How the Tier List Is Calculated

We use only the weekly data (win rate, pick rate, ban rate) that Starward's official data site publishes for the 6000–8000 rating band. We never estimate or fill in data we can't obtain, such as card pick rate or match counts.

We use the 6000–8000 band because it's the only one where the pick-weighted average win rate comes out to exactly 50% (i.e., matches within the band are self-contained), which is what makes "probability of winning when played" a well-defined win rate.

• Win rate — The probability that the character wins when played.
• Pick rate — The probability that a randomly chosen player's played character is this one (its share of all match slots).
• Ban rate — When in hand, the conditional probability of being banned. Since the total across all characters comes out to about 1400%, we've confirmed this is a conditional probability, not a share.

Score = win rate (shrinkage-adjusted) × 0.45 + ban rate × 0.30 + ban-adjusted pick rate × 0.25

• Win rate shrinkage — A character's win rate gets inflated by statistical noise and by an "only committed players use it" bias when its pick rate is low. We pull it back toward the theoretical 50% in proportion to how little data there is, so a lucky high win rate doesn't land a character at the top on its own.
• Ban-adjusted pick rate — Pick rate ÷ (1 − ban rate). A character with a 70% ban rate can only appear in matches that "got past the ban," so its raw pick rate reads lower than its true strength. This corrects for that ban-driven thinning to show its true presence.
• Within-cost standardization — Since the three metrics differ hugely in scale, we align their medians and spread within each cost before combining them (a robust z-score). Every evaluation is always "relative standing within the same cost."

We split the absolute score (in robust-z units) into 5 tiers — S / A+ / A / A− / B — using fixed thresholds. Since it's not a fixed percentage like "the top 20% is always S," a week with no standout characters can leave S empty. The table groups the A tier (A+/A/A−) into a single row, distinguished by frame color and a badge in the card's top-left corner.

Characters with a displayed pick rate of 0.1% or lower are marked with * because their win rate carries a lot of statistical noise. Shrinkage pulls their score back toward neutral, but their evaluation is still less reliable than other characters'.

• The weight split (0.45/0.30/0.25) is an editorial choice. That said, we've confirmed that shifting any weight by ±0.15 barely changes the rankings.
• Ban rate includes a reputation component — "banned because it's famous" — which may cause some characters to be over-rated relative to their actual strength.
• We can't get data on team composition (which characters/costs are paired together), so we can't evaluate synergy.

The full method spec and statistical validation notes are in docs/tier-methodology.md and docs/data-observations.md in the repository.