The methodology

How rankings work

Rankings use binary criterion recommends, recommend-rate scoring with a stabilizing prior, origin-connected lenses, and editorial guardrails. No paid placement — curated entry, trusted member signal.

Recommend-rate scoring

Members recommend a restaurant per criterion (authenticity, taste, value, vibe) or leave it blank. Vibe includes service and atmosphere. Each recommend is yes/no — not a 1–5 star score. Overall rank is computed from those four rates using transparent weights; it is not directly votable.

Sparse listings start with a small prior weight so one loud recommend cannot dominate. As more trusted members recommend, the percentage reflects real community conviction.

Prior weight (default)

5 virtual recommends

Display

Recommend %

Origin-connected lens

Members with demonstrated cultural connection to a cuisine can recommend in an origin-connected segment. Lenses let you compare community-wide signal with heritage-weighted insight.

Vibe & room

Vibe is one signal for how it feels to eat there — service, atmosphere, and energy together, not a separate service score.