We will soon learn which parts of Alabama need to be nuked from orbit.

Roy Moore winning would probably help dems flip the house next year. But Jones winning, and we might be able to start fantasizing about scenarios where dems could flip the Senate as well next year.

https://forumserver.twoplustwo.com/s...postcount=1564 :/

PredictIt has Jones at 30% chance to win. I was surprised he was that high, he really have that much of a chance?

Feel like the deplorables bring this home for Moore 10 times out of 10. This is Alabama.

Heh, I think he was about 33% right when the scandal broke, at which point I rushed to buy some shares and then he jumped up to high 40's. Then, ofc I got greedy and held on to the shares and they dropped back down so I guess I'm in for a penny, in for a pound at this point.

30% sounds fair to me.
The whole point of nuking from orbit is it's the only way to be sure. Leaving parts of the state intact is risky.

That is some punk rock **** right there. Much respect to this guy.

I don't have too much to add, I've never actually tried to model anything like this.

But unless I'm reading it incorrectly, most of those individual state Dem % being spit out seem insane.

Would it?

If people thought elections were in play then who knows what type of loons they would vote for.

I think the model encounters a problem because things are *so* favourable for Dems. i.e it keeps giving the Dem candidate extra points because she's an incumbent & then because she has the midterm advantage & right now it's about fine, but then it gives her extra points because the President, who is GOP, is massively unpopular. At a certain point there's diminishing returns which the simple model doesn't capture well.

That said, in my dataset (1990-2016) there are only 4 instances of incumbents losing when they had the midterm advantage. It's ... rare for incumbents to lose when the opposing party holds the White House. Two of the four instances were in 1998 and two were in 2002. Both times the opposing President was massively popular. The President of the United States is unlikely to be massively popular in November 2018.

So while for the aforementioned reason I suspect some of the Democratic numbers are optimistic, I also don't think the Virginia race is remotely close. Like, most people seem to assume because it's a purple state Republicans must have a 20% or so chance of holding the seat. I'm confident their chances are way less than that. CW massively underestimates the advantage of incumbency & of the midterm advantage (model says they're worth 9.25 & 3.25 points respectively). GOP winning Virginia requires Kaine to get caught in a massive scandal after winning the primary or some similar event. He's a massive favourite.

The one state I'm confident the model has wrong is Florida. If Rick Scott runs, which is likely, Bill Nelson isn't a 96.7% favourite. Other than I really do think it's a matter of adjusting numbers down a little. Dems win Hawaii 99% rather than 100%. Virginia 96% instead of 98.3%.

Think of this way: All Dem incumbents won in 2012 which was i) a general election year ii) they were running under a so-so popular President, iii) Some of them (Kaine) weren't even incumbents. Now it's a midterm year where they'll likely run against a massively unpopular president. In what world do they perform worse in 2018 than 2012? They could get caught in a scandal ... or run against a much stronger opponent, but .... yeah. Ohio isn't much of a nailbiter, even though it's PVI R + 3.33.

Edit: I think it also overvalues Claire McCaskill in Missouri. She arguably only won because of Akin's "legitimate rape" comment - so it's reasonable to assume she's less strong than the average incumbent. And her approval rating isn't great. I'd cap that race somewhere between Toss-Up & Lean Dem. But Dems are clear favourites even in North Dakota, Montana etc.

Too many incumbent D's are running in Trumpland for me to think anything other than R gain is reasonable despite the #'s.

The bama election doesn't really have to do anything with antiTrump per se as the only reason a D is within 10 points is because of how bad Moore is.

Nelson's not 96% or whatever in florida no matter who his opponent is.

Having states like Montana/Indiana/midwest etc that voted Trump being 85%+ Dem win this go around is completely absurd.

^^^ That's why the model has Dems winning an average of .28 seats, despite the incredibly favourable environment.

In my eyes most of the D incumbents in Trump states are way overvalued. Brown may be a favorite in Ohio but 96% is crazy. Indiana, Missouri, Montana all way too high. I almost think you need to incorporate state by state presidential approval but I don't know how much good data is out there state by state.

If you can get me state by state approval rating on the date of every general/midterm since 1990 I'll add it to the model. I'd also like state by state approval rating of each senator too - that would be really useful.

I'd like to know why you think Indiana & Montana are "way too high", other than "instinct". My instinct was the same as yours, but the data is extremely powerful.

BTW re. state by state approval ratings:

Down most in red states - https://fivethirtyeight.com/features...in-red-states/

and

https://morningconsult.com/2017/10/1...upport-remain/

Barely above water in IN & MN (Donnelly & Tester are both ~+20 while Trump is ~+5). Below water in OH (Brown is +19). Holding up surprisingly well in FL (+0, Nelson is +24). Very popular in WV, but then again, so is Manchin.

(FiveThirtyEight has Drumpf below water in IN & FL).

David Lee Roth lights the menorah
So do James Caan, Kirk Douglas, and the attorney for Roy Moore-ah

Timely. Hanukah starts tomorrow.

Just waiting for Kayla Moore to break out of the guitar at that presser:

Hanukah is
the festival of lights

Instead of one paid Soros Fake News plant fabricating lies, Roy Moore has eight crazy accusers

This is also assuming there's a libertarian candidate in those races--who takes up a decent % of R votes. Montana election recently Gianforte won by 6, plus another 6 libertarian. Yes the dem gov held on barely but I think he was extremely popular.

personally I don't think any of that matters as much as it used to, people are still voting trump's and moore's and not giving a **** as long he isn't the other guy.

Guess we'll find out shortly whether the FOX NEWS poll is right with you or all the other ones were right.

scott vs nelson polls were a flip last I looked a month ago, not 75%.

Luther Strange was +10 in approval and they still primaried his ass for ROY MOORE. That's a real threat to the centrist dems running like donnelly/manchin/mccaskill. Many dem voters want purity even if it means higher probability of losing the seat.

It's also making an assumption that voting won't be restricted (or further restricted) and I would be surprised if they didn't do it in at least a few states by then.

The NR is against Moore, so that means he is winning.

#NeverTrump

With regard to Dems like Tester and Manchin, it's worth pointing out that they aren't running as a random Dem in that state. While they both benefited from 2012 being a presidential election year, Tester and Manchin have been there longer than that and won several statewide elections.

Manchin won twice for governor, and twice for the Senate (special election in '10, regular in '12).

Tester won a squeaker in '06 and a comfortable won (4 points) in '12, although he did have the benefit of a Libertarian siphoning off 6.6%.

Nate thinks that's about right.

If someone wants a sweat for pennies, I'll take Moore over Jones, my $40 to your $20. Paypal.

McCaskill and Donnelly are the most vulnerable Dems. McCaskill is just kind of a meh candidate and Donnelly will have to fight against Pence, etc. campaigning hard against him. Dems will likely lose one of these seats at least.

Heitkamp/Manchin/Tester/Brown/Nelson are all popular enough and are definitely favorites to win again, but probably even money at least one of these loses. Tennessee will be in play because a really popular former Dem governor is running and will likely face off against someone super far right.

+1 sounds like a good bet, given Nevada and Arizona are likely pickups, there will be at least 1 loss from the above. It's just going to be impossible to hold all of those seats.

If Franken/Baldwin are in play, we are ****ed, so no need to even worry about those seats.

My model doesn't make any predictions about "off year" elections. And even if did, it's far better to use actual polling this close to the race. I agree that it's somewhere between a flip and a small edge to Moore. Lean a little closer to flip that most. Since Alabama has basically the same PVI as Tennessee, I'd make a generic Republican a ~2/1 favourite v. a generic Democrat in the current climate. But, Tennessee is significantly more "elastic" than Alabama, which the model doesn't account for. There's no way a generic Republican is only a 2/1 favourite in Alabama.

The model is off on Florida but there's no reason to think Nelson isn't a favourite v. Scott. Or he could be a flip. There's minimal polling and we're eleven months out. But I'd take Nelson v. Scott straight up if anyone wants.

Incumbents generally don't get primaried. The exceptions are normally unpopular guys like Flake & Heller. Primarying Joe Manchin is an incredibly stupid idea, but someone is doing it.. PredictIt thinks that she, or someone else has a 22% chance of winning the primary, which is why I included it in the model. McCaskill should arguably stand down for Kander but there's no evidence she, or Donnelly or Tester are getting primaried.

Extreme candidates winning primaries is a much bigger problem for the GOP than Dems.

QUOTE=wheatrich;53229066]It's also making an assumption that voting won't be restricted (or further restricted) and I would be surprised if they didn't do it in at least a few states by then.[/QUOTE]

This is true. A big X-factor.

Pence is not popular in Indiana. His reelection prospects as Governor were not great.

yup

Probably, yup.

Yup, Tennessee is one of the few states I think Dems are more likely to win that the model. Arizona + Nevada + Tennessee is a realistic trifecta for Dems to pick up the three seats they need. But then they need to hold on to all 26 seats they currently hold which is possible but unlikely. Too many bullets to dodge. Which is why they'd really like Jones to win tonight, to give them flexibility. A second vacancy is Arizona is excellent for Dem's chances. Jones winning + a second AZ vacancy still doesn't make Dems favourites, but if the political climate continued to improve the Senate would start to look like a flip.

Another X factor is the rumoured expose of a large number of congressmen (as alleged harassers). Let's see what happens. There could be a scenario where Dem (alleged) harassers resign and GOP don't. It could be alleged harassers refuse to resign in seats where the opposing party holds the governorship. So we end up with the same 34 races + an extra couple of open seats in like KY and MS. Who knows. The GOP has more male senators so in some sense they're more vulnerable. It could be a nothinburger but there could be like an extra six Senate elections. And anyone with credible allegations of sexual harassment in a scheduled 2018 election is probably likely to get primaried right? That's the sort of scenario where if Sherrod Brown steps down, Ohio becomes competitive (I'd still make Dems favs). Or Nelson steps down and Scott is a favourite. Let's see.

listening to morning joe on the way in to work (lol me), they had someone from Alabama state government on, something like secretary of commerce or the equivalent state level position. Morning joe played a montage of Moore's statements that should have been disqualifying, anti-mulsim, anti-gay, etc. The commerce guy says "these things Moore says are not what the majority of Alabamans believe". This kind of annoyed me. no dude. We dont know that yet. Alabamans are voting today on what they believe, and there is a pretty good chance that the majority due believe that nonsense.
I've heard a lot of interviews etc from Alabama where people are talking about how much progress has been made in the last decade, racially, economically, socially, etc. Maybe it has. So you know what, this is your chance to prove it.