How Our Algorithm Works

Mostly Magic doesn't just match you on one thing. We look at five dimensions of your character sheet, each weighted differently, to calculate how compatible you are with a game and its players.

The biggest factor in compatibility. We measure three axes — Tactics, Authorship, and Immersion — to capture how you like to play. Players with similar stat profiles tend to have better experiences at the same table.

• Tactics — Competitive play, optimization, and mechanical mastery
• Authorship — Shared narrative control, co-creation, and GM-less play
• Immersion — Deep immersion in world or story, sandbox exploration, and emergent fiction

Safety compatibility matters. We check that no player's "lines" (hard no's) conflict with the content of the game or other players' expectations. Shared veils indicate similar comfort levels.

• Line conflicts are weighted heavily — a single mismatch is a dealbreaker
• Overlapping veils indicate shared comfort zones
• Custom safety items are included in the match

Players who share favorite systems are more likely to enjoy the same style of play. System preferences also serve as a proxy for broader taste — PbtA fans often share narrative instincts, for example.

• Exact system matches are weighted highest
• System family overlap (e.g. all PbtA games) contributes partial weight
• Breadth of systems played signals flexibility

The fiction that inspires you says a lot about what you want at the table. Shared touchstones — books, movies, shows — suggest compatible taste in tone, genre, and narrative style.

• Genre overlap contributes to the match score
• Shared specific titles are a strong signal
• Helps match players on tone and theme, not just mechanics

Not a gatekeeping measure — experience level helps us suggest games where everyone will be comfortable. New players benefit from patient tables, and veterans can find the depth they're looking for.

• Similar experience levels reduce friction at the table
• GMs can set experience preferences for their games
• Weighted lower so it never excludes — just fine-tunes

The Research

Our matching model is grounded in research on player compatibility and the GNS framework originally described by Ron Edwards. We've adapted and expanded it with real-world playtesting data.