Will Democrats keep the Senate?

publishedabout 1 month ago
4 min read

Welcome to Nonrival, the newsletter where readers make predictions.

How it works

  1. Read the newsletter
  2. Make a forecast by clicking a link at the bottom
  3. In the next email, you’ll see how your forecast compares to other readers’
  4. Over time, you’ll get scores based on how accurate your forecasts are

In this issue

  • Follow-up: Readers don’t think Elon will lay off most of Twitter
  • Make a forecast: Will Democrats hold the Senate?

Thanks for forecasting. Send feedback to

Don’t bet on Elon laying off most of Twitter

Elon Musk owns Twitter now, and the firings have begun: CEO Parag Agrawal and several other top executives have already departed. There was also apparently a stunt where actors pretended to be laid off Twitter employees, carrying around boxes outside the office. But Musk has assured employees that he does not, in fact, want to lay off 75% of the company — as the Washington Post had reported last week.

And Nonrival readers believe him: Most of you think there’s very little chance that half or more of the staff is laid off by April.

Crowd forecast

Most readers put the chance at <15%:

But a few readers thought it was more likely than not, dragging the average forecast up to 27%:

Readers’ reasoning

4%: Cuts like this would destroy the value of the company. Not gonna happen
5%: Musk has a history of effectively manipulating public perception solely by making dramatic announcements--with zero follow-up (whether or not follow-up was ever intended).
10%: Musk will want a productive workforce - 70 or 50% will create antagonism, not productivity. I might believe a 30% cut, but 50% is quite unlikely.
10%: Dude's full of hot air most of the time. Why would this be different? It also seems just impractical as a strategy for "fixing" whatever is broken at the company. (Also, revenue per employee is a terrible measure that only encourages layoffs.)
10%: 10% is how much I trust anything that Musk says, and that may be 10% more than I should.
45%: I would've picked a much lower figure for Musk's stated 75% cut (I think that's pretty unlikely) but twitter was going to have to downsize anyway. I expect that Musk can cut more deeply into their content moderation crew than some outside commentators suspect without losing much revenue.

How your forecast compares

  • You said there was a [102322_FINAL GOES HERE]% chance of Twitter laying off half its staff or more.

  • You predicted that the average of readers' forecast would be [102322_CROWD GOES HERE]%. The actual average was 27%. You were closer than [102322_CROWD_RANK GOES HERE]% of readers.

Programming note: I’ve tried this follow-up section both as part of the Sunday email and on its own. I think it worked better as its own email so expect this to come separately going forward. Sundays will continue to be a new forecast, and the follow-up will come a few days later.

Will Democrats hold the Senate?

Make a forecast by clicking a link at the bottom of this email.


The US midterm elections on Nov. 8 will determine the composition of Congress for the next two years. The Senate is currently split 50-50, with vice president Kamala Harris providing Democrats the tie-breaking vote. 35 Senate seats are on the ballot—if Republicans pick up even one seat, they’ll hold the majority. (Republicans are favored to win a majority in the House.)

The president’s party almost always loses seats in the midterms. “Since the end of World War II, the president’s party has lost House seats in all but two midterms,” per Fivethirtyeight. But “the pattern is a bit more inconsistent in the Senate. Since World War II, the president’s party has either gained seats on net or at least avoided losing ground in six out of 19 midterms.”

Polls tilted toward Democrats over the summer, but that trend is fading. The Supreme Court's abortion decision and falling gas prices have been cited as explanations for Democrats’ polling surge; concern over inflation and a cooling economy could explain Republicans’ fall resurgence.

Polls are imperfect and in 2016 and 2020 they overestimated Democrats’ chances. For more on the ins-and-outs read Nate Cohn and Nate Silver.

Senate control will likely come down to close races in Georgia, Pennsylvania, Arizona, Nevada, and Wisconsin. Republicans have put forward several inexperienced, scandal-ridden candidates like Herschel Walker.


“In four of the most crucial races on the Senate map—Georgia, Arizona, Pennsylvania and Ohio—quirky political newcomers blessed by Trump are underperforming. The combined picture has many Republicans privately grousing that the former President may cost them the majority despite otherwise favorable conditions—again.” —Molly Ball, Time
“Republicans could have put themselves in a much better position to win in Georgia, Arizona, and Pennsylvania just by nominating real professional politicians… But despite it all, I think Republicans are going to take the Senate.” —Matt Yglesias, Slow Boring
“If you’d asked me a month ago — or really even a week ago — which party’s position I’d rather be in, I would have said the Democrats. Now, I honestly don’t know.” —Nate Silver, Fivethirtyeight
“We continue to rate Senate control a toss-up… Republicans still win voters who prioritize the economy but haven't been able to grow their share of them. Meanwhile the Democrats have not been able to grow the number of people prioritizing the abortion — an issue that favors them.” —Kabir Khanna, CBS News (part 1, part 2)


Nonrival collected eight different forecasts that use different methods, from betting markets to statistical models to pundits' opinions, and put them in one chart:

For more data, check out Fivethirtyeight’s “generic ballot” roundup and its Biden approval roundup.


How likely is it that Democrats maintain control of the Senate (50 seats or more)?

​~10% chance​ ​​​​

​​​~30% chance​ ​​​​

​​​~50% chance​ ​​​

​​​​~70% chance​ ​​​​

​​​~90% chance​​​​

Deadline: Make a forecast by 9am Tuesday Nov. 1

Fine print: In the case of any ambiguity here I’ll follow Good Judgment Open’s decision on resolving this question.

​Want to learn more about Nonrival and about crowd forecasting? Click here.