living together · metaphor 31 of 100
Meet a stranger in New York tomorrow — no time agreed, no place agreed, no way to communicate. Most people solve this "impossible" problem identically: noon, Grand Central. Coordination without communication runs on salience, not optimality — and then the obvious starts to rule us.
Nothing makes Grand Central correct. It is not the biggest place in the city, not the most central, not the oldest, not the first in any alphabet. Its entire power is that each of you expects the other to think of it — and expects the other to expect the same in return. Prominence, passed once around that loop, becomes self-fulfilling: the place is obvious because everyone treats it as obvious, and everyone treats it as obvious because it is.
Once you see the pattern, it is everywhere. Custom, precedent, round numbers, borders that follow rivers, "we split it 50/50" — enormous stretches of social life are settled not by argument or authority but by the quiet magnetism of whatever sticks out. Thomas Schelling called these landing sites focal points, and their magnetism has a mathematics. Knowing it shows you something uncomfortable: who chose your defaults, and how hard they will be to move.
Game 1 · the meeting
Noon tomorrow, somewhere on this island, a stranger from the crowd below is waiting for you. Click (or tab to the map and use arrow keys + Enter) where you'd wait. The crowd's heat map reveals only after you commit. Your score is the chance a randomly drawn crowd member is standing exactly where you are.
Game 2a · name the same number
You and a random crowd member each name a number from 1 to 100. If you name the same number, you both win. No number is better than any other — except that some are obvious.
Game 2b · split $100 with a stranger
Each of you writes down the share of $100 you claim (a positive amount). If the two claims describe the same split, the deal happens; otherwise both of you get nothing.
Game 3 · precedent mode — a town with no custom, yet
A new town: four plausible meeting spots, almost perfectly tied in the crowd's mind. Play the meeting game three times. After each round, the crowd's expectations shift toward wherever the most meetings happened — computed, not narrated. Watch an arbitrary early choice harden into "the obvious place." Your round-1 vote is enough to break the tie.
Game 4 · tie-break lab — what makes a place salient?
If salience were a geometric or historical property, a formula could find the focal point. Below, four reasonable tie-breaking rules are each run against the meeting-game map (computed live from the landmark data), and their picks are scored against the crowd.
| tie-break rule | its pick | crowd there |
|---|
Every defensible rule loses to "the place where movies set meetings." Salience is a property of the culture reading the map — which is why it can be owned, inherited, and sold.
coordination without a channel
In a pure coordination game, you don't want to be right — you want to be together. Any of a hundred meeting places would do, provided you both pick it; and that "any" is exactly the trouble. Reason can rank options by quality, but it cannot rank them by being chosen twice, because every candidate is equally good the moment both of you land on it. Deduction runs out before the decision does. Something must settle the residue that reasons cannot reach, and what settles it is prominence: the option that sticks out, for people like these, here, now.
Schelling's deep observation is that this is a skill of social imagination, not logic. The question you actually answer is "what will they think I'll think they'll think?" — a hall of mirrors that converges only if some image in it is brighter than the rest. The focal point is that bright image. It carries no argument for itself. It doesn't need one.
what to try
The meeting: pick the spot that feels cleverly non-obvious. Commit, watch the heat map, and feel the arithmetic scold you: cleverness scores near zero, because your score isn't a property of your choice — it's the crowd's density at your choice. Then take Grand Central like everyone else and notice how it feels: not smart, just correct in the only sense that counts.
Name a number: commit to your number before peeking. The crowd's spikes — 1, 7, 50, 100 — are not better numbers; they are more imaginable ones. Then try the money split and watch a moral norm appear out of pure prominence: 50/50 towers over every other division, not because it's fair but because fairness is findable.
Precedent mode: in round 1 the four spots are tied and your single vote breaks the tie. By round 3, the spot you happened to pick holds most of the crowd, and deviating would cost you almost everything. Watch your round-1 whim become round-3 law — then read the log's last line slowly.
who owns the obvious
Prominence feels like a property of things. It is a property of minds — and minds are furnished. Grand Central is focal because a century of films staged reunions under its clock. Round numbers are round in base ten, an accident of fingers. Rivers make "natural" borders only to people raised on maps that ink them darkly. The tie-break lab makes the point computable: every neutral-sounding rule — biggest, oldest, most central, alphabetical — misses the true focal point, because the true focal point was installed by culture, not geometry.
This is why salience is a form of power, and a strangely deniable one. Whoever makes an option come to mind first — the platform's default setting, the anchor price, the candidate "everyone's talking about" — never has to win an argument. Agenda-setting is focal-point engineering: you don't force the choice, you furnish the imagination that the choice will be made with. The obvious has owners, and they are rarely on the ballot.
the conservatism of convenience
Precedent mode shows the mechanism in miniature: once expectations concentrate, every individual is better off conforming to a focal point they never endorsed. The QWERTY keyboard, the awkward meeting time, the constitution's strange clause — these persist not because anyone still argues for them but because coordination is expensive to renegotiate. Deviating alone is worse than useless, so critics comply, and their compliance is read by everyone else as endorsement, thickening the very precedent they despise.
Which is why focal points don't move by private persuasion. Convincing people one by one that the default is bad changes nothing, because each convert still expects the others to show up at the old spot. What moves a focal point is a public, common-knowledge event — a coronation, a devaluation, a front page, a moment everyone knows everyone saw — that resets what everyone expects everyone to expect, all at once. That machinery has its own page: common knowledge.
the mapping
| Mathematics | Life |
|---|---|
| coordination game | Any choice whose worth depends on others choosing alike — where to meet, what to drive on, what counts as polite. |
| salience | What comes to mind first, for people like these — trained by maps, media, movies, and round numbers. |
| the focal point | Custom, precedent, the default: the option that wins by being findable, not by being good. |
| self-fulfilling expectation | Obvious because expected, because obvious — the loop that turns prominence into fact. |
| precedent dynamics | How one successful coordination hardens into an institution nobody remembers choosing. |
| renegotiation cost | Why the obvious keeps ruling long after it stops being good — and why moving it takes a public event, not a private argument. |
where the metaphor tears
Focal points presume a shared culture doing the imagining. Across cultures — or subcultures, or generations — the "obvious" diverges, and coordination fails precisely where it feels most natural, because neither side suspects the other's obvious is different. The metaphor's sharpest warning is against projecting your own salience: what "anyone would think of" is always anyone like you.
Prominence is a tie-breaker, and it is strongest when the options are genuinely near-tied. Make one option clearly better — a bigger prize, a safer route, a cheaper standard — and payoff differences override salience. Not every custom is an arbitrary focal point; some defaults survive because they earn it. The metaphor explains inertia, not everything that stands still.
The population you played against is a declared toy: its distributions were authored to echo classic experimental findings (Schelling's informal polls; Mehta, Starmer & Sugden's coordination studies), not sampled from anyone. The computations against it are exact, but the crowd itself is a portrait, not a measurement — and a mid-century, Anglophone portrait at that. Real crowds must be asked.