I expect to see some convergence in our estimates here once Nate Silver publishes. 
Let it be known that I will not be changing my guess. Why? Because I am the laziest person you know.
Speaking of which, some analysis from 538 about House polling. You’ll want to read through it yourself, of course, but I read the bottom line as House races are currently following the generic ballot. Which is bad news for Republicans in districts gerrymandered to be barely in their favor.
I’m afraid it will shatter the prevailing optimism but here are 20 of the closest House races which the Democrats must win. (Sorry for the inevitable error(s) in this list.)
Key: Name of D challenger where known, State-District Incumbent (R/D) % last received by R, status of incumbent if not running (R challenger in parens when different from incumbent) *-Conor Lamb is incumbent in another district.
?? Arizona-2 Martha McSally R, running for Senate (??)
Josh Harder California-10 Jeff Denham R
Katie Hill California-25 Steve Knight R
Gil Cisneros California-39 Ed Royce R, retiring (Young Kim)
Harley Rouda California-48 Dana Rohrabacher R, 58.3%
Mike Levin California-49 Darrell Issa R, retiring (Diane Harkey)
Jason Crow Colorado-6 Mike Coffman R
?? Florida-26 Carlos Curbelo R
Abby Finkenauer Iowa-1 Rod Blum R, 53.7%
?? Michigan-11 Dave Trott R, 52.9%
?? Minnesota-1 Tim Walz D, running for Gov (??)
Angie Craig Minnesota-2 Jason Lewis R
?? Minnesota-8 Rick Nolan D, running for Lt.-Gov. (??)
Tom Malinowski New Jersey-7 Leonard Lance R, 54.1%
Antonio Delgado New York-19 John Faso R, 54.0%
Anthony Brindisi New York-22 Claudia Tenney R
Susan Wild Pennsylvania-7 Charlie Dent R, resigned (Marty Nothstein)
Conor Lamb* Pennsylvania-17 Keith Rothfus R
The D’s need not win all 20 of these seats to win the House, but they must win an overwhelming majority of them — 15 out of 20 will certainly NOT be enough.
As shown, 18 of these 20 seats are presently held by an R.
I hope those in the know will keep us informed of D chances in these key elections.
There’s a gaping ambiguity in the rules.
Will the election results be based on polling results announced by CNN and other news sources in the day(s) immediately after the election? Or will the results be based on election winners finally certified by their states and permitted to take their seats in January of 2019? In the former case I’d guess 220 D’s in the House, but 217 in the latter case. I’ll wait till the rule is clarified before amending my contest submission.
Do you really believe this? What kind of shenanigans do you think are going to happen between election day and early January?
This would be absolutely Earth shattering if it went down like this with control of the house slipping away like that.
Furthermore, I think even one race being miscalled like that is pretty unlikely let alone three. I’m curious to hear why you think this might happen.
Note that the contest calls for me to pick a modal value, not an arithmetic mean. If we suppose the GOP will cheat on one seat if and only if they need it to get to 218, then a 217-total for the D’s becomes about twice as likely as, say, 216 or 219.
There are various ways for them to cheat. Most simple would be to claim fraud and that the incumbent must remain seated until a secure election can be re-held, but I think there are other ways. The incentive for R’s to retain House control is huge; for one thing this is the House that will approve census procedures and R’s already plan to subvert its counts for political gain.
I’m not sure why you find Republican mischief unlikely. I recall the apocryphal tale of the frog who didn’t notice the water was boiling hot because it warmed so gradually. The GOP is now openly operated by criminals to serve criminal goals but some Americans haven’t caught on because this developed gradually over decades.
FWIW 538’s now released House analysis is pretty much exactly in line with what I said almost 7 months ago (228D with a 54%D i.e. roughly D+8 ) … their “lite” is currently at 228D with a D+7.6, their “classic” is 230D with D+7.6, and their “deluxe” is 226 with the same D+7.6.
So clearly their current prediction is not getting me to move from my current bet 
… anyone wanting to alter theirs based on 538’s take?
I had 228 and 54% as well…so, no, I won’t change it based on 538. 
Wait… Jonathan Chance’s summaries list you at 56%, not 54%. You were referring just to Nate Silver (and me), then.
My optimistic guess of 241D and 55% is still within the 80% confidence range, so my fingers are still crossed. Of course, it was only a guess.
Time to update my estimate.
House = 252 D vs 183 R
Senate = 49 D vs 51 R
Popular Vote = 56% D vs 44% R
My first estimate looks pretty good, all things considered, but I think I got the scaling between House seats and vote margin slightly off. I’m keeping a similar “blue tsunami” prediction, but fine tuning it to be more self consistent.
Yes, I noticed that at the beginning. Here was the response:
Dang if you aren’t right! I misremembered. Still sticking with my current bet (and my early position in case of a tie) but you do get the honor of hitting the 538 number dead on!
When Silver predicts D+7.6 for the popular vote margin, that means something like (53.0 D; 45.4 R; 1.6 Other), right? Which translates to 53/(53+45.4) = 53.9% for this contest?
223 is pretty close to Silver’s median/mode. I don’t think Silver has posted a clear-cut Senate prediction yet, but one article makes 50-50 look quite likely.
I’m sitting very pretty with my 223/50/53 I think. How do you like my new title? :—
septimus
2018 Midterm Prediction Winner
No change to my contest submission, please; but thanks for asking.
House: D=221
Senate: D=49
Total Vote: D-52% R-48%
I want to wait a while, but if I were making an entry right now, that would be very close. But too many crazy things can happen between now and November. Agree on your second tie breaker.
You can change your guess at any time. Earlier post is used as a tiebreaker. No penalty for guessing early.
That’s what I get for not reading closely.
OK
House
D 220 R 215
Senate
R 53 D 47
Popular Vote
D 54% R 46%
You also could have used governorships as a tie breaker.
Eh, I chose to keep it federal.
So…
Does the recent unpleasantness in the courts, vis-a-vis Trump, cause anyone to rethink their position? Or is it just already built into the brew, so to speak?
Which unpleasantness do you speak of? I can’t keep track.
If it’s the illegal North Carolina districts, then you have a decision to make, Sir Moderator. Do you wait for the Supreme Court to allocate N.C. seats, but then punt if they split 4-4? Or just go with somebody’s (whose?) best guess on Worrisome Wednesday?