About the Method

Archetypes first. Card relationships second.

The site starts from real draft and game data, learns recurring deck families, and then looks inside each family to see which cards define it, which cards need support, and which combinations win more than you would expect.

At a high level, the method uses full decklists to learn archetypes, then measures card and pair performance inside each learned shell instead of treating the whole format as one big pool. Draft-side information is used separately to estimate urgency, ALSA, and whether a card is likely to come back.

The goal is not to replace normal draft instincts with a single ranking. The goal is to make the format easier to read: what the real decks are, what holds them together, and which picks matter most once you are in a lane.

Archetype maps Decks are grouped into recurring shells from the deck data itself, so card relationships are judged in context rather than in one global table.

Synergy and support Pairs are compared to an archetype baseline, so the graph can separate cards that are strong on rate from cards that mainly get paid off when the right support is present.

Draft utility Pick-focused views combine archetype win contribution with table signals like ALSA and wheeling, so the output is closer to a drafting decision than to a pure win-rate table.

Back to formats Privacy