As a political commentator, I take delight in forecasting outcomes. I have no special talent for it and am frequently spectacularly wrong: unlike many fellow pundits, I once stubbornly clung to the ridiculous notion that former President Joe Biden would never step aside as the Democrats’ 2024 nominee. I nailed the 2020 election (and got most states right, only missing Georgia and Florida), but I bungled the 2024 projection. (That said, I still beat my co-host in the challenge, who didn’t seem to treat it with much seriousness.)

The hardest part about making predictions is separating what you want to happen from what is actually likely to occur. Commentators often feel compelled to advocate for their preferred outcome, as if shouting about it on television could make it come true. In some ways, this is a temptation libertarians can more easily resist, since most outcomes tend to feel suboptimal anyway. Once again, this holds true: consider the forthcoming 2026 midterms as an example.

It seems undeniable that President Donald Trump has squandered a lot of political capital by not narrowing his focus to the economy and by pursuing another unpopular conflict in the Middle East. It also seems obvious that the Democratic Party, once back in power, will do everything in its power to obstruct the administration’s most prudent policies—especially in tech and AI.

But that’s my forecast: I expect Democrats to reclaim both the House and the Senate.

I don’t have a precise House map in mind (the intricacies are overwhelming), but these are my Senate projections:

Here is my REVISED prediction for the Senate. For the last several weeks, everything has moved in Dems’ direction, and I now expect them to take it.

I still say Collins holds on. Michigan and Ohio seem lost for GOP, though. And I’m not super confident in the Iowa call. https://t.co/oph3Swtgzw pic.twitter.com/FtsrGxmukJ

— Robby Soave (@robbysoave) October 20, 2026

When I surveyed the map back in August, I assumed Ohio would be a Democratic long shot and Michigan could stay competitive for the GOP due to Abdul El-Sayed’s leftward stance. Now, it appears Republicans are headed to lose both states, in addition to Texas and Alaska. I remain convinced Maine’s formidable incumbent Senator Susan Collins (R) will hold on—the polls have misread her before—and I doubt Democrats will pull off Iowa. If Republicans manage to defend those two seats, the Senate would tilt 51–49 in Democrats’ favor.

But hold on…Trump isn’t out of the woods yet.

In a scenario where Democrats sit at 51–49, it seems plausible to me that Senator John Fetterman (D–Pa.) could switch his allegiance, effectively restoring Republican control thanks to Vice President J.D. Vance’s tie-breaking vote. Given his stance on Israel and similar topics, Fetterman has no future as a Democratic officeholder in Pennsylvania—he could never win a Democratic primary—and would likely need to negotiate with Trump before running again in 2028. One option would be for Fetterman to become an independent who caucuses with Republicans, in exchange for Trump ensuring no Republican challenger emerges against him in the future. Absent such a deal, Fetterman is likely to serve only one term, so if he wishes to keep his seat—an open question—he may need to take action. A 51–49 Senate would provide him the leverage he needs.

Those are my predictions. What about yours?

This Week on Free Media

I’m joined by Amber Duke to discuss the riots in France, Vivek Ramaswamy’s terrible campaign ad, and more.

Worth Watching

Here’s another unabashed plug for my upcoming debate with Katherine Mangu-Ward, Mehdi Hasan, and Jason Stanley! (Katherine is aligned with me.) If you’re in Washington, D.C., you can grab tickets here.