How we calculate

Where the data comes from

None of this is opinion. The site reads the battle history of top global players through the official Supercell API and counts what happened: who played which card, in which deck, and who won.

Window considered

last 21 days, since October 2, 2026

Cards with measured battles

123

The bias, measured and owned

The collector reads the history of players at the top. When one of them faces someone outside the top, only their side enters the count — and that side wins more often. The result is that the average win rate is not 50%.

Ranked

50.0%

Top 200

53.4%

That is why the score never compares a card against a fixed 50%: it compares against the average of its own band. A score of 50 means "exactly the average of that environment". In Top 200 that average is 53.4%, in ranked it is 50.0%.

What is a decision and what is provisional

The Meta Score mixes win rate with consistency across decks. Not everything in that formula is settled:

• Decided by Allan: the heavier weight of Top 200, the idea of lowering confidence for small samples, and the consistency criterion (a card makes DIFFERENT decks win, and each deck counts as one vote).
• Provisional, chosen by us: the remaining weights, the smoothing, the consistency penalty and the tier cut-offs. Do not treat them as final.

9 cards have too small a sample for the tier to be reliable. Their badge appears dimmed in the ranking.

What we cannot know

• How much gold or how many gems you have. The Supercell API does not report it, which is why no site in the world shows it.
• Low arenas do not appear in the main ranking: nobody in the Top 200 plays with starter cards only. Arena decks come from a separate collection, from players in that range.
• Hitpoints, damage and DPS come from the Clash Royale Wiki, at level 11, and have not yet been checked inside the game.