History's Cheat Code: How to Predict What Happens Next

In 2010, a researcher named Peter Turchin published a paper predicting that the United States would experience a spike in political instability around 2020. He didn't base this on polling data, political commentary, or gut feeling. He based it on historical cycles -- specifically, the same structural dynamics (elite overproduction, declining real wages, political polarization, fiscal crisis) that preceded instability peaks in prior centuries. Ten years later, in 2020, the United States experienced a pandemic, economic shutdown, mass protests, a contested election, and an attack on the Capitol.

Turchin didn't predict COVID. He didn't predict George Floyd. He didn't predict any specific event. He predicted the conditions under which specific events would become explosive rather than contained. That's the difference between prediction and prophecy, and it's the reason historical pattern recognition is the closest thing to a cheat code you'll find for understanding the present.

Why This Exists

Most people think predicting the future is either impossible (too complex, too many variables) or the domain of mystics and cranks. Both are wrong. What's true is that predicting specific events is nearly impossible. Nobody can tell you which stock will rise tomorrow or which country will have a revolution next year. But predicting the conditions that make certain types of events more likely -- that's not only possible, it's been done repeatedly, and it uses the same historical patterns we've been building through this series.

The distinction matters because it changes what you're trying to do. You're not trying to become a fortune teller. You're trying to become a better assessor of probabilities. Philip Tetlock, in Superforecasting, studied thousands of people who make predictions about world events and found that the best forecasters share specific habits: they think in probabilities rather than certainties, they update their estimates when new evidence arrives, they break big questions into smaller answerable ones, and they look for historical base rates. That last one is the history connection. When you ask "what's likely to happen next?" the most useful starting point is "what happened the last ten times a country looked like this?"

The Core Ideas (In Order of "Oh, That's Cool")

How Turchin actually made his prediction. Turchin's method isn't mystical. It's statistical. In Ages of Discord, he identified four structural indicators that correlate with periods of instability in American history: declining real wages for workers, rising elite consumption (inequality), increasing intra-elite competition (too many elites fighting for too few positions), and declining state fiscal health (debt, deficits, inability to fund public goods). He tracked these indicators from 1780 to the present and found that they form a recognizable wave pattern. The last major peak was the 1850s-1870s (Civil War era). The indicators suggested the next peak would arrive around 2020, give or take a decade. His later book End Times, published in 2023, updated the analysis and argued the instability window was open and hadn't closed.

The key insight isn't that Turchin got 2020 right. It's that the method works. He identified structural pressures, measured them, found historical parallels, and made a probabilistic forecast. You can argue about whether his specific indicators are the right ones, whether his cycle lengths are too rigid, whether he overweights some factors. Those are legitimate debates. But the general approach -- using historical patterns to assess current conditions -- produces better predictions than either "nobody can predict anything" or "I have a feeling about this."

The prediction toolkit. If you wanted to apply Turchin's approach yourself (and you can -- the data is public), here's the basic framework.

Step 1: Identify which cycle phase you're in. Review the secular cycle model -- are you in an integrative phase (social cohesion rising, inequality falling, institutions strengthening) or a disintegrative phase (the opposite)? The indicators are measurable: real wage growth vs. elite wealth growth, political polarization metrics, institutional trust surveys, fiscal balance.

Step 2: Check for elite overproduction. This is Turchin's signature indicator. Are there more people with advanced credentials than there are positions that match those credentials? Proxy measures include: law school enrollment vs. available legal positions, PhD production vs. tenure-track openings, the ratio of people with master's degrees to jobs that require them. When this ratio increases, the frustrated aspirant class grows, and political instability follows within a generation.

Step 3: Measure institutional trust. When people trust institutions (government, courts, media, universities), they resolve disputes through those institutions. When trust declines, they resolve disputes outside them -- through protest, political extremism, or violence. Institutional trust is measurable through surveys (Gallup, Pew, Edelman Trust Barometer) and has been declining in the United States and many Western nations for decades.

Step 4: Watch inequality velocity. As we covered in the inequality ratchet article, it's not the level of inequality that predicts instability -- it's the rate of change. When inequality accelerates while expectations are rising, you're in the J-curve danger zone. The relevant data: wage growth by income quintile, housing affordability over time, wealth concentration trends.

Step 5: Assess vulnerability to external shocks. A society in the integrative phase can absorb shocks (wars, pandemics, financial crises) because institutions are trusted and social cohesion is high. A society in the disintegrative phase can't -- the shock doesn't cause the problem, but it reveals and accelerates problems that were already there. COVID hit the United States during what Turchin identified as a disintegrative phase, which is why a public health crisis immediately became a political crisis. The same pandemic hit countries with different structural conditions and produced different political outcomes.

Practice exercise: mapping current events onto patterns. Try this with three current headlines. For each one, ask: which historical pattern does this map onto? Is this an inequality story, an elite overproduction story, an institutional sclerosis story, a technology disruption story, or a propaganda story? You don't need to be right. The point is the practice. You're building the habit of looking at current events through a structural lens rather than treating each one as a unique, unprecedented occurrence. Most aren't unique. Most have happened before in different clothing.

For example: rising housing costs and declining homeownership among young adults. Historical pattern: inequality ratchet (asset prices rising faster than wages, concentrating wealth among those who already own property). Precedent: pre-revolutionary France (land concentration), Gilded Age America (real estate trusts). This doesn't mean revolution is coming. It means the structural pressure that historically precedes instability is present in this specific sector.

Another example: increasing political polarization and the inability of Congress to pass legislation. Historical pattern: institutional sclerosis and elite competition. Precedent: the late Roman Republic (senators blocking each other's legislation to protect factional interests), the pre-Civil War U.S. Congress (unable to resolve the slavery question through normal legislative processes). Again, not a prediction of specific outcomes. An identification of the structural pattern.

Why prediction isn't prophecy. This is the caveat that makes the whole framework honest. Historical pattern recognition gives you probabilities, not certainties. Turchin didn't say "there will be an insurrection on January 6, 2021." He said "the structural conditions that historically produce instability will peak around 2020." The specific events are contingent -- they depend on individual decisions, random chance, and the interaction of countless variables. The structural conditions are more predictable because they emerge from large-scale dynamics that involve millions of people and change slowly.

Nate Silver made a similar point in The Signal and the Noise: the best predictions come from combining a base rate (what usually happens in this situation, historically) with specific evidence about the current case. The base rate without the specifics gives you a generic forecast. The specifics without the base rate gives you a story without context. You need both. Historical pattern recognition provides the base rate. Paying attention to current events provides the specifics.

The Bayesian connection. If you've encountered Bayesian reasoning -- the idea that you start with a prior probability and update it as new evidence arrives -- then you already understand the mathematical engine behind historical prediction. Your "prior" is the base rate from historical patterns. Each new piece of evidence (a policy change, an economic report, a political event) updates your estimate. You're never certain. You're always updating. The best historical thinkers are comfortable with this kind of thinking, and Tetlock's research confirms that the best forecasters are explicitly Bayesian in their approach.

How This Connects

The prediction toolkit connects directly to probability thinking, which you'd encounter in a math or statistics class. The concept of base rates, Bayesian updating, and distinguishing signal from noise -- these are mathematical concepts applied to historical data. If you can do this with history, you can do it with anything: evaluating college admissions odds, assessing career paths, or deciding whether a news story is significant or just noise.

The pattern recognition skill itself is the same skill you use on standardized tests, where recognizing the structure of a problem matters more than memorizing the content. A reading comprehension question on the SAT follows patterns. A passage structure follows patterns. The skill of recognizing the type of problem before you solve it -- that's what historical prediction trains you to do, just at a larger scale.

The School Version vs. The Real Version

The school version: History is about the past. It tells you what happened. The value is cultural literacy -- knowing the stories that shaped the world you live in. Prediction is not part of the curriculum because prediction seems unscientific, speculative, or outside the scope of what a history class should do.

The real version: History is a dataset, and datasets can be used to make forecasts. The forecasts aren't perfect -- they're probabilistic, not deterministic. But they're far better than the default approach most people use, which is either "nobody knows what'll happen" or "I have a strong opinion based on how I feel about the current situation." The cliodynamics approach -- measuring structural indicators, finding historical parallels, making probabilistic assessments -- is genuinely scientific. It's falsifiable. Turchin made predictions that could have been wrong. They weren't (at least the big ones). That doesn't make the framework infallible. It makes it useful.

History doesn't repeat, but it rhymes -- and the rhyme scheme is learnable. You've now spent six articles learning the patterns: empire cycles, inequality ratchets, geographic constraints, propaganda techniques, technology traps. This article is about putting them together into a toolkit that you can actually use. Not to predict the future with certainty, but to navigate it with better odds than most people have. That's not a small advantage. In a world where most people's model of the future is either anxiety or wishful thinking, having an evidence-based framework is a genuine edge.

This article is part of the History: Pattern Recognition series at SurviveHighSchool. [QA-FLAG: footer series line format — expected "Part of the History: Pattern Recognition series." with no "This article is" or "at SurviveHighSchool"] [QA-FLAG: footer related reading label — expected "Related Reading:" (capital R), got "Related reading:"]

Related reading: The Same 5 Things Keep Happening, The 200-Year Pattern, The Inequality Ratchet