At the moment, this is a nail biter. There is a theory that a strong showing by one party at the top of the ticket brings along those lower down. However, that theory does not apply this year for several reasons. I think it works better for Republicans than for Democrats, for many the top of the Democratic ticket is less inspiring than ideal for this to work (though for no good reason), and this is the oddest election year ever, so in expecting the expected, expect the exceptional.
The Democrats hold 46 seats, and the Republicans 54 (slightly simplified numbers). This is a year in which the Republicans stand to loose a seat or two even if they are doing well because so many of theirs are up for grabs, and a good number of those are in states where Democrats have a chance now and then. Remember when the Democrats lost the Senate a couple of years back? The same thing happened to them that year, but in reverse.
In order for the Democrats to have simple majority in the Senate (which will mean nothing unless the new leader grows a pair of tacos and leads the way to change the rules of the Senate to allow a simple majority for most purposes) four seats need to shift, and Hillary Clinton needs to be elected president along with her Vice President, because he will be the tie vote. Indeed, if the Senate ends up being 50-50, Kaine can say good bye to doing much else while the Senate is session because he'll be breaking ties every single day.
The Republicans running (as incumbent or otherwise) in the Republican seats representing Illinois, Indiana, Wisconsin, and Pennsylvania are generally regarded as toast. They will lose. But, Harry Reid's seat in Nevada is not so clearly going to remain a Democratic seat. Remember, people are stupid. People hate the status quo. They will therefore vote against Harry Reid even if he isn't running. Plus, it has been kind of an unexplained miracle that Nevada even had Harry Reid, Democrat, representing them.
The prudent thing to assume is that Democrats will take Illinois, Indiana, Wisconsin and Pennsylvania, (Yay, we win the Senate!) and lose Nevada (Oh wait, we didn't win the Senate :( ).
The remaining seats that are in play include Arizona and Florida. You may have heard North Carolina mentioned a while back. Fuggetaboutit. They will all remain in the R column. (But do see below.) So, this leaves New Hampshire and Missouri. If Democrats take Illinois, Indiana, Wisconsin and Pennsylvania but lose Nevada, they will have to win New Hampshire or Missouri in order to be even in the Senate. If they win both, they have a simple majority sans Veep.
Election Night Senate Bingo
Time zones, how fast counting happens, how close it is in a given state, how uneven the precincts that are counted early vs. late are, across the political spectrum, will all matter to the level of political entertainment to be gleaned from this year's Senate race, on election night.
As the evening develops, look for a Democratic takeover in Pennsylvania, Indiana, Illinois and Wisconsin, Eastern and Central time zones, fairly early polls. Pennsylvania and Wisconsin will be where you will be chewing off some nails, because they will not be clearly decided early. Or maybe they will be.
Nevada is in the west, they count up their votes slow in that state, and the race may be close, so we may not know about Nevada, which I'm predicting will be Republicans, until Wednesday. If I'm wrong on that, we'll have a nice bit of news on Wednesday AM.
New Hampshire and Missouri should have reasonably early results. DEMOCRATS WILL LIKLEY NEED ONE OF THESE TO WIN THE SENATE. And, if they get both, that's extra. So, look for that to be resolved by midnight. We'll know about Florida, likely Republican, early. If, on the other hand, Florida happens to go D, then all bets are off on everything else. Stay up late and watch Arizona, just in case. You'll want to see McCain's concession live.
I assume that New Hampshire will be a) very close and b) polls will be available early but c) did I mention very close? Also, New Hampshire has a steep a gradient across geography, so however counting works out there, expect a fairly high degree of volatility in the meaning of the count as the evening progresses. We will be biting some nails over New Hampshire.
Are these predictions based on any kind of model, or are you pulling them out of your nether regions?
I agree that Illinois, Indiana, and Wisconsin are very likely Dem pickups, Pennsylvania is a tossup, and Arizona is a longshot. But beyond that that, the polling I have seen disagrees with your predictions. In Nevada, Cortez Masto has opened up a lead on Heck, and that one looks like a better chance to stay blue than for any of Missouri, New Hampshire, North Carolina, and Pennsylvania. Meanwhile, polling in New Hampshire has been oscillating, but recently Hassan has opened a lead over Ayotte, so that one is leaning toward a pickup.
In North Carolina, Burr has a small (within the MOE) lead over Ross, but there is another factor: Gov. McCrory is up for reelection this year, and I am hearing that lots of voters, even some who normally vote Republican, hate him. Whether they will split their ticket, or take out their frustrations against other Republicans (many of these people also don't like what the state legislature is doing, but the state legislature is so heavily gerrymandered that it will almost certainly stay Republican, possibly with a veto-proof margin).
Rubio's lead in Florida is about the same as Blunt's in Missouri. But Rubio has been losing ground this month, and Florida is much more likely than Missouri to go blue at the Presidential level. I'd still rate Florida as leaning Republican, but it is still potentially flippable. I'd be surprised to see Missouri flip unless it's an across-the-board Republican rout.
Nevada is in the west, they count up their votes slow in that state
'cause all their fast counters work for the casinos.
And the fastest ones are busy counting cards.
or are you pulling them out of your nether regions?
No, Eric; Republican predictions are pulled out of one's nether regions.
"At the moment, this is a nail biter. There is a theory that a strong showing by one party at the top of the ticket brings along those lower down. However, that theory does not apply this year for several reasons. I think it works better for Republicans than for Democrats, for many the top of the Democratic ticket is less inspiring than ideal for this to work (though for no good reason), and this is the oddest election year ever, so in expecting the expected, expect the exceptional."
Isn't this the sort of thing math could be useful for?
I would ask Nate Silver that question...
Sam Wang at PEC gives the Democrats an 85% chance.
"Are these predictions based on any kind of model, or are you pulling them out of your nether regions?"
Are those my only two choices?
Anyway, nether regions.
For those interested in math, models, and if I bow down sufficiently to Nate Silver, these estimates are mainly based on Nate Silver's assembled polling data and a few current or recent articles. In other words, all poll based. I used 538 to eliminate all the races where nothing is going to change, and the races I talk about here are all the races that are "in play" in one way or another. Beyond that, I'm using current polling data and trends, nothing more. Simply hoping to get a discussion going.
This is very different from the presidential race. For the Senate, we have more or less one population that is being roughly and not very accurately estimated, as opposed to fifty. The Senate is not going to be too different this year than an other year: Unknown until it happens. A discussion of what might happen mainly serves to prime us all for election night.