How We Score

We start with the awards, and only the awards. Published, professional honors from the experts: inspection guides, judged competitions, critics' lists. No crowd stars, no votes, no paid placement, and none of our own taste. If the experts haven't weighed in on a place, it isn't here.

These are the publications whose honors count right now (the list keeps growing):

Then we weigh each honor, because a Michelin star and a spot on a listicle aren't the same thing, and we're not going to pretend they are. Four questions set the weight. Who gave it: an inspection body like Michelin or Forbes carries the most trust, then a judged competition like the James Beard Awards, then an editorial list, then a marketing one. How far it reaches: the world, a country, a city, a neighborhood. Where it landed, exactly: a Winner outranks a Finalist outranks a Semifinalist, and #2 on a ranked list outranks #7 (interpolated, never rounded up). And how old it is: durable wins fade slowly, "best new restaurant" fades fast, and a current inspection rating doesn't age at all, because it gets re-checked.

Then we add them up, with one rule doing most of the work: an award can only help. The strongest honor sets the floor, and every other honor closes part of the remaining gap to 100. Nothing a place wins can ever lower its score. (We used to average. Averaging punished a place for winning a small thing alongside a big one, which was a bug, so it's gone.) Prestige beats volume: a pile of local lists barely moves the number, one big honor moves it a lot, and a second honor from the same publication counts a fraction of the first. Reaching 100 takes a world-number-one peak with the full slate of elite corroboration behind it. It's been done.

Then there's distance, which does exactly nothing to the number. When you sort by what's near you, a good pizza place next door beats driving 45 miles to a great one on a Thursday night, so distance reorders the list. That's all it does. The printed score is the raw score, everywhere, always: distance changes the order, never the score.

We take the point and weigh every scored place to eat and drink around it by distance. Next door counts more than across town, and across town counts more than a highway drive. Then we average the best 25. If there aren't 25, the empty slots count as zero, so a couple of great places can't carry an area on their own. Museums and parks go through the same math, but for now they barely move the number. We haven't found enough honors for them yet. What comes out is the good stuff within reach.

That's sixty percent of the score. The other forty is the setting. Most of it is the country the point sits in, measured the only way we trust: by how life actually turns out for the people who live there, never by the policies that are supposed to produce them. How residents rate their own lives (the Global Flourishing Study, Eurostat, and the OECD). How long and how well they live (the UN's inequality-adjusted Human Development Index and World Bank life expectancy). How safe they are (the UN's homicide count, conflict deaths from the Uppsala Conflict Data Program, and the World Risk Poll, which asks people whether they were actually harmed). And the rest is the environment right at the point: the air (fine-particle pollution, PM2.5) and the weather (monthly temperature and humidity from CHELSA climate data), scored month by month for how comfortable it feels. Doesn't count: our own taste in restaurants.

One honest note about the weather. Clean air is better for everybody, but a good temperature is a preference, not an outcome: we score it against a mild, comfortable band, so a place with punishing summers or hard winters comes out lower even where the people who live there wouldn't trade it. That part is a judgment, not a measurement, so we keep its weight light: the good stuff nearby and how life actually turns out there carry far more.

A point with no scored place within reach still gets the country share, so the countryside in a country that's good to live in isn't a zero. It's a low number that says exactly what it means: nothing recognized nearby, in a place where life is good. It's the same score behind the map's heat, and it's the number members see on their own home address.

How strong its very best are, worth 40 of 100: the top five in eating, drinking, and staying, averaged across the modes the city actually has. Doing is not in the blend yet, because we haven't found enough honors for it. It joins once we have.

The highest Best Place Score any address in it reaches, worth 20.

How many places score 80 or better, worth 25, counted up to a hundred, past which a city is simply deep. The cap is the point: a raw count would measure how much we have gathered, not the city.

How many neighborhoods it has that are worth the name, worth 15.