Guides · Marvel Rivals data

What a winning game looks like for each role

After a loss it's tempting to check the scoreboard and blame whoever had the lowest damage. But how different are the numbers in a win and a loss? We compared typical stats for each role across more than 32,000 Competitive player games recorded by Lunar. The biggest difference isn't damage or healing. It's deaths.

How we measured it

We used the full scoreboards of 2,832 Competitive matches (six against six, no leavers, no AI players, December 2025 to November 2026). Each player was counted under the role of the hero they played the longest, and only players with at least five minutes on that hero were included so very short games don't skew the rates.

Every number is per 10 minutes of play, so long and short matches can be compared. We show the median, the value in the middle: half of players were above it and half below. Medians aren't thrown off by one player's huge game the way averages are.

Vanguards

Per 10 minutes	In wins	In losses	Difference
Hero damage dealt	15,462	14,264	+8%
Damage taken	26,722	27,221	about the same
Deaths	4.4	7.2	39% fewer
KOs	20.9	12.7	+65%
Assists	4.2	2.6	+62%

Based on 5,152 winning and 5,052 losing Vanguard games. Vanguards took almost exactly the same amount of damage in wins and losses, which makes sense: absorbing damage is the job either way. The difference is that in wins they took it and survived. A winning Vanguard died about four times per 10 minutes, a losing one about seven.

Duelists

Per 10 minutes	In wins	In losses	Difference
Hero damage dealt	16,700	15,229	+10%
Damage taken	8,910	9,903	10% less
Deaths	5.3	7.9	33% fewer
KOs	23.0	14.3	+61%
Assists	3.0	1.8	+67%

Based on 5,479 winning and 5,451 losing Duelist games. Winning Duelists did only about 10% more damage than losing ones. If you lost a game with 15,000 damage per 10 minutes, you were dealing almost as much as the typical winner. Damage on its own barely separates the two.

Strategists

Per 10 minutes	In wins	In losses	Difference
Healing	22,355	20,905	+7%
Hero damage dealt	7,944	7,862	about the same
Damage taken	8,142	9,678	16% less
Deaths	4.1	6.8	40% fewer
KOs	15.9	9.5	+67%
Assists	24.8	14.4	+72%

Based on 5,819 winning and 5,940 losing Strategist games. Healing went up only 7% in wins. Strategists heal a lot in both cases, because in a losing game there's plenty of damage to heal. What changed was that winning Strategists took less damage themselves and died far less.

The pattern across every role

Put the three tables side by side and the same thing shows up each time:

Damage and healing barely move. Winners did 7 to 10% more of their main job, which is small next to everything else.
Deaths drop by a third or more. Winning players died between a third and 40% less often per 10 minutes, depending on the role.
KOs and assists jump. Winning teams got about 60 to 70% more KOs and assists. That partly follows from the deaths: a team that isn't respawning has more players in each fight.

There's a chicken-and-egg problem here. Winning teams die less partly because they're winning: they hold the objective and the other team has to come to them. So this doesn't prove that dying less wins games. But it does show that the stat people argue about most, damage, is one of the weakest signs of who won.

What to take from this

Judge a game by deaths before damage. If you died eight times in ten minutes, that's a bigger warning sign than a low damage number.
Don't blame the lowest damage on the board. The difference between a typical winning and losing Duelist was about 1,500 damage per 10 minutes.
For Strategists, staying alive is healing. A dead Strategist heals nobody, and winning Strategists took noticeably less damage.
Compare against yourself. These are medians across thousands of players and many heroes. Lunar's tilt check compares each game with your own usual numbers on the same hero, including deaths per 10 minutes, which is a fairer yardstick.

Limits of this data

Stats vary a lot between heroes in the same role, so a role median hides big differences between, say, a dive Duelist and a long-range one. KOs are counted as the in-game scoreboard counts them. The matches come from Lunar's users and their lobbies, across many patches. These are aggregates of real matches, not official statistics from the game's developer.