← Overview

Knockout

Split Difference

54,742 battles collected on this map.

OPEN — more choice than on most maps: 59 of 105 brawlers above 50%.

Recommendation

Hidden gems & traps

What gets played here

The most common three-brawler line-ups in this mode, sorted by frequency — not by success. That is deliberate: there are around 180,000 different line-ups, and picking the best out of them mostly turns up flukes. The win rate tells you whether what everyone plays actually works here.

Which line-up carries

Not individual brawlers but their classes — 82 combinations with enough games in this mode. The portraits show who most often fills each class on this map; they are examples, not a recommendation. The bold number is not the win rate but its distance from the value you would expect from the individual strength of the three brawlers. An example: if a class combination contains three brawlers who are strong anyway, it will often win even when the roles do not fit together at all. That is exactly what the comparison subtracts — what remains is the share of the win that depends on the interplay of the classes and not on who happens to be strong. Below it stand the actual win rate, the expected value and the number of team appearances — that is, how often this combination stood on the field as a team at all.

All brawlers on this map

Including the thin ones — the games column tells you how seriously to take a row. The win rate tells you how good a brawler is. The trust score (0–100) tells you how much you can believe it — it depends purely on the number of games. Sorting uses both together: 52% from 75 games can be luck, 51% from 3,000 games cannot. Die Pfeile zeigen die Verschiebung gegenüber dem Stand von gestern, hier gerechnet nur für diese Map.