> ## Content Index
> Fetch the complete content index at: https://longtalkshort.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# When the Forecast Misses
- URL: https://longtalkshort.com/when-the-forecast-misses-jiang-xueqin-revises-predictive-model/
- Published: 2026-11-06T15:14:56.000Z
- Updated: 2026-11-06T15:14:56.000Z
- Description: You follow a forecaster, then watch them miss. Jiang Xueqin admits he was wrong on Brazil, Alberta and the midterms, and explains how his predictive model is supposed to revise.
- Author: Truman
- Tags: Psychology, Jiang Xueqin, Forecasting, Midterms

**If you follow online forecasters, you eventually meet the harder part of the job: admitting a miss.** In Geo-Strategy Update #10 on his Predictive History channel, Jiang Xueqin does that for the US midterms, and for Brazil and Alberta along the way. The useful material is not another round of alarmist geopolitics. It is his account of how a predictive model is supposed to change when the world refuses to cooperate.

*In short: Jiang says he got three calls wrong, walks through why his "globalists versus nationalists" frame led him there, and argues that predictive history means revising the analysis rather than defending an old forecast. He then pivots to what a Democratic Congress will do. Treat the midterm outcome as reported fact; treat the motive stories as his speculation.*

## 1\. What he says he got wrong

Speaking from Beijing, Jiang opens with a list. He was wrong about Brazil's election, wrong about Alberta's independence referendum, and wrong about the midterms. The point of the talk, he says, is not to excuse those misses. In his method, a worldview is tested by predictions. When predictions fail, the analysis has to move.

His baseline model is familiar to his audience: a conflict between "globalists," whom he associates with figures such as Mark Carney, and "nationalists," whom he associates with Donald Trump and Vladimir Putin. He says he treated local elections as a scoreboard for that fight. That is the frame that then steered him through Brazil, Alberta and the US vote.

## 2\. How the revision path runs

On Brazil, he says he assumed Luiz Inácio Lula da Silva had a lock on the presidency. The first round shocked him when Flávio Bolsonaro finished ahead, and he expected a wide runoff win for the Bolsonaro camp. After looking for explanations, including evangelical turnout and social-media targeting, he concluded that nationalists were ascending. That, he says, led him to expect a strong Alberta independence showing, "maybe" around 40%, and to doubt the consensus that Democrats would take Congress.

Alberta did not do what he expected. Then Lula won the Brazilian runoff, which Jiang reads as evidence that "globalists" still have tools. In this talk he says that after that he shifted and predicted a Democratic sweep. The midterms, he says, became a "blue tsunami": Democrats taking the House and Senate, most governorships, and flipping Texas into what he calls a "purple" state.

![Four cards. 1. Worldview: globalists versus nationalists drive elections. 2. Brazil shock: first round surprised him; nationalists seemed to rise. 3. Alberta miss: expected about 40% for separation; official share about 25%. 4. Midterms: says he then flipped toward a Democratic sweep. Banner: a revision story told after the results were known.](https://storage.ghost.io/c/d7/93/d793216c-cf13-441f-bdae-4f3aef998c0e/content/images/2026/11/fig1-32.png)

Jiang's revision path as he tells it. The path labels are ours. Made for Long Talk Short.

## 3\. The method claim

The cleanest idea in the video is epistemic, not geopolitical. "You don't have a prediction \[and\] stick to this prediction," he says. You have a worldview, derive forecasts from it, and change the forecast when fresh information changes the analysis. That is a useful standard for any public forecaster. It is also easier to recite after election night than to practice before it.

He is more candid than many YouTube predictors about prior speculation. He recalls wondering whether the White House had a "secret plan" to rig or delay the midterms, and whether social-media algorithms could deliver a Republican surprise. Those are presented as discarded wrong turns, not live claims.

## Weighing it: what holds up

**Framing:** A revision story told after results were known, used to illustrate how his model is supposed to update.

**Holds up:** Major news organizations, including the Associated Press and NPR, report that Democrats won control of both the House and the Senate. Chamber control is what those calls establish; this piece does not add seat totals. "Blue tsunami" and "Texas is purple" are Jiang's labels. Statewide Democratic wins in Texas were widely reported; the color metaphor is his.

**Partly holds:** Elections Alberta's official results for Question 10 show about 75% choosing to remain in Canada and about 25% choosing to start the legal process toward a binding separation referendum. Jiang's "about 20%" understates that share, and the ballot was not a direct independence vote. On Brazil, contemporaneous reporting matches the outline he gives: Flávio Bolsonaro led the first round and Lula won the runoff.

![Six rows checking Jiang's claims: Democrats take House and Senate, major news calls yes with some races open; Alberta independence about 20 percent, official start-process share about 25 percent; prior call that GOP keeps Senate, meetup captions; his Democratic-sweep claim, no dated recording found here; behind-scenes war cooperation, attributed speculation; Texas as purple, statewide Democratic wins, the label is his.](https://storage.ghost.io/c/d7/93/d793216c-cf13-441f-bdae-4f3aef998c0e/content/images/2026/11/fig2-32.png)

Assertions versus verification status. Made for Long Talk Short.

**Corrections:** A meetup recording from Predictive History's Global Meet-Up #2 in Los Angeles has him saying, in a midterms Q&A, "It's possible Democrats take the House, but they won't take the Senate" (auto captions). We did not find a dated public recording of the Democratic-sweep call he cites in this talk; that remains his assertion here.

**Not checked:** Forward-looking motive talk in the second half—donors, Iran, AI bailouts, the NDAA, stablecoin law, Epstein-file leverage, a possible Trump–Putin–Xi summit—stays attributed speculation. Our transcript is YouTube's automatic English captions. The figure groupings are ours. Distinct from our Jiang piece on Iran and end-times belief.

**Disclosures:** None commercial in the clip. Predictive History is his channel.

**Bottom line:** the update-the-model ethic is the useful part; treat chamber-control news calls as reporting and his motive stories as speculation.

**Related reading:** [Jiang on Iran and end-times belief](https://longtalkshort.com/interests-or-prophecy-jiang-xueqin-iran-war-end-times-thesis-checked/) · [Jiang on the occult of the meritocracy](https://longtalkshort.com/jiang-xueqin-occult-of-the-meritocracy-test-built-on-a-belief/) · [Annie Duke on judging the bet, not the result](https://longtalkshort.com/annie-duke-thinking-in-bets-judge-the-bet-not-the-result/)

## Long Talk Short

- Jiang admits misses on Brazil, Alberta and the US midterms, and uses them to explain how his model is supposed to update.
- His method claim is sound on paper: revise the analysis when evidence changes, rather than defend an old forecast.
- Democrats' control of Congress is supported by major news calls; this piece does not add seat totals.
- Alberta's official separation-process share was about 25%, not the \~20% he cited, and the question was not a direct independence vote.
- A meetup recording has him saying Democrats "won't take the Senate"; his Democratic-sweep claim in this talk is unverified here.

**The one thing to remember:** a forecaster's value is less in never being wrong than in showing how the model changed when it was.

*This summary is for general information only and is not professional advice.*

Source: Predictive History, "Geo-Strategy Update #10: I Was Wrong. The Democrats Take Congress," on YouTube. This is an independent summary and is not affiliated with Predictive History or Jiang Xueqin.