Back to news

Marvel Rivals Season 9.5 Tier List, and What the Letter Actually Means

Who is winning this half season, where the letter comes from, and the thing a tier list can never answer.

Aug 22, 2026 · figures from the Aug 31, 2026 snapshot

Tier listSeason 9.5

You want the short version: if you are picking blind and want the best odds this half season, the heroes below are where the numbers point. The longer version is that the tier letter is a consensus, not a measurement, and it is worth ninety seconds to know what went into it.

The S tier right now

4 heroes sit at the top of our consensus. Every figure in this table is read from the current snapshot, so it moves when the data moves.

HeroRoleWin ratePick rate
MantisStrategist57.1%22.9%
Peni ParkerVanguard55.7%17.9%
UltronStrategist54%12.7%
The HoodVanguard53.3%33%

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.

Where the letter comes from

The tier is a Borda count over 5 ordered inputs, and being honest about what they are matters more than the number of them: the stats snapshot (53 heroes), the official feed (52 heroes), the overlay stats (53 heroes), the profile aggregate (53 heroes) and the community ranking (53 heroes). Four of those measure matches. One is a named person's ranking, which is admitted because it can be attributed rather than because it tracks outcomes.

Do not read that as four independent readings of reality. Two of them run their own game client, one is the game itself, and one samples public player profiles that the game serves. That is four measurements from three origins, and the reason we aggregate at all is that they genuinely disagree at the ranks competitive players care about.

A hero missing from one input is averaged over the inputs that do have them, never scored zero. That sounds like housekeeping and is not: score the gap and a newly released hero reads as the worst in the game, purely for being new.

The bottom, and why it is less useful than it looks

Cyclops, Phoenix, Jeff the Land Shark and Squirrel Girl close out the list. Before you bin them: a low tier says a hero wins less across everybody's games, not that they lose yours. Tier is a prior. The matchup in front of you is evidence, and evidence beats a prior.

Which is the honest limit of any tier list, including this one. It cannot tell you what to do about the specific hero currently ruining your game. That is what the counter tables are for, and it is the question we would rather you asked.

The full board, with every hero and the letter each source gave them, is on the tier list.

Statistical data is based on external APIs and publicly available information. It may differ from official data.