The Uncomfortable Math
Happy Valentine's Day. Here is the most unromantic thing you will read today.
In 1962, two mathematicians named David Gale and Lloyd Shapley sat down and solved one of the most fundamental problems in economics: how do you pair people up when everyone has preferences? Their answer was an elegant algorithm that always produces a stable matching, one where no two people would rather ditch their partners for each other.
Shapley got a Nobel Prize for this work in 2012. And buried in the math is an uncomfortable fact that polite society has been ignoring ever since: whoever does the asking gets systematically better partners.
Not a little better. Provably, mathematically, optimally better. The proposing side gets their best possible partner out of every stable matching that could exist. The receiving side? They get their worst.
So if you accept the cultural norm that men do the asking and women do the waiting, and if you look at the dating market through the lens of algorithmic game theory, men have been running an incredibly favorable protocol this whole time. Yes, I am a single man writing this on Valentine's Day. My motives are exactly what you think they are.
The Mechanism Is Simple. The Asymmetry Is Not.
The algorithm takes a group of proposers and a group of receivers, each with a ranked preference list of everyone on the other side, and runs in rounds:
while some proposer is unmatched: proposer asks their next-highest-ranked receiver if receiver is unmatched: they tentatively accept else if receiver prefers this proposer over current partner: they trade up (old partner is now unmatched) else: they reject the proposalProposers work down their list. Receivers can only trade up or stay put. The process always terminates, always produces a stable matching, and always gives the proposers their best possible outcome. Three guarantees from a procedure short enough to fit on an index card.
Same People, Same Preferences, Different Outcomes
The simulator below runs the algorithm both ways. Hit Play, and watch the average ranks diverge. Then flip who proposes and run it again. Same people, same preferences, completely different outcomes, determined entirely by who had the nerve to ask first.
Preferences are randomly generated each time you reset, so run it a few times to see the pattern hold across different draws.
Run the simulation to see how match quality differs between proposers and receivers.
Step Log
At Scale: 100 vs 100
Ten people is useful for watching the mechanism, but small enough that individual runs can be noisy. Below, the same algorithm runs on 100 men and 100 women, showing both "men propose" and "women propose" simultaneously with the same preferences. Each dot is a person, positioned by the rank of the partner they ended up with. Left is good (#1 choice), right is bad (#100 choice).
The clustering difference is the whole argument in one picture. Hit Regenerate a few times to watch it hold.
The Side That Bears the Search Cost Captures the Surplus
The core insight survives every real-world complication you can throw at it: the side that bears the search cost captures the matching surplus. People do not have neatly ordered preference lists. The market is not perfectly two-sided. Preferences shift, information is incomplete, and nobody is running a formal algorithm over brunch. All true. But the structural claim does not require clean inputs to hold. It requires only that one side initiates more than the other, and that initiation is costly.
Asking someone out is expensive. It costs ego, time, and the risk of rejection. But those are upfront costs. The payoff (ending up with a higher-quality partner by your own subjective ranking) compounds for the duration of the relationship.
This is not a subtle effect. The theorem does not say proposers do slightly better on average. It says proposers get their best possible stable partner and receivers get their worst possible stable partner. The maximum asymmetry that stability allows.
The Economics: Search Costs vs. Lifetime Returns
The intuitive objection is obvious: "Sure, proposers do better, but asking people out sucks. You're paying in rejection, awkwardness, and ego damage every time you shoot your shot." Fair enough. But let's actually do the math on that tradeoff.
The simulator below models 40 years of romantic happiness specifically. Each run generates a fresh 100-person Gale-Shapley matching to get the actual quality gap between proposers and receivers. Hit Play and it randomly determines how many years the proposer spends searching, then animates the two lines diverging over a lifetime.
Neither side accumulates romantic happiness until they're in a relationship. The proposer actively pays a search cost (rejection, ego, time), while the receiver's romantic happiness is simply zero until they're matched. This is defensible: we're tracking the returns from a romantic partnership specifically, not life satisfaction in general. Nobody's getting partner-quality companionship from someone they haven't met yet.
Hit Regenerate a few times. The search duration changes with each play, the exact quality gap shifts with each matching, but the proposer almost always comes out ahead by age 60. And note: this is a positive-sum game. The receiver still ends up happy. Both lines go up once the relationship starts. The proposer just ends up happier, because they found a better-matched partner by bearing the search cost upfront. The cost of rejection is loud and immediate. The cost of settling is quiet and permanent.
The Case for Women Asking Men Out
If you accept (even loosely) that the Gale-Shapley framework maps onto dating, the prescription is straightforward: women should be doing more of the asking. And critically, this is not a collective action problem. You do not need to wait for society to change norms. Any individual who switches from receiver to proposer improves their own outcome immediately, regardless of what everyone else does.
The current norm essentially has men paying a per-attempt cost (rejection, awkwardness, the cost of shooting your shot) in exchange for a massive lifetime benefit: ending up with a partner who's closer to the top of their preference ranking. Men have been socialized into bearing the search cost, and the math says that's the advantaged position.
Meanwhile, women who adopt the "receiver" strategy (waiting to be approached, filtering from whoever shows up) are provably leaving value on the table. Not because the men who approach them are bad, but because the algorithm's structure means receivers converge toward their worst stable partner, not their best.
The fix does not require a movement or a manifesto. It requires one person deciding to ask. If you want better outcomes, be the proposer. Don't wait for the market to come to you. Go find your #1 choice and ask. Get rejected. Ask your #2. Valentine's Day celebrates the people who already found each other. The algorithm says it should celebrate whoever had the nerve to go first.
The Rational Move Is Obvious
A few caveats, because the internet loves to assume you don't know about nuance unless you spell it out. Real life is not a bipartite graph. People are bisexual, polyamorous, uncertain, contradictory. Preferences are not fixed or fully known; you do not carry a ranked list of every person on Earth, though you do carry a rough sense of who you'd rather be with, and you update it as you go, which is actually close to what the algorithm models. And rejection is real. Nobody is disputing that. The argument is that the cost of rejection is finite and upfront, while the benefit of a better partner is large and ongoing. But simplified models that win Nobel Prizes tend to capture something real. The structural claim (that initiative confers advantage in matching markets) holds across enough variations to take seriously.
A Nobel Prize-winning algorithm says the proposing side gets the best deal. Cultural norms currently hand that advantage to men by default. But this is not about what women as a group should do. It is about what you, specifically, reading this right now, should do. The advantage accrues to the individual who acts, not to the gender that collectively decides to. One person switching strategies is enough to improve that one person's outcome.
Ask them out. It is Valentine's Day. You will never have a better excuse. The worst they can say is no. And then you move to #2 on your list, who was probably great anyway.