Predictive History: Jiang Xueqin's Method for Forecasting Geopolitics
Predictive history is a forecasting method that treats the recorded past as data. The premise: if empires decline in recognizable stages, wars follow recurring patterns, and geopolitical structure constrains leaders more than personality shapes events, then the future can be reasoned about in advance — not with certainty, but with dated, checkable predictions. The term is now most visible as the practice of Professor Jiang Xueqin, the educator and commentator whose YouTube channel, Predictive History, has turned the method into hundreds of public, timestamped calls about wars, elections, and industrial decline since 2024. This tracker exists to check those calls against the news, one at a time.
Three claims underneath the method
Strip away the production and the framework rests on three claims.
First, history rhymes. Conflicts replay recognizable patterns — the opening strikes, the escalation ladders, the way empires overextend — so the past supplies the priors. Jiang's June 2024 capstone lecture, Geo-Strategy END: Psychohistory, takes its name from Isaac Asimov's fictional science of forecasting mass behavior from historical law. The name is the mission statement.
Second, civilizational decline follows a pattern. His Civilization series walks the late-imperial checklist — fracturing domestic coalitions, inflation, military overstretch — and closes with Civilization END: The Decline and Fall of the American Empire (June 2025), a title that leans deliberately on Gibbon. Decline, in this telling, is a process you can place on a timeline, not a mood.
Third, structure outweighs personality. Leaders come and go; lobbies, geography, alliances, and debt endure. A prediction grounded in structure can survive being wrong about which individual occupies the office.
What the method produces in practice
The lectures are not commentary; they are commitments. In April 2024, with President Biden still the presumptive nominee, Jiang predicted across his Geo-Strategy series that Trump would win the 2024 election (P001) — his stated reasoning was structural: a fracturing Democratic coalition, inflation, immigration, and foreign-policy failures. The call was confirmed when Trump won that November. The same May 2024 lecture, Geo-Strategy #5: Why Trump Will Win, shows both the granularity and the limits: he was wrong that Trump would pick Nikki Haley as his running mate (P002a) and right that JD Vance was the strong alternative (P002b). One lecture, one split verdict.
The Iran war is the method's stress test. As early as April 2024 he predicted a US war with Iran within two to four years (P011), driven by three structural forces he named repeatedly: the Israel lobby, Saudi Arabia's rivalry with Tehran, and what he describes as America's addiction to empire. The follow-on calls stacked up over two years — the US would bomb Iranian nuclear facilities (P012); shock and awe would fail against a dug-in adversary (P015, argued in Geo-Strategy #6: America's Imperial Hubris); nuclear weapons would stay holstered (P016); Iran would close the Strait of Hormuz (P026) and strike the Gulf states (P027); the war would run many weeks, possibly years (P018). When the war began in early March 2026 — inside his window — those calls started resolving against real events, and most have held. One did not: his May 2024 prediction that Putin would declare a nuclear umbrella over Iran (P017, from Geo-Strategy #8: The Iran Trap) was proved wrong.
The method also runs outside war. From Secret History END: Pax Judaica (December 2025) he predicted an American attack on Venezuela in the short term (P082); it was confirmed when the attacks followed within months. From Game Theory #24: The AI Apocalypse (May 2026) he predicted AI companies would ship AI girlfriends to drive engagement (P129); the launches arrived on schedule.
The scoreboard, honestly
As of 2026-08-26, this tracker holds 382 predictions drawn from 178 source lectures (177 transcribed so far). Only 22 have resolved: 16 confirmed, 6 wrong — a headline accuracy of 73%. Two caveats belong next to that number. The denominator is small, and the resolved set is dominated by a single event, the Iran war, so these are not independent trials of the method. And 343 predictions remain pending; most of the record is still ungraded. Scoring is binary — a settled call is either confirmed or wrong — which keeps the arithmetic honest but allows no half-credit.
Where to verify it yourself
Every claim above is traceable. The full archive of Jiang Xueqin lectures, with dates and the predictions extracted from each video, is browsable by lecture, and who Jiang Xueqin is — and how he argues — is documented separately. Most readers first meet the term through predictive history YouTube lectures; his channel is the format's most prolific practitioner. The point of an independent tracker is that you should not have to take the channel's word for anything: watch the lecture, note the timestamp, and wait.
Browse all tracked predictions · Accuracy analysis · All blog posts