What Is Predictive History? A Framework for Geopolitical Forecasting

Predictive history is a methodology for forecasting geopolitical events by treating historical patterns as structural rather than偶然. It operates on a simple premise: societies follow recognizable cycles of rise and decline, and understanding these cycles makes the future less opaque. Professor Jiang Xueqin, a historian who applies this framework on his Predictive History YouTube channel, has spent nearly four years using it to call wars, elections, and succession crises — many of them months or years in advance. The tracker at this site has documented 328 of these predictions across 156 lectures. Of the 26 that have resolved so far, 23 have been confirmed or partially confirmed — an 88% hit rate on resolved predictions.

The framework rejects personality-driven political analysis in favor of structural constraints. Leaders matter less than the forces they channel. When Jiang forecast in April 2024 that the United States would go to war with Iran within 2 to 4 years (P011), he wasn't guessing about presidential temperament. He was tracing a familiar pattern: declining hegemons seek peripheral victories to arrest their loss of influence, and regional powers like Saudi Arabia cannot survive without their patron fighting their battles (P025). The structural incentives for war were already baked in; the trigger was a matter of time. The prediction was confirmed in March 2026 when airstrikes began.

Three core principles animate predictive history. First, history rhymes rather than repeats — same incentives, different costumes. The American withdrawal from Afghanistan in 2021 didn't just signal incompetence; it telegraphed the credible end of unipolarity, a precondition for the Iran war prediction. Second, civilizational decline follows a recognizable sequence: fiscal strain, military overextension, domestic polarization, and the search for restorative conflict. Third, personality is subordinate to structure — autocrats and presidents alike are constrained by the circumstances they inherit. You can see this in Jiang's December 2025 call that America would attack Venezuela (P082), a move consistent with hegemonic decline and the search for accessible wins.

The evidence so far is not definitive, but it is suggestive. Jiang called Trump's 2024 victory in April 2024 (P001), correctly identifying the structural conditions for a comeback. He predicted JD Vance as a strong VP pick the following month (P002b); Vance joined the ticket in August. He foresaw Mojtaba Khamenei's ascent to Supreme Leader within five years (P041), a dark horse bet that looks increasingly probable as the current Supreme Leader ages. Perhaps most striking was his early 2026 prediction that China would stay out of the Iran conflict (P071), a contrarian call that treated Beijing's restraint as predictable given its economic entanglement with the West and its aversion to direct confrontation with American firepower.

But the framework also admits uncertainty. Three predictions have been marked wrong, and 302 remain pending — including some, like P021's claim that the Iran war constitutes World War III, that are still developing. The tracker deliberately treats partial confirmations as hits: Iran has not yet used nuclear weapons (P016), but the war is young, and civilian infrastructure attacks are escalating (P020). Dubai's long-term decline as a financial center (P028) will take decades to judge. Predictive history is not about precision timing; it's about direction. Structural forces point the way, but the path is messy.

What distinguishes Jiang's lectures from typical geopolitical punditry is the systematic nature of the analysis. Each prediction is timestamped, categorized, and tied to a specific video source. When the tracker marked P088 — the decapitation strike that killed Ayatollah Khamenei — as confirmed in March 2026, it did so because the event occurred, not because it confirmed a bias. The methodology demands that claims be falsifiable. That is why this site exists: to document whether the framework actually works, in public, with the numbers updating daily.

Predictive history remains an experiment. Twenty-six resolved predictions out of 328 is a small denominator, and the 88% accuracy rate could regress as more predictions reach their timeline dates. But the early returns are compelling enough to take the framework seriously — and to watch what happens next. As more predictions from Jiang Xueqin's lectures resolve over the coming months and years, the tracker will continue to score them. History may not repeat exactly, but if predictive history is right, it rhymes clearly enough that we can see the next measure coming.

Browse all tracked predictions · Accuracy analysis · All blog posts