Correction: Jiang's World War III Forecast Is Still Open
On July 29, 2026, this tracker marked Professor Jiang Xueqin's prediction P021 as wrong. That verdict was a mistake, and we have withdrawn it. The prediction is open again.
What Jiang said
In a lecture published on June 12, 2025, before the US-Iran war began, Jiang said: "This conflict is World War III. I cannot overstate how brutal this conflict will be. It will lead to fundamental changes in the world." That is a forecast about where the conflict would go, and it can only be settled by what the conflict becomes.
What went wrong
Our automated resolver ruled that the war had "conclusively ended via a June 2026 ceasefire" and therefore could not become a world war. But the ceasefire had already broken. On July 28, about five hours before the verdict, the US reported that Iran had resumed missile attacks on American forces, and on July 29 Bloomberg reported that President Trump was promising to strike Iran hard as the war escalated again.
The verdict rested on facts that were already out of date when it was written. The same day, the resolver described a different prediction's window as still open because of the escalating war, so the two findings contradicted each other.
What we changed
The resolver searched the news using the wording of the prediction itself, which surfaced op-eds about whether the war counts as World War III rather than reports of what was happening. It also let a single model's verdict go straight to the public score.
Since September 2026, every proposed verdict has to pass several checks before it can be written. A second, independent model must agree. The evidence must be dated after the lecture it concerns. An adversarial audit then tries to break the verdict. Final verdicts are also reviewed by a person before they are published.
Where the prediction stands
P021 is back to "not yet". Whether the US-Iran conflict becomes World War III is an open question, and the tracker will judge it on the evidence as it comes in. Removing this verdict also changes the headline accuracy, which now counts one fewer wrong prediction.
Follow the prediction on its [tracker page](https://jiangpredictions.com/p/P021.html), and see how verdicts are made on the [methodology and accuracy page](https://jiangpredictions.com/accuracy.html).
See the full prediction record at P021 on the tracker · All blog posts