan (D) vs. Tillis (R). The Democrat incumbent should still be considered the favorite, but as in New Hampshire, the Republicans could pull it out if things break late for the party. Leans Democrat.

If Republicans win all of the states that are leaning Republican at the moment, and win in Kentucky and Kansas, they will pick up seven seats and secure the majority in the Senate. I think that is the most likely outcome as of now.

There is some reason to think, however, that polling is not capturing Democrats very well. The party’s core supporters, racial minorities and young voters, are difficult to reach, so pollsters must extrapolate more heavily to estimate behavior for those groups. If those extrapolations are biased against Democrat turnout, the Republicans could be in for a surprise on Tuesday. This may not matter in Arkansas or Iowa, but it could be decisive in Colorado and Georgia.

An additional concern for the GOP is that Democrats have dominated in the turnout game during the last several election cycles. They have harnessed technology more effectively to target potential voters, reach them, and get them to the polls. Republicans have spent money to close the gap, but until the technology hits the real world on Tuesday, we will not know whether those investments will lead to real dividends.

Finally, Republicans do have demonstrable advantages heading into Tuesday. The party’s base is older, more engaged, and more likely to vote. The GOP has, then, a built-in advantage in mid-term contests when turnout is usually lower. Given President Obama’s relative unpopularity, there seems to be little the White House can do to buck that trend.