What this number is not

Let's do the disclaimer first, at full volume, because the number in the title looks like a statistic and is not one.

The 3:10 ratio — the claim that three well-provisioned, coordinated households can carry ten through most common disruptions — is a design hypothesis. It is not a measurement. There is no dataset behind it, no study, no after-action archive from which it was computed. We have never measured it, and as of this writing, neither has anyone else that we know of. An earlier version of this site presented the ratio as though it were an established fact of neighborhood dynamics. That was wrong of us, and this rewrite exists partly to say so in public.

So why keep it at all? Because a hypothesis, honestly labeled, is one of the most useful objects an engineering project can own. It tells you what to build and — just as important — what evidence would prove you wrong. Here is ours, with its work shown.

Why we think the hypothesis is plausible

The 3:10 shape didn't come from nowhere. It comes from three observations about how capability and need are distributed on an ordinary street.

Capability is lumpy; need is shared. A 2,000-watt generator is dramatically more machine than one household needs and roughly enough for ten, run as a rotation — a refrigerator only needs an hour or two of power a day to hold safe temperatures. A countertop gravity filter that produces 30 to 40 gallons a day is a street's drinking water, not a family's. A deep pantry cooks more efficiently by the potful than by the portion. Most resilience hardware comes in street-sized lumps, which means the surplus already exists; the missing ingredient is the routing.

Specialization beats spreading thin. Three households investing, say, $2,000 each — one deep in power and tools, one in water and food, one in medical and communications — produce far more usable capability than ten households each spending $600 on a shallow copy of everything. Depth also carries the skills along with the objects: the household that owns the radio is the household that learns to use it.

Three is the smallest honest mesh. We keep returning to this number across the whole project — it's the same three as who are your three. One anchor household is a favor; two are a lifeline that fails during someone's vacation; three form a network that survives losing a node. And ten is roughly the number of front doors within a shout on a typical block. So 3:10 is really a topology claim dressed as arithmetic: the smallest resilient core, serving the largest naturally-occurring social unit. Plausible. Elegant, even. Still unproven.

What would confirm it — and what would kill it

A hypothesis you can't imagine losing isn't a hypothesis; it's a slogan. So, concretely:

Evidence that would confirm it. After-action reports from real outages on streets using the shared inventory: did roughly three capability-dense households, in practice, keep roughly ten functional — fridges cold, water clean, medications managed, everyone informed? Coverage data across many declared street inventories: when three households' declared surplus is laid against twelve capability axes, how much of a street's need does it actually blanket? If the pattern holds across dozens of streets and several kinds of disruption, the ratio graduates from hypothesis to finding, and this post gets its third rewrite — a much happier one.

Evidence that would kill it. Several failure modes would, and we're watching for all of them. Anchor burnout: if carrying seven other households reliably exhausts the three within days, the ratio is a recipe for resentment, not resilience. Free-riding: if visible surplus teaches the seven to invest nothing, coordination could actually lower a street's total capability over time. Coordination overhead: if routing surplus to need costs more attention than it saves past four or five households, the mesh frays exactly when it's needed. Or the number is simply wrong — the real ratio might be 3:6, or 5:10, or it might vary so much by disruption type that no single ratio deserves a name. If the data says any of that, we will change the number, and the title of this post, and say why.

The honest current status: we hold anecdotes and topology, not data. Anecdotes and topology are enough to design by. They are not enough to advertise with — which is why every appearance of this ratio on this site now carries the word hypothesis, enforced by an automated audit that fails our own build if we forget.

The ledger of retired claims

This isn't the first number we've had to walk back, and the pattern is worth naming because it's the actual subject of the honest-engine stream.

Earlier versions of Terrarium displayed a neighborhood readiness figure of "about 70%." It was an estimate a model produced once, and it looked exactly like a measurement. We retired it, published the computed Readiness Score that replaced it — three components, every weight on the table — and wrote up the whole story of the retirement so the new number could never pull the old trick. The 3:10 ratio is the same policy applied one door down: a number that can't show its work doesn't get to wear a percent sign, or a ratio's confident little colon, without a label around its neck.

Why be this fussy? Because this store's entire pitch is that the score means something. Precision theater — fake numbers with real fonts — is the cheapest trust to borrow and the most expensive to repay. We'd rather tell you "we believe this and haven't proven it" than be caught believing our own marketing.

Run it on your own street

Here's the good news buried in all this epistemic hygiene: this hypothesis is testable at the scale of one block, by you, without our permission.

Find your three — the capability-dense households on your street, which may include yours. Get the inventories declared, so the surplus is visible. Then let the next ordinary disruption — the six-hour outage, the boil-water notice — run its course, and take notes. Did three carry ten? Did it cost the three too much? Send us what happened; street-level after-action notes are exactly the evidence this number is waiting on.

Either way, you'll have built the mesh. And whatever the true ratio turns out to be, we're confident of its shape: coordinated streets outperform siloed ones. That part isn't a hypothesis. That part is just what neighbors are for.