HOW WE CRUNCH THE NUMBERS

01 Where the Data Comes From

When you look up a player tag, we request that player's public profile and recent battle log from the official API in real time. To build meta-level stats (win rates, pick rates, tier ratings), we aggregate large numbers of these public battle logs over a rolling window and compute the figures ourselves.

This means our numbers are community estimates built from a sample of real matches, not the global ground truth that only Supercell can see. We label them as estimates on purpose, and we explain below exactly how each one is calculated.

02 Win Rate

Win rate is the simplest figure: the percentage of recorded matches a brawler won, in a given mode or map, within our sampling window.

win rate = wins / (wins + losses)

On its own, win rate can be misleading. A brawler that is only picked by experienced players, or only on maps that suit it, can show an inflated win rate that does not reflect how it performs in an average game. That is why we also compute an adjusted figure.

03 Adjusted Win Rate

The adjusted win rate is our attempt to correct for that bias. It pulls a brawler's raw win rate toward the population average in proportion to how small or skewed its sample is, so a brawler with a handful of cherry-picked games does not leap to the top of a list on noise alone.

In practice, a brawler with a very high raw win rate but a small or biased sample will show a lower, more realistic adjusted win rate. When the raw and adjusted figures disagree, trust the adjusted one. Our tier ratings are based on the adjusted figure, not the raw one.

04 Pick Rate

Pick rate is the share of recorded matches in which a brawler appeared, within a mode or map.

pick rate = matches with brawler / total matches sampled

Pick rate measures popularity, not strength. A high pick rate can mean a brawler is genuinely strong, or simply that it is fun, free, or trending. We show it for context, but it never drives a tier rating by itself.

05 Tier Ratings

Tier letters (S, A, B, C, and so on) are a readable summary of a brawler's adjusted performance in a given context (overall, by mode, or by map). They are derived from the adjusted win rate, weighted by how reliable the sample is.

Because tiers are context-specific, the same brawler can be S-tier in one mode and B-tier in another. That is the point: a tier is only useful next to the mode or map it describes.

06 Counters & Matchups

A matchup compares how two brawlers perform when they face each other, relative to how each performs on average. A brawler is a "counter" when it wins that head-to-head meaningfully more often than expected.

We treat the data as the starting point, not the final word. On counter pages we add a written explanation of why a matchup goes the way it does (range, dash, shield, reload, area control) so the advice holds up in a real game even when a single sample is noisy. Where the data is thin or the explanation is unclear, we say so rather than presenting a shaky number as fact.

07 Sample Sizes & Thresholds

We use minimum-match thresholds before showing a figure so that tiny samples do not masquerade as reliable stats:

• Brawler-by-mode stats require a minimum number of recorded matches before they are displayed.
• Map rankings use a higher minimum, because the data is split across more buckets.
• When a combination falls below threshold, we either hide the figure or mark it clearly rather than guessing.

If you ever see a result that looks too good to be true, check the match count next to it. A small sample is the usual reason, and it is why the adjusted figures exist.

08 How Often We Update

Player lookups are live. Aggregated meta stats are refreshed on a rolling basis, and the data pipeline is rebuilt after every balance change so the meta you see reflects the current patch. Each stats page shows an "updated" date so you always know how fresh the numbers are.

09 Limitations We Are Honest About

• Our figures are estimates from a sample, not Supercell's full global dataset.
• Samples can be biased toward higher-skill players, who are more visible through rankings and clubs.
• A fresh patch or a brand-new brawler will have thin data until enough matches accumulate.
• Tier lists and matchup verdicts are guidance for decisions, not guarantees of a result.

If a figure on the site looks wrong to you, you are probably onto something. Tell us and we will look into it.

10 Questions or Corrections