The surprising power of simple predictions
Why Read This
What Makes This Article Worth Your Time
Summary
What This Article Is About
Tim Harford recounts psychologist Robyn Dawes‘s story of a psychiatric patient locked away for a supposed delusion that turned out to be a genuine genetic condition—a parable for how experts overcomplicate simple problems. This leads into research by Ted Sarbin and Paul Meehl showing that simple linear regression models often outperform expert clinical judgment when predicting outcomes like college GPA or offender recidivism.
Harford then explores Dawes’s concept of “improper linear models”—formulas with arbitrarily chosen weights rather than mathematically optimized ones—applied to investment portfolios and even relationship happiness, where a crude formula comparing sex frequency to arguments predicted couples’ happiness nearly as well as far more sophisticated approaches like Harry Markowitz‘s portfolio theory.
Key Points
Main Takeaways
The Breast-Growth Misdiagnosis
A psychiatric patient was locked away for a supposed delusion that turned out to be a genuine genetic condition, illustrating how experts overlook simple explanations.
Simple Models Beat Experts
Research by Sarbin and Meehl found that basic linear regressions often predicted outcomes like GPA more accurately than trained clinical psychologists.
Improper Linear Models
Dawes proposed using arbitrarily weighted formulas instead of mathematically optimized ones, finding they performed surprisingly well across many domains.
Equal-Weight Portfolios Work
Even Nobel laureate Harry Markowitz split his own pension equally between stocks and bonds, and simple portfolios often rival optimized ones.
Sex Versus Fights Predicts Happiness
A crude two-variable formula comparing how often couples had sex to how often they argued accurately predicted relationship happiness.
Datasets Rarely Capture Everything
Complex optimized models can overstate their accuracy because real-world data is noisy, incomplete, and constantly changing over time.
Master Reading Comprehension
Practice with 365 curated articles and 2,400+ questions across 9 RC types.
Article Analysis
Breaking Down the Elements
Main Idea
Simplicity Often Outperforms Sophistication
Harford’s central argument is that crude, arbitrarily weighted statistical models—what Dawes called “improper linear models”—frequently match or beat complex expert judgment and optimized formulas, because real-world data is noisy and ever-changing, making elaborate precision less valuable than a robust, simple baseline grounded in the right variables.
Purpose
Championing Practical Statistical Humility
Harford writes to persuade readers—particularly those in finance, medicine, and policy—that chasing mathematically optimal models can be counterproductive, advocating instead for simple, transparent rules of thumb that are easier to apply, harder to overfit, and often nearly as accurate as their complicated counterparts.
Structure
Anecdotal → Evidential → Reflective
The piece opens with a memorable anecdote about a misdiagnosed patient, moves through escalating examples of “improper” models outperforming experts in academia, criminal justice, finance, and relationships, then closes by reflecting on why such crude models work and where their limits lie.
Tone
Witty, Curious & Persuasive
Harford’s tone is conversational and wry, using anecdotes—the misdiagnosed patient, the panellist’s marital challenge—to entertain while methodically building a persuasive case for statistical humility, balancing playful storytelling with genuine intellectual rigor drawn from decades of academic research across psychology, finance, and criminology.
Key Terms
Vocabulary from the Article
Click each card to reveal the definition
Build your vocabulary systematically
Each article in our course includes 8-12 vocabulary words with contextual usage.
Tough Words
Challenging Vocabulary
Tap each card to flip and see the definition
Decided on or chose something decisively, often after some indecision.
“He plumped for half in stocks, half in bonds.”
Introduced secretly or indirectly, here meaning hidden expertise embedded within a model’s design.
“there is already some expertise smuggled into the choice of variables”
Intense emotional conflicts or dramatic episodes within personal relationships.
“regardless of the psychodramas surrounding a couple’s relationship”
A member of a panel discussion, typically an expert invited to share views publicly.
“a fellow panellist challenged Dawes”
The informal act of adding up or tallying a series of numbers or instances.
“otherwise the totting-up may provoke a self-fulfilling crisis”
Strong, resilient, and able to withstand variation or stress without breaking down.
“A simpler, cruder method may be a bit more robust.”
Reading Comprehension
Test Your Understanding
5 questions covering different RC question types
1According to the article, the psychiatric patient’s breast growth turned out to be caused by a genetic condition rather than psychological trauma.
2What did Sarbin’s simple linear regression predict more accurately than clinical psychologists, according to the article?
3Which sentence best explains Dawes’s concept of an “improper” linear regression?
4Evaluate the following statements about the relationship-happiness study described in the article:
Of the unhappy couples studied, all 12 fought more often than they had sex.
All 30 happy couples had sex more often than they argued.
The relationship formula used three equally weighted variables: sex, fights, and income.
Select True or False for all three statements, then click “Check Answers”
5Based on the article’s discussion of DeMiguel, Garlappi, and Uppal’s 2009 paper, what can be inferred about complex optimized investment models?
FAQ
Frequently Asked Questions
He uses the anecdote as a memorable parable for his central theme: that experts, eager to find elaborate explanations, can overlook simple and direct answers that turn out to be correct, setting up his broader argument about the surprising power of simple statistical models.
It’s a predictive formula where the weights given to each input variable are chosen arbitrarily—equally, randomly, or by simple judgment—rather than mathematically calculated to best fit historical data, yet Dawes found such crude models often performed nearly as well as carefully optimized ones.
It serves as Harford’s most striking and memorable example, showing that even something as emotionally complex as marital happiness could be predicted with surprising accuracy using just two crudely weighted variables—sex frequency versus fight frequency—reinforcing his broader case for the power of simple models.
Readlite provides curated articles with comprehensive analysis including summaries, key points, vocabulary building, and practice questions across 9 different RC question types. Our Ultimate Reading Course offers 365 articles with 2,400+ questions to systematically improve your reading comprehension skills.
This article is rated Intermediate because it explains statistical concepts like linear regression and portfolio optimization in accessible, anecdote-driven language, requiring some familiarity with basic data and decision-making concepts but no advanced mathematical background to follow Harford’s argument, making it approachable for general readers interested in statistics and behavioral science.
Robyn Dawes was a psychologist whose research on “improper” linear models demonstrated that crude, arbitrarily weighted statistical formulas could rival or outperform expert judgment across diverse fields, from clinical diagnosis to criminal justice and personal relationships, making his work foundational to behavioral decision science.
The Ultimate Reading Course covers 9 RC question types: Multiple Choice, True/False, Multi-Statement T/F, Text Highlight, Fill in the Blanks, Matching, Sequencing, Error Spotting, and Short Answer. This comprehensive coverage prepares you for any reading comprehension format you might encounter.
