Marvel Rivals: 106 Team-Ups and the Balance Problem the Developer Cannot Win

The day The Hood entered Marvel Rivals in Season 10, I was sitting in a PC cafe in Hongdae, Seoul, watching two Korean students open the hero select screen and argue for forty seconds before locking in. They were not arguing about who was stronger. They were arguing about who should pair with whom, which duo matched a Team-Up, and whether it was worth sacrificing a comfort pick to gain the enhanced effect. At second forty-one, both locked in. The match began in a state where neither of them was truly sure they had chosen correctly.

When I cut that footage back at the studio, what struck me was not a teamfight. It was the silence before the match began. Marvel Rivals has changed the underlying question of the hero-shooter genre: from "which hero is strongest" to "which pairing is strongest inside this web." When the underlying question changes, everything behind it has to change too — balance, drafting, coaching, and the way a professional team builds a composition.

Marvel Rivals is a six-versus-six hero shooter, and Team-Up is its primary differentiator against the rest of the genre. The mechanic runs on two layers. The first is a base effect, always available even when you lock your hero alone. The second is an enhanced effect, which only activates when a teammate locks the correct partner hero. Season 10 brought The Hood into the game, along with that character's new Team-Up pairings.

The compilation I read offers three quantitative facts, and I record them verbatim: the game currently contains 106 Team-Ups; every hero holds exactly two Team-Ups; new heroes arrive at a cadence of roughly one per month. The compilation also states that no character is ever released without a Team-Up, and specifies that new heroes will pair with old ones rather than living in a separate ecosystem. At the end, the author recommends bookmarking the page for regular reference.

Those three facts are the entire quantitative foundation of this story. No win rate. No pick rate. No ban rate. Not a single line of performance data. Technically speaking, this is an inventory: it answers the question "which pairings exist," not the question "which pairing is distorting the match." I say this before discussing the number 106, because an inventory has no right to be read as an analysis — even when the person who wrote it believes otherwise.

Balance in a traditional hero shooter is a problem over points. Team-Up turns it into a problem over edges. In a point problem, each hero is an independent node, and designers can raise or lower that node's power without touching the rest of the scale. In an edge problem, a hero's value no longer lives in the hero itself but in the connections running from it to other heroes. Adjusting one edge can move the value of two heroes, three compositions, and an entire downstream draft sequence. With 106 edges in play and two edges per hero, the number of variables a balance team must manage is many times larger than the number of characters.

That is before accounting for speed. Every new hero adds at least two edges to the graph, and the new hero itself also needs a place on the scale. At roughly one character per month, the graph grows by about twenty-four edges per year, plus the two edges belonging to the new hero. No design team can thoroughly test every edge at that cadence. What they can test are the most prominent edges; the rest survive on luck and on player tolerance.

The base layer and the enhanced layer form a double-edged design. The fact that the base effect is always present — regardless of what your teammate picks — is an intelligent pressure valve: it keeps a hero from becoming useless when played solo, and it keeps players from being punished too harshly for an uncooperative teammate. But that same mechanic creates what I call conditional dependency: the reward vanishes when the pairing does not match, and in solo ranked, that reward vanishes far more often than it appears.

The second notable point is backward compatibility. The developer states plainly that new heroes will pair with old ones. This is a design commitment meant to protect the value of existing heroes and avoid the obsolescence that abandons an entire old roster every season. But it has a flip side: each new hero can inadvertently raise the power ceiling of an old hero through a new Team-Up edge, and that increase happens without touching the old hero's base stats. The designer changes an edge, but players feel the change on an entirely different hero.

In Seoul, strategy is always written after the match is already over — and Team-Up is how Marvel Rivals writes it for you in advance. I once sat listening through the headsets of a Korean amateur team during scrims, and what they discussed was not individual skill. They discussed lock-in order. Who locks first to open a path, who locks later to secure the pair, who has to take the loss because their pool lacks the partner hero. Hero selection has turned from a matter of preference into a negotiation over the distribution of advantage.

And once knowledge becomes a competitive asset, the entry barrier rises with it. 106 pairings is a number far beyond the reflexive memory capacity of an ordinary player. The author of the compilation has to tell readers to bookmark the page, and that recommendation is itself a confession: the system can no longer explain itself in-game. New players walk into a web they cannot see, while veterans and coached players walk in holding a map.

A monthly hero cadence produces what I call a permanent adjustment period. A meta is only considered solved when the community agrees on which options are strongest, and that agreement takes time. When a new hero appears every month and each appearance adds two edges to the graph, the window in which a meta can be solved becomes shorter than the window required to solve it. The result is a community in permanent argument, professional teams in permanent re-learning, and nobody with enough time to prove conclusively that pairing A beats pairing B.

What stops me from concluding decisively is the total absence of performance data. I have watched tournaments and scrims in this genre long enough to know that the felt strength of a pairing often diverges from its actual win rate. A pairing can be branded overpowered simply because it is easy to play and easy to produce flashy fights, while a pairing that controls vision and holds map tempo posts a higher win rate that nobody notices. With 106 edges and not a single data column, any judgment about which pairing is breaking the match is structural, not statistical.

Now comes the part where I have to audit myself. The strongest argument against me is this: what I am calling an escalating balance burden may in fact be the product itself, not a defect. A system with 106 edges ensures that veteran players never truly leave their old heroes — they have a reason to return every time a new hero pairs with a hero they used to play. Read that way, Team-Up is a well-engineered retention mechanic, and complexity is the game's moat against a market full of clones.

I also have to admit a weakness in how I gathered my data. This entire piece rests on a community-compiled summary, not on official patch notes. If the number 106 is wrong, or if the description of the base and enhanced layers does not fully match the in-game mechanic, then most of the argument above still holds structurally but loses precision on detail. I chose to write on because the structure is worth discussing, but I mark it clearly: this fact has not been independently verified.

The whole world chants about combination play, while I only see a hero select screen locking itself out. That may be the pessimistic view of someone sitting at the edge observing the Korean market rather than playing it every day. A solo ranked player might tell me he never thinks about 106 pairings, he just picks the hero he likes and hits the queue button. If most players really are like that, then every concern about the entry barrier I just constructed belongs to a very thin slice of the player base.

The Hood will not break the meta. It will only make the maze two tiles longer. And those added tiles are the thing worth tracking, not the new hero itself. My verifiable prediction: within the next twelve months, Marvel Rivals will have to adjust or temporarily disable at least one Team-Up mid-season, and that adjustment will not come with any major stat change to the heroes involved — because the problem lives on the edge, not on the node.

So when a game reaches the threshold where understanding it is harder than playing it, does that knowledge map still serve the players, or is it only serving itself?