Marvel Rivals Season 9.5 Ban Guide: Who Gets Banned, and Who You Should Ban

The ban phase is the one moment you get to remove a problem instead of solving it, and most players spend it banning whoever beat them last game. Here is what everybody else is banning, and the more useful question underneath it.

The most banned heroes

Ban share is the percentage of every ban cast that named that hero, across the ranked divisions where the ban phase runs. It is counted from 2,926,875 bans.

Hero	Ban share	Win rate	Tier
Black Cat	13.36%	50.9%	A
Black Panther	7.90%	52%	B
Elsa Bloodstone	7.87%	49%	C
Peni Parker	6.71%	55.7%	S
Gambit	6.59%	49.1%	C
The Thing	6.48%	52.1%	A
Cloak & Dagger	6.08%	49.2%	B
Spider-Man	5.16%	51.3%	B

These win rates are shrunk toward 50% in proportion to sample size at the source, so a rarely played hero reads closer to even than it actually played.

Ban share is not a ban rate, and the difference matters

A ban rate would tell you how often a hero is banned in a given match. Ours is a share of all bans cast, which is a different denominator and a different number. We are careful about the wording because the two get used interchangeably and they should not be.

The reason this column is worth having at all is that bans are the cleanest signal on this site. They happen before the match, deliberately, with no information about who is losing. Almost every other behavioural number in this game is contaminated by people reacting to a game already going badly.

Why the popular ban is often the wrong ban

Look at the table again. The heroes at the top are banned because they are strong in general, which is a reasonable thing to care about and a bad way to pick your ban. What you actually want is the hero who is strong and hard for your comp to answer.

That is why every hero page ranks bans by counter strength against that hero combined with how often they are banned, rather than by popularity alone. On all 53 heroes, those two metrics name a different top ban. If they agreed, the panel would not be worth showing.

One practical note: the combined ranking damps ban share before combining it, because the spread across the roster is roughly 240 times from top to bottom while pick rate spans under nine. Undamped, a mediocre counter that everybody bans would outrank a genuine hard counter that nobody does.