2015 · Philip Tetlock & Dan Gardner · Crown · ~20 min read
Super forecasting.
Why can an ordinary person see the future more accurately than seasoned intelligence experts? The secret isn't IQ — it's how you think.
Read · Measure · Predict better
FOX MINDSET
10 PRINCIPLES
BRIER SCORE
Central thesis
Forecasting isn't an innate gift. It's a skill you can train to master.
Philip Tetlock spent twenty years dissecting expert predictions. The result was humiliating: deep professional opinion did no better than rolling a die. In a four-year tournament run by US intelligence, the top 2% of ordinary people beat seasoned experts by 30% accuracy. They weren't smarter — they had a different way of thinking.
Every January, experts parade on TV predicting stock and real estate markets. By year-end, almost no one bothers checking results. Tetlock proves that if you measure honestly, most experts forecast no better than a dart-throwing chimp. This book teaches you to predict in concrete numbers and update them constantly. It doesn't teach fortune-telling.
"
Beliefs are just hypotheses to keep testing — not treasures to stubbornly defend.
Philip Tetlock, Superforecasting
Framework 01 / Two thinking models
Are you a fox or a hedgehog?
Tetlock borrows the image from ancient Greek poetry to split two types. Hedgehog experts cling stubbornly to one big theory and filter all information through that lens. Foxes pull from many models and never commit to a fixed pattern. The truth: every superforecaster has a fox's brain.
Model 01 (weak)
The Hedgehog model
"I have 1 big idea that explains everything."
Always clings to one theory or one ideology.
Filters all information through their narrow lens.
Issues strong, definite, very ear-pleasing opinions.
Media's favorite face — always with sharp closing lines.
Predictions tend to be extreme — and very high error rate.
When reality slaps them, they blame circumstances instead of revising their theory.
Model 02 (strong)
The Fox model
"I pull from many models depending on the situation."
Never imprisons themselves in a single theory.
Knows how to combine strength from many opposing views.
Always predicts cautiously with clear confidence intervals.
Often ignored by media — doesn't know how to make sensational claims.
Predicts modestly — but accuracy is far higher.
When stumbling, immediately adjusts beliefs instead of finding excuses.
Media and society celebrate hedgehogs because of their fake confidence and bombastic statements. But if your goal is real accuracy, you must learn to think like a fox. Self-test: in the past 5 years, how many times have you bravely changed your mind in light of new facts? If zero — you're trapped in the hedgehog cage.
Framework 02 / 10 principles of the Superforecaster
Ten principles to become a superforecaster
The thinking habits of the top 2% aren't family secrets. They're skills you can start training today. Apply just half of these ten principles and you'll already beat most experts out there.
01
Triage
Focus on Goldilocks-zone questions — not too easy, not too hard. Easy questions everyone gets; questions too hard (5-year stocks) randomness wins.
02
Break down
Break a big question into smaller answerable ones. "Will there be war?" → "Will crisis X occur?" → "Does side Y have motive?"
03
Outside in
Before noticing situation specifics, ask: base rate for this kind of event? Most skip this step — jump straight into detail.
04
Update, don't entrench
When new info arrives, adjust the probability — don't jump too far or ignore. Probability should move in small steps, based on the strength of evidence.
05
Synthesize many angles
Multiple ways to analyze the same question from different angles. If 5 methods reach the same conclusion — high confidence. If they conflict — cautious conclusion.
06
Balance confidence
Overconfident: stop learning. Too timid: can't decide. Find the balance — willing to say "60%" instead of "maybe maybe not".
07
Balance judgment
Two errors: ignoring intuition (usually has value) or fully relying on it (easily biased). Use both — cross-check.
08
Look for errors to learn
After every wrong prediction, no excuses. Ask: what info could I have had? What process can I change? This is the #1 difference.
09
Teams are better
Superforecasters work 23% better in teams. Condition: the group dares to disagree, respects multiple perspectives — not a yes-man group.
10
Perpetual beta
No "already good". Always in beta — ready to change models, learn new tools, fix bad habits. Growth mindset is the #1 driver.
Framework 03 / Brier Score — measure predictions
The Brier Score: how to grade your judgment
If you don't measure, you'll never improve. The Brier Score is how you weigh the accuracy of your predictions. The smaller the score, the closer to perfect. Zero is absolute; 0.5 means you're just rolling a die for luck.
Brier Score formula (simple)
Brier Score = (Predicted probability − Actual outcome)²
Example: You predict 80% rain — it rains: (0.8 − 1)² = 0.04 (good) You predict 80% rain — no rain: (0.8 − 0)² = 0.64 (bad) You predict 50% — nothing to learn: (0.5 − x)² = 0.25 (average)
This score punishes the overconfident-but-wrong heavily. At the same time, it rewards those both accurate and decisive. After tracking ~30 predictions, you'll know exactly where you sit on the thinking map.
12 chapters — 4 big themes
Detailed map of the book
This book doesn't get lost in airy theory — it leans on 20 years of real combat data. The 12 chapters strip away why expert long-term forecasts are so terrible, and more importantly, show you how to flip the table.
PART I — THE PROBLEM
The paradox of professionalism
Tracking down why deep expertise doesn't pair with good forecasting.
CHAPTER 01
The optimistic skeptic
Tetlock asserts that good forecasting is a skill you can acquire. Data shows the average expert predicts as poorly as die-rolling — but some individuals climb to peak through the right thinking method.
Optimistic skeptic
CHAPTER 02
The illusion of understanding
Most experts are stuck in narrative bias. They weave beautiful stories about the past and falsely believe they understand the rules. This cognitive trap drives them to wrong confidence about the future.
HindsightNarrative
PART II — MEASUREMENT
The power of measurement
Without an honest scoring system, we'll never know who's actually skilled.
CHAPTER 03
Scoring forecasts
The Brier Score lets us compare individual accuracy in the most objective way possible. For the first time in history, we can quantify the ability to see the future.
Brier score
CHAPTER 04
The superforecasters
Out of 20,000 participants, the top 2% beat the entire US intelligence apparatus. Their secret isn't sky-high IQ — it's a clear thinking roadmap.
Top 2%
PART III — TRAITS
Portrait of the forecasting superhumans
Smashing 3 popular myths and outlining the real traits of the superforecaster group.
CHAPTER 05
Is intelligence everything?
The average IQ of the superforecaster group is 130. High but not genius. In fact, beyond this threshold IQ adds only an extra 5% to success.
IQ myth
CHAPTER 06
Superhuman calculation ability?
They don't need advanced statistics — they just need to be fluent in basic addition and subtraction. The most important skill is breaking giant problems into answerable questions.
Fermi
CHAPTER 07
A news addiction?
They read widely — not heavily. The difference: they kneed the information themselves and never let themselves get nailed to a single stream of opinion.
Fox vs Hedgehog
CHAPTER 08
Permanent beta state
The most precious trait is growth thinking. After every stumble, they don't seek excuses — they dive into investigating the holes in their own process.
growth mindset
PART IV — APPLY
Expanding the view and its limits
How to build a team, the role of leaders, and the boundaries this method can't cross.
CHAPTER 09
The power of super-teams
A superforecaster team can work 23% better than an individual. The prerequisite: the team must value disagreement and aggressively reject groupthink.
Superteams
CHAPTER 10
The leader's paradox
Leaders need decisiveness — but forecasters need skepticism. The only solution: be aggressive about the goal but extremely open about the method.
Leader paradox
CHAPTER 11
Are they really superhumans?
The method is extremely effective short-term — but for visions over 5 years, everything goes blurry. You must know when to trust the numbers and when to humbly admit life's uncertainty.
Limits
CHAPTER 12
The future is waiting
The forecasting revolution is quietly underway. Don't treat the future as fated destiny — treat it as a probability problem you can absolutely calculate.
Future
Tetlock hides a brutal truth: becoming a superforecaster eats a massive amount of your time. You must stay up watching news, scrutinize every number, dissect data daily. An office worker with 8-hour days and dozens of deadlines has no chance to play this game. It's the territory of venture investors — not those who must decide in a flash.
Flashcards / 10 core concepts
Ten core flashcards
Ten thinking codes distilled from thousands of forecasting experiments. Flip the cards and train your brain to see the world through probability numbers.
Card 1 / 10
Memorized: 0
Question
Click to see answer
Answer
Click to see question
Quiz / 6 comprehension questions
Did you actually get Tetlock?
6 questions — Tetlock has many overlapping ideas easy to confuse. This quiz separates: foxes vs hedgehogs, system 1 vs 2, Bayes update. Miss 2+ → ideas are blurred in your head.
Question 1 / 6Score: 0
15 Actions / Apply to your decisions
Fifteen concrete actions
Tetlock doesn't teach you fortune-telling — he teaches you to build a disciplined judgment system. Pick three actions and practice them like a daily ritual. Accuracy doesn't appear by itself — it's forged from the smallest habits.
Private journal / Write to understand
Write to sharpen prediction
Writing is the forecaster's sharpest weapon. The five questions below force you to face your overconfidence honestly. Don't write to perform — write to see the holes in your mind clearly.
1
2
3
4
5
Discussion / Community
Your take
I value every share and sharp counter-argument from you. Your email is only used for verification — never displayed publicly. After you confirm via email, your view appears immediately.