Click a chosen card to compare replacements · Fit = its measured contribution to the deck score · Missing evidence scores zero · Contributions overlap, so do not add them together
Synergies within your deck
0 measured pairs out of 1 possible · No qualifying pair evidence in this selection
Best next cards for your whole deck
Ranked across every selected card, not just your last pick
No qualifying suggestions for these cards in this snapshot · You can still add any catalog card
Deck-fit score = sum of measured signed pair lifts in percentage points ÷ selected card count. Negative associations reduce the score; insufficient evidence contributes zero, not proof of incompatibility. Every card form is compared with every other form independently. These are pairwise associations, not a measured full-deck win rate.
Excluded cards · 0
Exclude suggested forms you don’t own or want · They stay out of the catalog, suggestions and auto-fill · Other forms remain available
Your deck and exclusions are saved in this page’s URL · Bookmark the page to keep them · Deck share links omit personal exclusions
Win rate counts draws as half a win. Every variant is compared with every other variant. Signed lift = pair win rate minus the larger of each card's win rate without the other. A pair is measured when together and both without-partner groups meet minGames; nonpositive measured lifts remain in pairStats. Forms are never pooled.
Observational association, not causal synergy or statistical significance. Top-player sampling, deck composition, repeated players, levels and tower troops can confound results. Both sides of a battle count, so observations are not independent. Same-base forms cannot coexist and have same_card status. Insufficient comparisons preserve available counts and rates but have null lift. The cards lists contain every measured positive partner. Recent battle-log snapshot only; no historical accumulation.
3,746 deck observations · Minimum 20 observations in each comparison group · Up to 181 partners per card