Welton Chang had actual combat experience in Iraq. If anyone had earned the right to trust his gut on a war forecast, it was him. In 2013, as part of a forecasting tournament called the Good Judgment Project, he was asked to estimate something with real stakes attached: would the Free Syrian Army capture Aleppo?
He didn’t start with Syria.
He started somewhere else entirely — with a question that had nothing to do with the specific war in front of him. Historically, across unrelated conflicts, how long does it typically take even a militarily superior attacking force to seize a city the size of Aleppo? The answer: it rarely happens fast, if it happens at all. Something like a 10-20% base success rate within a short timeframe. Only after establishing that cold, borrowed number did Chang turn to the actual details of the Syrian civil war — and there, he found something that mattered: the Free Syrian Army didn’t come close to qualifying as a militarily superior force. So he adjusted his estimate down from there. His final forecast ranked in the top 5 percent of forecasters’ accuracy on that question.
To most of the analysts working this same problem, that approach would have looked backward, maybe even irresponsible. Their war wasn’t a spreadsheet. It was hour-by-hour — troop movements, morale, alliances shifting overnight. To someone who’d spent years mastering the specific, granular reality of Middle Eastern conflict, reaching for a database of unrelated battles from unrelated decades feels almost insulting to the seriousness of what’s actually happening. The battlefield was right there. Not in a dataset.
That’s not a foolish instinct. It’s a completely understandable one — when the thing in front of you is loud, urgent, and alive, backing away from it to consult cold historical averages can feel like looking away at the exact moment you can least afford to.
Chang did it anyway. Not because the details of Syria didn’t matter — they mattered plenty, and he used them, just seconds. He used the outside view as an anchor first, then let the inside view adjust it. Not one or the other. Sequence.
There’s a cleaner, more clichéd version of this story that gets told sometimes: trust the data, not your gut. That’s not quite it, and it actually undersells what Chang did. The data wasn’t the answer. It was the discipline of not letting the loudest, most immediate details in the room make the decision before anything else got a say.
Underneath the war and the numbers, this is a story about discipline — the kind it takes to step back from something urgent, real, and in motion, and look outside it anyway, before deciding what it means.