Welcome to FiveThirtyEight’s weekly politics chat. The transcript below has been lightly edited.
micah (Micah Cohen, politics editor): Greetings, colleagues! For your consideration today: Do Democrats need to win the special election for Alabama’s Senate seat in order to have a chance to win control of the chamber in 2018?
Implicit in that question, obviously, is: Can Democrats win the Senate in 2018?
So let’s start off with this, from friend-of-the-site Sean Trende of RealClearPolitics:
He says he would buy Democrats to win the Senate at 30 percent. And that it would be 50-50 if Democrat Doug Jones wins in Alabama.
Would people buy at 30 right now?
clare.malone (Clare Malone, senior political writer): Yes. While I don’t think it’s a sure thing by any means, this environment is a lot friendlier for Democrats than Republicans.
harry (Harry Enten, senior political writer): Ohhh boy. I tell ya. This is where a formal model would really help. Here’s what we know: Incumbent senators of the opposition party (the party that doesn’t control the White House) rarely lose in midterms, and Democrats have two clear pickup opportunities in addition to Alabama (Arizona and Nevada). They need a net gain of three seats to get the majority, so the math is there. Of course, the Democrats are defending seats on some very red turf, including in Missouri, North Dakota and West Virginia, to name just a few.
natesilver (Nate Silver, editor in chief): Answer the question, dude.
micah: You answer it, Nate!
natesilver: A-N-S-W-E-R. T-H-E. Q-U-E-S-T-I-O-N. H-A-R-R-Y.
micah: Clare is the only brave one here.
natesilver: I’d hold at 30.
micah: 
harry: Holdin’ Nate.
micah: HARRY!
natesilver: HARRY WHAT WOULD YOU DO?
harry: Fine. Fine. I’ll sell at 30 percent and buy at 25 percent.
micah:
natesilver: Oh, give me a break.
micah: You’re holding, in other words.
natesilver: You’re like, “We don’t have a formal model, blah blah blah,” and then you’re parsing the difference between 25 and 30 percent.
micah: Clare, why would you buy?
And how contingent is your buy on Jones winning in Alabama?
Like, if I told you that Republican Roy Moore is going to win, would you sell at 30 percent?
natesilver: I’d just point out that if Moore wins, he’d probably get expelled, which would compel another special election.
clare.malone: As I said above, before all the bullshit equivocating, I think the Alabama thing, if it happens for Democrats, could really build some momentum.
Would I sell?
No.
micah: My prediction: Moore wins, he gets expelled. He runs again in the new special election and wins.
natesilver: Then gets expelled again?
micah: No.
I’m not convinced Republicans would even expel him to begin with.
clare.malone: It feels a bit like an intimidation tactic right now.
micah: Yeah.
natesilver: Moore would be a huge problem for Republicans if he stays in office. He’s not going to be cooperative at all with the GOP leadership. And he’s basically every liberal’s worst stereotype of a Republican, which isn’t great for the GOP brand. I think the expulsion threat is pretty real.
harry: Read up on your Powell vs. McCormack. Adam Clayton Powell was expelled by the House of Representatives, then elected again and seated.
micah: But here’s my argument for buying Democrats at 30 percent: They basically need one seat in addition to Arizona and Nevada. They might get that in a month. And even if they don’t, if it’s a super Democratic-leaning year, as we think it will be, I’d bet Democrats in red states will be mostly safe.
Moreover! I think people think too narrowly about what states could be in play.
Like, if Democrats have a +10 advantage on the generic ballot and it’s an anti-incumbent year, who’s to say Ted Cruz won’t be in trouble in Texas?
clare.malone: Welcome to Team Buy, Micah.
micah: TEAM BUY!
Defend yourselves, Team Sell!
natesilver: I’d buy at 30 percent on Democrats winning three or more seats. But they also have a lot of their own seats to defend.
I’m not Team Sell, by the way, I’m Team Hold.
micah: You’re Team Sell, Nate.
natesilver: I’m Team Hold, Micah.
clare.malone: The coward’s choice.
micah: You’re Team
harry: I mean, it’s pretty simple why you wouldn’t buy. Other than Arizona and Nevada, the most Democratic-leaning seat that’s up in 2018 and has a Republican incumbent is Texas. Beyond that, where can Democrats pick up a seat? There aren’t good choices. Maybe they have a shot in Tennessee if the Republicans nominate an archconservative and Phil Bredesen, the former governor, wins the Democratic nomination.
natesilver: Given that there are approximately 6 jillion Democratic seats up for re-election and only a few Republican ones, I think Democrats having a 30 percent chance of taking the Senate is pretty good.
harry: What Nathaniel Read just said.
natesilver: The fact that it’s as high as 30 percent indicates that things are going pretty bad for Republicans.
micah: The Cook Political Report rates the Texas Senate race as more solidly Republican than the one in Tennessee.
natesilver: So, traditionally people place a lot of weight on incumbency. Tennessee is an open seat, and Texas isn’t.
I think that incumbency is maybe a little overrated. The incumbency advantage has been dwindling. (We’ll have an article on this soon.)
micah: What if Utah Sen. Orrin Hatch retires, then Mitt Romney wins in Utah and caucuses with Democrats?
clare.malone: He won’t caucus with Democrats.
harry: What if my mother’s dog starts talking, Micah?
micah: So, Nate, you basically think Democrats need Alabama?
natesilver: No, I don’t. I think Texas is plausible-ish. And I think a second Arizona seat could open up for obvious reasons.
clare.malone: OK, what are seats that Democrats hold that we think they are in danger of losing?
Missouri.
micah: Indiana.
West Virginia.
I mean, here are the Cook ratings on Democratic-held seats:
clare.malone: Yeah, I mean … they could DEFINITELY get cut up in some of those.
natesilver: I think Missouri and, to a slightly lesser extent, Indiana are the biggest problems for Democrats.
The reason being that Claire McCaskill and Joe Donnelly had the benefit of running against really poor opponents last time.
clare.malone: Yeah. “Legitimate rape.”
micah: Joe Manchin may have enough of his own brand in West Virginia to be in OK shape. Right?
natesilver: Manchin is still fairly popular there, yeah.
clare.malone: But there’s also the fact that Team Trump could go all out and pool support against Manchin in that state if they wanted.
harry: I wouldn’t discount Florida myself, given that Rick Scott could spend more money than most Americans dream of seeing in a lifetime.
natesilver: Yeah, Florida — you could see Bill Nelson blowing that race, somehow.
Or Sherrod Brown in Ohio.
And to the point earlier about the incumbency advantage diminishing — that could hurt “generic” Democratic incumbents like McCaskill who don’t have their own brands carved out.
micah: I don’t really buy Florida. If it’s a really Democratic-leaning year, why would they lose Florida?
Ohio is more believable. But I think even that would go blue in a blue-wave year.
clare.malone: The Republican candidate in Ohio (Josh Mandel) is … not super likeable. But who knows.
harry: Democrats in Florida are heavily dependent on a Latino vote that may not show up for a midterm.
natesilver: Democrats probably wouldn’t lose Florida, Micah. And it’s true that all of these outcomes are correlated. But sometimes individual races deviate from the trend, like when Democrats won by a huge margin in Michigan in 2014 despite having a really bad year everywhere else.
New Jersey is also a trouble spot for Democrats with the Bob Menendez trial.
clare.malone: Yeah, Menendez is very unpopular there.
natesilver: He’d probably lose a primary challenge — New Jersey Democrats tend to be fairly pragmatic — but it’s a wild card.
harry: My guess is New Jersey ain’t gonna happen for Republicans. It has a powerful state party that will get Menendez to lose the primary or step aside if necessary.
natesilver: So maybe there’s a 50 percent chance that Democrats win 3+ seats, but some of those times, they also lose one or more seats of their own. Which puts us at 30 or so.
If I were actually betting on this stuff, I’d also want to know how the contracts handle post-election party switches.
If Maine Sen. Collins switched parties, for instance, I think it would be right after the midterm and not before.
Sorta like Jim Jeffords or Arlen Specter after the 2000 and 2008 presidential elections.
clare.malone: Do we think Alaska Sen. Lisa Murkowski would switch parties post-midterm? Or caucus with the Democrats?
micah: That would be exciting!
harry: And to think, in 2011, I believed that politics had become boring and predictable.
natesilver: Collins is more likely, just because she’s more out of step with the Maine GOP, which is getting more like Gov. LePage.
clare.malone: Yeah, that’s true.
We’re just considering WILD CARDS!
natesilver: But if the Senate balance is 50-50 after the election — and let’s say the Republicans have really gotten slaughtered in the House, so the mood of the country is pretty clear — it becomes very tempting to switch if you’re a pivotal Republican senator. You arguably have more power that way (although you have a lot of power either way).
micah: OK, so party-switching and wild cards aside, does your read on Democrats’ Senate chances basically come down to how you think red-state Democrats will fare?
harry: I think that’s right.
natesilver: I mean, it also comes down to Alabama, where I’m slightly more skeptical of Jones’s chances than the consensus.
And it comes down to where you come down on Arizona and Nevada, on the spectrum between “toss-up” and “leans Democrat.”
I think both are leans Democrat, for what it’s worth.
micah: So, my read, to take just one example, is that Harry and Nate are looking at these races like gubernatorial races, which tend to be more about state-level concerns. But we know Senate races tend to be more nationalized, and the odds are that Trump is going to be super unpopular in 2018, perhaps with a much-ramped-up Russia investigation.
clare.malone: I’d agree with that.
micah: Why wouldn’t we expect the 
to hit the Senate?
natesilver: Even if you have a wave, Democrats might only gain two seats.
That’s the point.
In a non-wave year, they might lose six seats.
So the wave is what takes them from -6 to +2.
It’s a realllllllllyyyyyyy bad map for Democrats
micah: That seems circular to me.
natesilver: OMG
harry: I’ll just drop this in here.
micah: “Democrats will have a really hard time winning races, so they won’t win many races.”
OK, so let’s go back to Alabama …
harry: I love Alabama.
micah: Imagine it’s Dec. 13, and Doug Jones is the senator-elect from Alabama. I give you even odds that Democrats take the Senate back in 2018. Buy, sell or hold?
And remember, if you hold, you suck.
clare.malone: I’m holding and then selling once all the suckers in the betting markets get amped.
micah: That’s the right answer.
natesilver: 50 percent seems in the right range to me, if Jones wins.
harry: Apparently, you suck.
natesilver: Apparently. But it’s actually pretty hard to estimate this stuff without a model
because of how the outcomes are correlated, etc.
micah: We should build a model.
natesilver: Too soon.
harry: We should build a motel.
clare.malone: OK. Let’s cut through the crap: Do you guys think Jones is going to pull this out?
micah: Yes.
natesilver: If there’s no write-in bid, then I think Moore is still the favorite.
harry: Wait. That’s trash. Do you think there will be a write-in bid?
natesilver: Well, I have a whole freaking article about that, which I’m filing to Micah.
Basically, I think all the other outcomes are so bad that it doesn’t hurt the GOP to try a write-in bid, even though it probably helps Jones.
harry: For those wondering, PredictIt has Jones shares selling at 42 cents, as I’m writing this. Moore is at 41 cents, Luther Strange at 6 cents, Mo Brooks at 2 cents and Trip Pittman at 1 cent.
So the conventional wisdom seems to be that Moore won’t win?
micah: Final question: Does the outcome in Alabama matter solely in terms of the seat math? Or would a Jones win tell us something about the political environment? Or would it have more nuts-and-bolts consequences?
harry: Moore would win this race if the environment was neutral to pro-Republican.
clare.malone: I think a Jones win — which might be dependent, as Nate said, on McConnell and national Republicans trying to screw over Moore — would be yet another little battle in the “establishment vs. Bannonites” or whatever we’re calling that emerging wing of the party.
The wing that now makes the tea party look moderate in tone.
natesilver: Right. It would tell us a little something. But, again, mostly it would tell us things that are consistent with what we already know. And people are liable to over-interpret the difference between, say, a 3-point Moore win and a 3-point Jones win — both of which would count as a really bad performance for Republicans, but either of which could sort of be blamed on Moore also.
So basically I think the actual consequences of the Alabama race are larger than the predictive info it contains. In contrast to, say, the Georgia 6th or something.
harry: Right, I think Jones winning is both a sign of the environment and of candidate quality. But yeah, one Senate seat is worth a whole lot when the majority party is at 52 seats.