Based on 9,370 Ranked deck observations tracked over the last 3 days, Minion Giant is currently CrownStats' highest-rated Clash Royale card. Crown Rating uses clean matchup performance and statistical confidence, while Usage, Win Rate and Clean Win Rate are shown separately.
This window includes some historical battles from before CrownStats began capturing exact Evolution/Hero forms. Knight, Valkyrie, Musketeer, and Wizard alternate-form battles from that period are labeled “Alternate form (historical)” rather than guessed.
Minion Giant
X-Bow
Hero Dark Prince
Berserker
Evolution Furnace
Mega Knight
Little Prince
Evolution Knight
Evolution Baby Dragon
Hero Knight
Evolution Archers
Hero Tombstone
Evolution P.E.K.K.A
Fireball
Ice Golem
Rascals
Barbarians
Suspicious Bush
Cannon Cart
Rage
Skeletons
The Log
Ice Spirit
Dark Prince
Freeze
Graveyard
Electro Wizard
Bowler
Evolution Goblin Barrel
Barbarian Barrel
Skeleton King
Evolution Cannon
Evolution Mortar
Hero Mini P.E.K.K.A
Evolution Minion Horde
Giant Skeleton
Hero Wizard
Evolution Skeleton Army
Evolution Battle Ram
Princess
Baby Dragon
Royal Chef
Electro Spirit
Zappies
Tornado
Balloon
Lightning
Goblin Gang
Usage Rate is the percentage of eligible tracked decks in the selected battle population and window that include the card — a measure of popularity, not performance.
Win Rate is wins divided by decisive battles (wins + losses). Draws are recorded separately and never counted as half a win.
Clean Win Rate excludes battles where both opposing decks use the same ranked card identity, removing mirror-match results that can pull heavily-used cards toward a misleading 50% ordinary Win Rate.
Crown Rating is a confidence-adjusted performance score based on the lower bound of a 95% Wilson confidence interval for Clean Win Rate — a score, not a percentage (it is never shown with a % sign). This is a standard, published statistical technique (the same idea behind many “best/top” sorts elsewhere on the web) — not a proprietary secret formula. It rewards a high Clean Win Rate only when there's enough sample to trust it, and never factors in Usage.
Sample Confidence bands (Very Low / Low / Medium / High / Very High) show how many clean, decisive battles back a card's numbers. Low-sample cards are never hidden from the rankings — Crown Rating's confidence adjustment already accounts for them.
Trends compare the selected window against the immediately preceding window of the same length (e.g. this 7 days vs. the previous 7 days) — never an overlapping range.
Why CWR matters, right now: in this exact selection, Tower Princess appears in 93% of its own matches as a mirror (both sides using it) — those matches are automatically excluded from its Clean Win Rate, so its 49.9% ordinary Win Rate and 49.2% Clean Win Rate can genuinely differ.
CrownStats ranks cards by Crown Rating — a confidence-adjusted measure of Clean Win Rate computed from real Ranked battle data CrownStats has crawled. The default view shows Ranked battles from the last 7 days; use the time and battle-population filters above to see other windows.
Crown Rating is the lower bound of a 95% Wilson confidence interval applied to a card's Clean Win Rate. It rewards cards that win consistently across a large, reliable sample, and penalizes cards whose high win rate comes from too few battles to trust yet. It never factors in how popular (used) a card is — Usage is shown as a separate number.
Clean Win Rate (CWR) is the win rate for battles where the opposing deck does NOT use the same card in the same form — mirror matches are excluded. This removes matches that are guaranteed to end near 50% regardless of whether the card is actually strong.
Ordinary Win Rate includes mirror matches (both sides using the same card), which pull heavily-used cards toward a misleading 50% no matter how strong they are. Clean Win Rate removes those specific battles so the number reflects performance against decks that don't share the card.
The percentage of eligible tracked decks in the selected battle population and time window that include the card. Usage measures popularity, not performance.
CrownStats crawls new Ranked/Ladder battle data multiple times per day and refreshes these rankings immediately after each crawl run that adds new data.
Crown Rating accounts for sample size. A card with a 62% Clean Win Rate over only 30 battles is far less certain than one with 53% over 200,000 battles — Crown Rating's confidence adjustment can rank the larger, more reliable sample higher.
Yes — Base, Evolution, and Hero forms of the same card are ranked as separate entities going forward, since they perform differently. For 4 cards that support both Evolution and Hero (Knight, Valkyrie, Musketeer, Wizard), CrownStats' historical data before this feature launched could only record that an alternate form was used, not which one — those historical rows are labeled honestly rather than guessed, and every new battle CrownStats crawls records the exact form.
A sample-confidence label showing how many decisive, non-mirror (clean) battles back a card's numbers. CrownStats never hides low-sample cards from the rankings — Crown Rating's own confidence adjustment already accounts for a small sample; the label just makes that visible.
Ranked includes Path of Legend battles CrownStats can reliably identify from its own crawled data. Sub-tiers like Ultimate Champion, Top 1000, or Top 200 aren't exposed yet — CrownStats doesn't yet have reliable data to identify them.
CrownStats' own statistically transparent ranking system — not affiliated with, endorsed by, or derived from any third-party Clash Royale stats site.
