Wins early — in Heist 4.3 points better in short rounds than in long ones (49.9% against 45.6%, split at 83 s). A pick for pace — if the game drags on, it works against him.
Prochange
Win rate per rank tier, across all maps — ranked battles only, the tier comes from the battle itself. The axis runs from Silver to Masters, one tick per tier: the stretch where this brawler has at least 300 battles per tier; 2 tiers fall below that and drop out. The stretch always fills the full width, three tiers or eight — so when comparing two curves, mind the names on the axis. The curve glows brighter where more was measured — the tiers differ by more than a thousandfold in battles. The faint horizontal line sits at 50 %.
No clear rank trend — from Silver to Masters the win rate stays practically the same (regression −1.4 points across the whole range). He plays the same at the bottom as at the top.
Where this brawler works (ranked, Pro).
Direct duels in Heist: how Nani does when that brawler is on the enemy team. Always one mode, never all of them at once — the same pairing can mean the opposite in two modes. Only listed where a real counter is behind it: whoever only wins because he is the stronger brawler is not shown here. Computed across all ranked tiers together — not for your tier alone.
Worst case: 29.6% against Starr Nova in Gem Grab. On average the worst case sits at 27.5%.
The word after the number says how hard: hard counter means the edge is clearly bigger than the strength of the two would give. A high win rate alone never appears here — whoever only wins because he is the stronger brawler does not make the list at all. See it in the counter network
On the same team, in Heist — which combination wins. The mode is the one chosen above. Computed across all ranked tiers together — not for your tier alone.
Nori
60.9%
940 games
Mico
55.9%
608 games
Edgar
53.6%
1,214 games
Chuck
53.9%
510 games
8-Bit
52.1%
864 games
Melodie
53.4%
268 games
Players with a profile and at least 20 games on Nani, sorted by the Wilson lower bound. To the full list