Table of Contents
John Maynard Keynes never actually organized a beauty contest, which is perhaps the first irony worth savoring. The man who would reshape modern economics drew his famous metaphor from a newspaper parlor game popular in 1930s Britain, where readers submitted their picks for the prettiest faces from a grid of photographs. The twist? Winners weren’t those who chose the faces they personally found most attractive, but those who correctly predicted which faces everyone else would choose. Prizes went to the prophets, not the poets.
This seemingly frivolous exercise contains what might be the most uncomfortable truth in all of game theory: in many competitive situations, getting the right answer matters far less than predicting what everyone else thinks is the right answer. And in the highest levels of this game, even that isn’t enough.
The Infinite Mirror
Imagine standing between two perfectly parallel mirrors. Your reflection bounces back and forth into infinity, each image reflecting all the images before it. This is the structure of reasoning in a Keynesian beauty contest, except instead of mirrors, there are minds—each attempting to reflect what all the other minds are thinking.
The newspaper game worked like this: a rational player wouldn’t simply choose the faces they found attractive. That would be level zero thinking, the intellectual equivalent of showing up to a chess tournament and moving pieces because they make pretty patterns.
Instead, a thoughtful contestant would choose faces they believed had broad appeal—level one thinking. But wait. If everyone else is also playing at level one, choosing faces they think others will like, then the winning strategy becomes choosing faces you think others will think others will like. That’s level two. And so on, spiraling into an abyss of recursive reasoning where each layer of thought attempts to encompass all previous layers.
This is where most explanations of the beauty contest end, but here’s what those explanations miss: the real lesson isn’t that you should climb to some higher level of reasoning. The real lesson is that the picture itself has become almost completely irrelevant.
When Fundamentals Become Fictions
Consider the stock market, which Keynes compared to his beauty contest. A company’s shares should theoretically reflect its fundamental value—its assets, earnings potential, management quality, and competitive position. These are the “picture” in our metaphor. An investor who studies these fundamentals is looking directly at what’s in the frame.
But in a sufficiently liquid market with many participants, this approach becomes naïve. What matters isn’t whether a company is genuinely valuable, but whether other investors will believe it’s valuable tomorrow, next week, next quarter. This creates a bizarre situation where shares can surge in price even as the underlying business deteriorates, simply because investors believe other investors believe other investors will buy. The picture has become secondary to predictions about predictions about predictions.
The dot-com bubble provides a perfect case study. By 1999, many investors knew that companies with no revenue, no profit, and often no viable business model were wildly overvalued by any fundamental measure. They were looking at the picture and seeing nothing of substance. Yet these same rational investors continued buying because they correctly predicted that other investors would continue buying—at least for a while longer. The game wasn’t about beauty anymore; it was about predicting when everyone else would simultaneously realize they were playing a game about predicting when everyone else would realize they were playing a game about… and so the tower of cards eventually collapsed, as such towers must.
The Paradox of Sophistication
Here’s where game theory reveals something counterintuitive: in a beauty contest, being too smart can be exactly as fatal as being too dumb.
The least sophisticated player looks at the faces and picks their favorite. They lose. The moderately sophisticated player thinks one level deep: “I’ll choose what I think others will find attractive.” In a room of unsophisticated players, this strategy dominates. But in a room of other moderately sophisticated players, everyone makes similar choices, and winners are determined by random variation.
The very sophisticated player thinks two, three, or four levels deep. And here’s the trap: if you’re the only person in the room thinking at level four while everyone else is at level one, you lose just as badly as the person thinking at level zero. You’ve outsmarted yourself right out of the prize.
The Nash equilibrium of this game—the point where no player can improve their outcome by unilaterally changing strategy—occurs when everyone converges on the same level of reasoning. But reaching this equilibrium requires accurate knowledge of the average sophistication level of all other players. Think you’re in a room of level-two thinkers? Better play level three. But if you’re wrong and they’re actually at level one, you’ve just thrown away your advantage.
This creates a peculiar situation where the theoretically optimal strategy is unknowable without perfect information about other players’ reasoning abilities. And since no one has this information, everyone is essentially guessing—which brings us back to ignoring the picture entirely and focusing instead on reading the room.
The Consensus Trap
Organizations face their own version of the beauty contest, often with devastating results. Picture a corporate strategy meeting where executives must decide which projects to fund. The “beautiful faces” are various business proposals, each with projected returns, risk assessments, and strategic rationales. A diligent executive might carefully evaluate each proposal’s merits. But a politically savvy executive knows that what matters isn’t which projects are objectively best, but which projects the CEO likes, or which projects the board expects the CEO to like, or which projects the board expects the CEO to expect them to expect him to like.
The memo that gets sent up the chain is shaped not by reality but by this infinite regress of anticipated expectations. Data gets cherry-picked, risks get minimized or exaggerated, and presentations get designed not to inform but to align with perceived preferences at each level of review. By the time a proposal reaches the decision-makers, it has been polished and reshaped so many times to match expected preferences that its connection to ground truth has become tenuous at best.
The same dynamic explains why committees so often make decisions that no individual member would endorse. No one is actually looking at the merits of the decision; everyone is trying to vote the way they think everyone else wants to vote. The result is a choice that perfectly reflects the group’s collective belief about the group’s preferences, which may bear no resemblance to anyone’s actual preferences or to what would genuinely serve the organization’s interests.
The Keynesian Casino
Keynes himself was troubled by the implications of his own metaphor. He worried that when markets become dominated by beauty-contest reasoning—when speculation about speculation overwhelms consideration of underlying value—capital gets misallocated. Resources flow not to the most productive enterprises but to whatever has momentum in the collective imagination.
He compared this to a casino, but the analogy undersells the strangeness. In a casino, the house edge and odds are fixed and knowable. The roulette wheel doesn’t change its physics based on what gamblers think other gamblers are thinking. But in markets structured as beauty contests, the game itself mutates based on the meta-reasoning of its players.
This creates what might be called “epistemic pollution”—a situation where the signal (actual value) gets so overwhelmed by noise (speculation about speculation) that price discovery fails. The market, which theoretically exists to efficiently allocate capital based on information about real economic activity, instead becomes a self-referential loop of guesses about guesses.
Breaking the Mirror
So what’s the solution? How does one win a game where thinking too carefully guarantees defeat, where the object of judgment becomes irrelevant, and where everyone is trapped trying to predict everyone else’s predictions?
Keynes himself suggested that the healthiest markets are those where participants have genuine long-term interests in underlying value—where investors actually care about whether companies will thrive over years and decades, not just whether their stock prices will tick up next quarter. When enough players refuse to play the beauty contest and instead stubbornly evaluate fundamentals, the game changes structure. The picture starts mattering again.
This is easier prescribed than practiced. Any individual who ignores market sentiment to focus on fundamentals will underperform in the short run whenever beauty-contest dynamics dominate. Being right about value while the market rewards speculation is a costly form of correctness. It requires either enormous conviction, enormous patience, or enormous capital reserves—ideally all three.
Warren Buffett’s famous advice to “be fearful when others are greedy and greedy when others are fearful” is essentially a prescription for breaking the beauty contest cycle. When everyone is playing the meta-game of predicting predictions, the contrarian move is to stop playing that game entirely and revert to level-zero thinking: just look at the damn picture. Is this company actually good? Does this asset actually have value?
The paradox is that this only works if you’re one of the few doing it. If everyone suddenly became a pure fundamentals-focused value investor, markets would stop functioning as vehicles for short-term profit through speculation. The beauty contest persists because it serves certain functions—price discovery, liquidity provision, risk transfer—even as it distorts others.
The Broader Implications
The beauty contest extends far beyond markets. Consider professional incentives in academia, where researchers must publish to advance their careers. The “picture” would be genuine contribution to knowledge. But hiring and tenure committees don’t have perfect information about intellectual merit, so they rely on proxies: publications in prestigious journals, citation counts, reputation signals. Smart academics therefore optimize not for genuine discovery but for producing work that fits the templates prestigious journals favor and that will attract citations from other academics who are themselves optimizing for citations.
The journal editors, meanwhile, choose papers based partly on quality but also on what they think will be most cited and discussed—because citations and attention determine a journal’s prestige. Reviewers recommend papers based partly on merit but also on their predictions about what editors want and what will generate controversy or attention. And so the whole system becomes an elaborate beauty contest where everyone is trying to predict what everyone else values, while the goal—advancing human knowledge—fades into the background like the faces in Keynes’s newspaper.
Or consider dating markets, where the picture is genuine compatibility and chemistry. But apps and social dynamics often transform courtship into a beauty contest where people signal the traits they think others think others want, choosing partners based not on authentic connection but on social proof and status markers that indicate other people find them desirable. The relationship you actually want matters less than the relationship you think will be validated by your social network’s expectations of what you should want.
The Unwinnable Game
Perhaps the deepest lesson of the Keynesian beauty contest is that some games can’t be won—only exited. When you recognize you’re in a situation where recursive reasoning about others’ reasoning has replaced actual evaluation of reality, the sophisticated move isn’t to think one level deeper. It’s to ask whether you want to play at all.
This doesn’t mean rejecting strategic thinking or ignoring social dynamics. It means recognizing when the meta-game has become so divorced from substance that participating means abandoning your own values and judgment. Sometimes the winning move is to look at the picture, decide what you genuinely think is beautiful.
Keynes invented his metaphor as a critique, not a playbook. He wanted to illuminate a dysfunction in markets, not celebrate clever speculation. The true lesson isn’t “think about what others think about what others think.” The true lesson is simpler and more subversive: when everyone else is playing a game of mirrors, the most radical act is to turn around and look at something real.
