Methodology
Playmatch treats game experience, persistent player taste, and what someone wants today as separate signals. This page explains how those signals become a recommendation without presenting inference as fact.
Game DNA is a versioned profile of qualities such as pacing, progression, exploration, difficulty, stress, social dependency, and session length. It describes how a game tends to play; it does not claim that every player will experience it identically.
Initial profiles can use Steam storefront metadata, developer information, assisted metadata analysis, and editorial curation. Player feedback and observed behaviour are separate evidence types. Playmatch does not label AI-assisted inference as observed player evidence.
Every trait and profile carries confidence. Public game pages require complete metadata, meaningful descriptions, established reception, a current Steam destination, and sufficiently supported active Game DNA. A database row alone does not make a page suitable for search.
A similar-games score compares two games through shared Game DNA. A player-match score also uses a person's taste, and an intent-aware match adds their current request. Playmatch labels these scores differently because they answer different questions.
Commercial availability may determine which buying options appear, but affiliate economics never increase an organic recommendation score. Today, Playmatch shows Steam as its verified store destination.
Recommendation versions, explanations, impressions, reactions, game opens, and Steam clicks can be measured. Synthetic tests verify mechanics only; they are not represented as real-player validation.
The evidence each of these signals comes from is listed on the source register, and the rules Playmatch writes under are in the editorial policy.