106 Team-Ups: When Marvel Rivals Turns Draft Into a Game of Go With No Endpoint

There's a moment I still remember vividly, and it had nothing to do with a teamfight or a pentakill.

It was the evening I sat in my small apartment in Seoul, playback of the 2026 LCK Summer Finals on the left screen — the moment Score stole Baron, something I rewrote seven times. On the right, I was opening a blank spreadsheet, trying to sketch out the 106 Team-Up connections in Marvel Rivals. I got about twenty lines in before I stopped. My hand was tired. And a question surfaced in my mind that I couldn't answer right away: if a game has more than a hundred resonance relationships between characters, who is the player actually being skilled?

The most skilled player, or the player who remembers the most?

An empty stadium is never empty, if we know how to listen. And in this case, that "stadium" is a game design system. It is telling us something very large about the future of esports.

Context: A New Season, A New Web

Marvel Rivals is a 6v6 hero shooter developed and operated by NetEase, positioned as a direct competitor to Overwatch. Its core differentiator — and the very thing that makes it a worthy subject of analysis — is the Team-Up system.

The mechanic works like this: every character in the game has Team-Up abilities. Every character has exactly two Team-Ups. When you play a character, you always have the base effect of that Team-Up. But when your teammate picks the correct partner character, the enhanced effect activates. In other words, half the power of a Team-Up is free, and half is conditional.

In Season 10, NetEase introduced The Hood along with that character's Team-Ups. This is a content update at the mechanic layer — not a nerf patch, not a systems restructure, but an addition to an already-large resonance web.

The current number is 106 Team-Ups.

And the cadence for new hero releases is roughly one per month.

Let me place those two numbers side by side, because I think we often read past them without seeing what's actually happening. 106 Team-Ups, two per hero, means the system currently has around 53 characters. Each month, when a new hero arrives, it brings at least two new Team-Ups. But more importantly: according to the developer's own statement, new heroes will team up with old ones. That means every release not only adds two new edges to the graph — it can also retroactively raise the power ceiling of characters that have existed for years.

Imagine a bridge. Each month you build two more branches, but those branches connect back to pillars that have stood there for three years. That's the problem NetEase is setting for itself.

Core Analysis: The Combinatorial Balance Problem

I want to start with a principle I learned from years of following competitive games with team-synergy systems: the strength of a system is not in the number of its elements, but in the number of relationships between them.

With a list of 53 independent characters, you have 53 variables to balance. That's a linear problem. Hard, but manageable.

With a system of 106 Team-Ups, you have 106 pair-relationships to balance. But that's not even the real number. Because in a 6v6 match, nobody plays just one pair. You have two players, each carrying the influence of their own Team-Up, and those two Team-Ups interact with the four remaining characters. What you're balancing isn't 106 pairs — it's the combination of pairs across different roster contexts.

This is what I call the combinatorial balance burden. The number of relationships doesn't grow linearly, but in a far more unpredictable way.

Every match is a chapter, and I'm just turning the page. But this page is thicker than I thought.

Look at how this system reshapes the core tactical question. In a conventional hero shooter, the draft question is: "Which hero is strongest in this patch?" That's a ranking question — you list heroes from strongest to weakest, and you pick from the top down.

In Marvel Rivals, the draft question becomes: "Which synergy web is strongest in this patch, on this map, against this enemy composition?" That's no longer a ranking question. That's a graph question.

This distinction isn't academic. It completely changes how players must prepare, how coaches must plan, and how analysts must read matches.

In a hero-ranking system, you can prepare by memorizing a tier list. Thirty rows. Updated each patch.

In a synergy-graph system, you must prepare by memorizing a web. With 53 characters, each with two Team-Ups, you're talking about 106 directed relationships, and each relationship has at least two states (base effect and enhanced effect). Multiply it out, and you have over two hundred states to track.

I used to print match stat sheets and underline the invisible numbers. I did that because I believed meaning lives in the places the data table doesn't say out loud. But with 106 Team-Ups, even printing it out becomes a meaningless act — the sheet would be longer than my desk.

That's when I realized the most important thing about this system: encyclopedic knowledge has become a competitive asset in the literal sense.

Quoting the original guide itself: the author recommends players bookmark the Team-Up list for frequent reference. It's a small detail with large symbolic weight. When a guide says "bookmark this list," the author is admitting one thing: nobody can memorize the entire system reflexively. You need an external reference document.

In football, nobody needs to look anything up to know that a winger cutting inside creates space for an overlapping fullback. That's foundational knowledge, absorbed through years of observation. In Marvel Rivals, the boundary between "foundational knowledge" and "knowledge requiring lookup" is being pushed back continuously.

Who Benefits, Who Suffers

The Team-Up system favors a specific type of player: one with a deep hero pool and flexibility in switching.

The logic is simple. If you only play one character well, you're limited to that character's two Team-Ups. If the opponent bans or picks away your partner character, your enhanced effect disappears — and you're left with only the base effect. You become a weaker version of yourself.

Conversely, if you play five characters across three different roles well, you can pick the Team-Up that fits your teammates' composition, instead of forcing teammates to adapt to you.

This is a scouting principle that I think any esports organization building a Marvel Rivals roster should carve into the wall: a player's value is not in their peak on one character, but in the number of peaks they can reach across different characters.

Some victories you have to read three times before you see the tears. And some defeats you have to read three times before you see the structure behind them. A roster of superstars each good at one character will lose to a humbler roster that can weave a dense resonance web.

This is a point where I want to pause, because it has deeper implications for competitive culture.

For years, esports has built its mythology around individuals. Faker's Zed vs Zed. Son Heung-min's dribbles from the left wing. Those moments are beautiful, and we need them. But Marvel Rivals' Team-Up system is a reminder that some games are designed to be unsolvable by an individual, no matter how talented.

The quiet hero — the one I always search for in every match — in Marvel Rivals is not the support player silently protecting teammates. The quiet hero here is the second player in a Team-Up. The one whose very existence activates another's full potential. The one whose name doesn't appear on the stat sheet, but whose absence would collapse their teammate's numbers.

The Contrarian Angle: When "Everyone Gets a Team-Up" Is a Dangerous Promise

This is the part where I want to challenge myself.

The developer has committed that no character will be released without a Team-Up. I read this commitment the first time and thought: good, it protects the value of old characters. A new character won't make old characters entirely obsolete, because the new character will team up with old ones.

But look again.

Such a commitment means every new character release expands the web by at least two edges. And because new characters team up with old ones, it also means every update can potentially change the value of existing Team-Ups.

This creates a phenomenon I call the "permanent adjustment period."

Think of it this way. In a game with a slow update cadence, the meta has time to be "solved" — players find the optimal composition, the community reaches consensus on it, and there's a period of stability before the next patch upends everything. That stability window is time for pro teams to prepare tactics, build chemistry, and practice.

With a monthly hero release cadence, that window is compressed. The meta is never truly "solved" before being disrupted. And in a system where every new character adds edges to the graph, the disruption doesn't just come from the new character — it comes from that character's interaction with the other 105 Team-Ups.

This is where I have to be careful with myself. Because I have no win-rate, pick-rate, or ban-rate data. The original guide gives me an inventory — a list of what exists — but no performance analysis. All my judgments about meta direction are therefore structural judgments, not data-backed ones.

I say this not to defend myself, but to draw a boundary around what I can and cannot claim.

What I can claim is structure. And structure says: a system with 106 Team-Ups growing monthly is creating a balance surface that expands faster than any balance team can tune.

What I cannot claim is which specific Team-Up is overpowered. Without that data in hand, any claim about a "broken duo" would be pure speculation.

And here's the truly counterintuitive point: perhaps the problem isn't that some Team-Up is broken. The problem is that the system is designed to always have a broken Team-Up, because its scale exceeds the developer's own capacity to control.

I've seen something similar in football. When youth academies chase short-term results and physicalize U18 players, they don't intentionally destroy the technical foundation. They're just reacting to immediate pressure, without seeing the compounding consequences. Those consequences only become clear ten years later, when you look back and realize a whole generation of creative players was lost.

Marvel Rivals may be at the early stage of a similar trajectory. Not because the developer lacks capability, but because the problem they've set for themselves grows exponentially in complexity, while balancing capacity grows arithmetically.

What the Number 106 Is Really Telling Us

I don't predict outcomes, I only read the story as it's being written. And the story here is a story about scale.

When a game crosses the threshold where individual knowledge no longer suffices, it forces the development of social structures to compensate. That's why we have coaches, analysts, and support staff in professional esports. Not because players are less intelligent than previous generations, but because games have become too complex for a single brain to process in real time.

Marvel Rivals, with 106 Team-Ups and a monthly cadence, is pushing that threshold higher still. And that means: if this title builds a professional competitive ecosystem, teams will need dedicated analytics staff not as a luxury, but as a condition of survival.

This is where I think about what I've learned from watching my own matches over years. I remember a match I observed in the LCK, where a team lost not because any individual played badly, but because they entered draft with a mental model of the meta that was two weeks out of date. The patch had changed, but their way of thinking about the patch had not.

With Marvel Rivals' monthly cadence, the gap between "actual meta" and "the meta your team believes is actual" will be a competitive weapon. The team that can update its mental model faster gains the edge. And that's not a mechanical edge — not someone playing better. It's a cognitive edge.

I used to think an analyst's job was to find the truth behind the numbers. Now I think an analyst's job is to find the numbers behind the truth. The truth here is a system growing faster than the community's ability to understand it. The number here is 106 — and it will be 108 next month, 110 the month after, and so on.

The Silence Between Connections

I return to my blank spreadsheet, the one I abandoned after twenty lines.

I don't finish it. But I understand why.

The silence in an esports match is the moment of waiting to respawn, the moment of moving between two objectives, the moment a player makes a decision nobody sees on camera. That's when true character shows, because nothing conceals it.

The silence in a Team-Up system is the same. It lives in the moment between picking your character and hitting the start button. That's the moment you must decide: what will I rely on to win? My individual skill on a familiar character, or my adaptability to a web I've never fully grasped?

The answer to that question isn't in the guide. It's in how each player defines their own worth.

And that, perhaps, is what 106 Team-Ups is really testing — not our memory, but our identity as players.

The value of a contract isn't in the number, but in the story it opens. And the value of a synergy system isn't in the number of Team-Ups, but in the question it poses about what we consider skill.

From the empty stands, I learned to write for myself first. And in this case, I write to remind myself: the most complex systems are often not defeated by the strongest, but by the one who understands them best.

The question I carry out of this piece isn't "which Team-Up is strongest."

The question is: when a game reaches a complexity that no individual can fully grasp, are we witnessing the end of the individual myth in esports, or merely its transformation into a new form we don't yet have words for?

I don't have the answer. But I'll keep turning the page.