plateluck.
DECISION LAB

Smart enough to help.
Random enough to be fun.

A plain-English explanation of how PlateLuck filters, scores and randomizes nearby restaurant choices without pretending to know more than the map data provides.

Step 1: hard filters

Your radius, selected cuisine, rejected restaurants, chain filter and Open Now setting can remove a restaurant before scoring begins. When Open Now is enabled, PlateLuck only keeps places whose mapped hours can be interpreted as currently open.

Step 2: fit scoring

Remaining places receive points for proximity and for matching the active mode or Smart Pick settings. Date Night, Family, Quick Bite, Healthy-ish and Local Night all use different soft signals. Complete map information such as an address, website or opening-hours value can add a small score because it makes the result more usable.

Step 3: repetition penalty

Restaurants that appeared recently get a temporary score penalty. A saved restaurant is not automatically favored; saving is for your short list, not paid placement.

Step 4: controlled randomness

PlateLuck ranks the candidates, then chooses from the strongest group instead of always taking number one. The Surprise Level changes how wide that group can be and how much random noise affects ranking.

What PlateLuck does not know

PlateLuck does not have reliable live menu prices, food-quality scores, wait times, health inspection grades, reservation availability or universal restaurant ratings. “Cheap,” “healthy-ish,” “date night” and similar settings are preference heuristics based on map categories, cuisine tags, restaurant names and available metadata—not objective claims.

No pay-to-win ranking

The core picker does not boost a restaurant because it paid PlateLuck. If sponsored placements are added in the future, they should be clearly labeled and kept separate from the ordinary spin.