When Nobel Prize-winning physicist Richard Feynman sat down to lunch, he didn’t just pick a meal based on hunger or whim. According to newly analyzed notes from his personal archives, Feynman turned the seemingly simple act of ordering food into a probabilistic optimization problem. His handwritten calculations, now studied by behavioral economists, suggest that humans—even without realizing it—approximate his method when faced with choices like restaurant menus, job offers, or even medical treatments.
Researchers who decrypted Feynman’s notes, housed at the California Institute of Technology Archives, found that his approach relied on expected utility theory—a framework that weighs outcomes by their likelihood and desirability. What’s striking is that modern studies confirm people naturally gravitate toward similar strategies when making decisions under uncertainty. The implications stretch far beyond dining: from healthcare policy to financial planning, Feynman’s lunch dilemma offers a blueprint for how humans balance risk and reward in everyday life.
But here’s the twist: Feynman’s method wasn’t about perfection. It was about satisficing—a term coined by economist Herbert Simon to describe decisions that are “good enough” rather than optimal. This aligns with recent findings in cognitive psychology, which show that humans often simplify complex choices using mental shortcuts called heuristics. Feynman’s notes, it turns out, were a rare glimpse into how a genius mind applied rigorous math to a mundane task—and how the rest of us do the same, albeit less formally.
How Feynman’s Lunch Math Works—and Why It Matters
Feynman’s approach hinged on three key steps, each rooted in probability theory:
- Assign probabilities: He estimated the likelihood of each menu item being “good enough” based on past experience or reviews. For example, if he’d ordered a dish twice and enjoyed it both times, he’d assign it a high probability of satisfaction.
- Assign utilities: He then rated each option’s desirability on a scale (e.g., 1–10), factoring in taste, nutritional value, or even the effort required to eat it.
- Calculate expected value: Multiplying probability by utility for each option, he’d pick the one with the highest expected score—a strategy now called Feynman’s Lunch Algorithm in informal circles.
This method mirrors what economists call expected utility maximization, but with a critical difference: Feynman’s notes reveal he adjusted for uncertainty. If he was unsure about a dish’s quality, he’d skew his probability downward, making the expected value more conservative. This aligns with modern behavioral economics research showing that humans overestimate rare events (like food poisoning) and underestimate common ones (like enjoying a familiar dish).
Why This Isn’t Just About Food
The broader significance of Feynman’s lunch math lies in its universality. Behavioral scientists have since applied similar frameworks to:
- Healthcare decisions: Patients weighing treatment risks vs. Benefits (e.g., vaccine hesitancy studies use expected utility models to predict choices).
- Financial planning: Investors balancing risk and return (Feynman himself was known to use probabilistic reasoning in his scientific work).
- Public policy: Governments modeling citizen compliance with health guidelines (e.g., COVID-19 lockdown studies applied expected utility to predict behavior).
Even Feynman’s adjustment for uncertainty has real-world parallels. In medicine, this translates to clinical decision-making frameworks like the number needed to treat (NNT), where doctors weigh a treatment’s benefits against its side effects using probability-based thresholds.
Do Humans Really Use Feynman’s Method?
New research suggests the answer is yes—but unconsciously. A 2023 study published in Nature Human Behaviour found that when given multiple-choice tasks (from restaurant menus to job applications), participants’ selections closely matched Feynman’s expected utility calculations, even when they couldn’t articulate their reasoning. The study’s lead author, Dr. Elena Rusconi of the University College London, noted that:
“Feynman’s notes reveal a cognitive shortcut that’s universal. We’re not solving complex equations in our heads, but our brains are approximating the same logic—just with faster, fuzzier heuristics.”
The study also highlighted a gender disparity in how people apply the method: Men tended to overestimate high-probability outcomes (e.g., “This dish is definitely good”), while women adjusted probabilities more conservatively—a pattern observed in previous risk-perception research. This could explain why women are often more cautious in medical decisions, such as vaccine uptake.
The “Good Enough” Factor: Satisficing in the Real World
Feynman’s method isn’t about finding the best option—it’s about finding the good enough one. This aligns with the theory of bounded rationality, developed by Nobel laureate Herbert Simon. In practice, this means:
- Menu choices: You might skip the “perfect” dish if it’s too spicy or unfamiliar, even if it has slightly higher expected utility.
- Medical tests: Patients may decline a 90% accurate screening if the side effects outweigh the benefit—a classic trade-off in expected utility.
- Investments: Feynman himself avoided high-risk stocks, preferring diversified, low-volatility portfolios that aligned with his satisficing approach.
This principle even extends to public health campaigns. When designing interventions—like flu shot reminders—policy makers use expected utility models to predict which messages will maximize compliance. Feynman’s lunch math, in other words, is the hidden algorithm behind everything from your grocery list to global immunization strategies.
How to Apply Feynman’s Method to Your Own Decisions
You don’t need a PhD in physics to use this framework. Here’s a step-by-step guide to applying Feynman’s approach to your next sizeable choice:
- List your options: Write down all viable choices (e.g., restaurant dishes, job offers, medical treatments).
- Assign probabilities:
- Rate each option’s likelihood of meeting your needs (e.g., “80% chance this salad is healthy”).
- Adjust downward if you’re unsure (Feynman’s “uncertainty discount”).
- Assign utilities:
- Score each option on a 1–10 scale based on desirability (taste, cost, convenience, etc.).
- Weight factors differently if some matter more (e.g., health > cost for a meal).
- Calculate expected value: Multiply probability × utility for each option. Pick the highest score—or the one that feels “good enough.”
- Check for satisficing: If the top choice is too risky or effortful, consider the next-best option that meets your minimum standards.
Example: Choosing a Restaurant
| Option | Probability (Good Enough) | Utility (Desirability) | Expected Value |
|---|---|---|---|
| Pasta (familiar) | 90% | 8 | 7.2 |
| Sushi (new but risky) | 60% | 9 | 5.4 |
| Salad (healthy but boring) | 85% | 6 | 5.1 |
The pasta wins—but if you’re feeling adventurous and adjust the sushi’s probability upward (e.g., because you’ve heard good reviews), it might become the satisficing choice.
What’s Next for Feynman’s Lunch Math?
Researchers are now exploring how Feynman’s method could improve:
- AI decision-making: Algorithms that mimic human satisficing (rather than pure optimization) may be more trustworthy for users.
- Healthcare algorithms: Tools that help patients weigh treatment options with Feynman-style uncertainty adjustments.
- Climate policy: Modeling public support for green initiatives by predicting how people balance personal costs vs. Societal benefits.
The next checkpoint in this research will be a 2025 conference at Caltech (tentatively titled *”Feynman’s Legacy in Decision Science”*), where behavioral economists and computer scientists will present new models blending expected utility with machine learning. Until then, Feynman’s lunch notes remain a testament to the power of simple math in complex choices.
Key Takeaways
- Feynman’s lunch math uses expected utility theory to balance probability and desirability in decisions.
- Humans intuitively approximate this method, even without formal training—a process called satisficing.
- The approach applies to healthcare, finance, and public policy, where uncertainty plays a key role.
- You can use the 4-step framework (probability × utility = expected value) for any major choice.
- Future research may integrate Feynman’s logic into AI and climate modeling.
Have you ever made a decision using a similar “math” in your head? Or do you rely on gut instinct? Share your stories in the comments—and don’t forget to follow World Today Journal for more on how science explains everyday life.