Friedberg’s prediction concerned the probability that in calendar year 2024 a nuclear weapon would be used in an armed conflict somewhere in the world. To evaluate it as of the end of 2024, we check whether any such nuclear use occurred.

Historical and contemporary records show that the only wartime uses of nuclear weapons remain the U.S. bombings of Hiroshima and Nagasaki in August 1945; there has been no subsequent use of nuclear weapons in combat, including during 2024. Major nuclear-weapons monitoring organizations, global news outlets, and international bodies (e.g., UN, IAEA) extensively track and report on nuclear incidents; none report any nuclear detonation in warfare in 2024. This includes the major conflicts of that year (e.g., Russia–Ukraine, Israel–Hamas and related regional tensions), where nuclear use would have been unmistakable and globally reported.

Because no nuclear weapon was used in an armed conflict in 2024, the specific event whose chance he quantified (1–2% probability of nuclear use) did not occur. In forecast-scoring terms, assigning a small probability to an event that does not happen is not a factual error; however, in this task we are judging whether the event he predicted would possibly happen actually did happen in that year. Since it did not, the outcome for the proposition “a nuclear weapon will be used in conflict in 2024” is false, so the prediction that such an event might occur is evaluated here as wrong.

(Note: If we were strictly assessing probabilistic calibration, a 1–2% forecast for a non-occurring event would be considered reasonable, not ‘wrong.’ But under the binary correctness standard requested—did the described event happen during 2024 or not?—the result is that it did not.)