106 Team-Ups in Marvel Rivals: A Balance Map Growing Faster Than the Hands That Tune It

In a late-night ranked match on a European server, I watched a player lock The Hood the moment Season 10 went live. He did not ask his team what they wanted to run. He typed one line into chat: who can pick my partner hero. Nobody answered. Ten minutes later his ability bar was still sitting in its base state while the enemy team had two Team-Ups firing on both flanks. The match ended at the twelve-minute mark. It had been decided before the loading screen finished.

That was the moment I understood Marvel Rivals is no longer a game that asks which hero is strongest. It is a game that asks which web of heroes is strongest. And that web had just grown another link.

Marvel Rivals, NetEase's 6v6 hero shooter, entered Season 10 with The Hood. But the real shift was not the character himself, it was the Team-Ups that arrived with him. Every hero in the game owns two Team-Ups. Each Team-Up has two layers: a base effect that is always active, and an enhanced effect that only triggers when a teammate picks the designated partner hero. Add it all up and the game currently holds 106 pairings.

What matters more is that the studio has no intention of stopping. New heroes ship roughly once a month, and every new hero must have Team-Ups, even if that means pairing them with characters released several seasons earlier. This is not a patch. It is a publicly stated long-term design commitment.

For a ranked player, that changes the entire preparation process. You cannot simply be good at one hero. You have to be good at a matrix. And that matrix does not live in your head when you sit down at the hero-select screen, it lives in a browser tab you opened beforehand.

This is also the core difference between Marvel Rivals and the rest of the hero-shooter genre. Comparable titles build depth around individual skill and role. Marvel Rivals builds depth around the relationships between roles. The distinction sounds small. It rewrites the entire logic of pick and ban.

People call that the meta. I call it digitised fear, the fear that you locked the wrong pairing and will now pay for it with twelve minutes of helplessness.

I spent two weeks tracking high-ranked matches and logging how players made their draft decisions. What stood out was not which hero someone picked, but who they picked it with, and what they were willing to give up to secure that pairing.

The Team-Up system turns the basic tactical question into a graph problem. In a traditional hero shooter you ask: which hero is strongest on this patch? The answer is a point. In Marvel Rivals you must ask: which duo is strongest, does that duo still leave room for the other three players, and if the enemy bans half of that duo, what is the fallback? The answer is a network.

That produces an effect I call combinatorial burden. With twenty heroes, a studio has roughly forty pairing edges to test. With more than a hundred edges, each new character does not merely add two fresh links, it can also retroactively raise the power ceiling of older heroes. Every small balance change in this system can ripple across multiple compositions, so tuning stops being about adjusting a number and becomes about adjusting a web.

The two-layer structure deserves its own note. A base effect that always exists means half of a duo's power is free and half is conditional. On the surface this is humane design: it protects a hero's value when played solo and prevents old heroes from being entirely invalidated by new ones. But it also opens a significant gap between players who queue with a coordinated group and players who queue alone. In solo queue you cannot choose your partner. You can only hope.

From a scouting perspective, the system rewards two things. The first is hero-pool depth: players with wide pools always have more pairing edges, and therefore more fallback options when something is banned. The second is pre-game scripting: knowing which pair stacks with which, knowing the order in which effects should come online, knowing when to sacrifice an enhanced effect in exchange for a better matchup on a side lane.

I once watched a team in a European community tournament win three straight games without changing a single hero for the first two. They did not swap because their plan A already contained two overlapping Team-Up paths, and their plan B was a role swap rather than a hero swap. Their opponents lost not through a lack of mechanical skill, but because they were forced into a defensive posture at the draft stage: every time they locked a strong individual hero, they detached themselves from the network, and every time they did, the gap widened a little further.

Here I have to pour cold water on my own enthusiasm.

The story that coordination will decide everything sounds beautiful. It also sells very well. But not a single performance number has been published to prove it. No win rate. No pick rate. No ban rate. The figure of 106 pairings is an inventory number, not an effectiveness number. We know how many edges exist. We do not know which ones actually win. A catalogue is not an analysis, no matter how it is dressed up.

Second, the assumption that more options means more strategy is a dubious one. A system that forces fixed pairings can produce the opposite: it narrows the creative space. When the reward only arrives for pre-approved combinations, optimal players do not explore, they copy. The meta does not expand. It sets into a list of mandatory pairs, and everything outside that list becomes a wrong answer.

There is also a detail about how this information reaches us. The list of 106 pairings I read is a guide that is updated continuously, carrying a timestamp but no year. For a game that ships a hero every month, such a list can drift out of date within weeks. The phrase complete list is a promise nobody can verify, including the person who wrote it.

I am not saying the system is broken. I am saying we are reading a catalogue and calling it an analysis.

The match begins when the coaching staff submits the roster, not when the referee blows the whistle. In Marvel Rivals that line is now literal: the roster is a web of more than a hundred pairings, and it thickens every month.

The question I leave behind is not how to memorise 106 Team-Ups. It is this: how far can a system designed never to stand still be balanced, before its own richness becomes a barrier to entry for new players? If the answer is never, then what is being built here is not a meta. It is a race with no finish line.