What I got wrong in 2025

And some new predictions for 2026

I’m going to have to ask you to trust me when I say that I wrote these predictions before school Christmas break and that therefore prediction 43 (“Maduro out in Venezuela”) was a bona fide prediction for 2026 even though it happened before the post ran. In terms of the substance of the Venezuela situation, I stand by what I said a month ago — namely that the Trump-Vance vision of a world where the US retreats from assisting allies like Ukraine but enthusiastically bullies our own neighbors is going to make the world a more dangerous and unstable place.

---

Each year, I try to make a bunch of predictions about the year to come and then check back to see how I did.

The main point of this exercise is to humble myself with a reminder that predicting the future is really hard. For the predictions I made about 2025, I offered 50 different theses and said I had 80 percent credence in each of them. Depending on who you believe, either 34 or 35 of them came true, which is a success rate of either 68 or 70 percent.

What really killed me was that I made 10 forecasts about foreign elections and only four of them came true.

I was much stronger on my American politics forecasts, which were marred by me stupidly not realizing that the Virginia and New Jersey midterm elections were for governor and the lower houses but not for Senate. There were special elections in both states, but none resulted in State Senate seats flipping parties. Given that there weren’t any non-special State Senate elections, it wasn’t a great idea to predict that Democrats would pick up seats in the elections that weren’t happening. In my defense, I think it’s pretty clear that if there had been more elections for State Senate in Virginia and New Jersey that Democrats would have picked up seats. That said, the main point is that I was dumb. Here’s how I did, and my predictions for next year: