The Match Was Balanced. So Why Wasn't It Fun?
On Deadlock, Call of Duty, the obsession with competitive balance, and the social experience that modern matchmaking forgot.
In this post
There was a period when I enjoyed Deadlock more than I do today, and it wasn't necessarily because the game was better balanced. I was playing at a relatively high rank, usually between 11 PM and four or five in the morning. The player population was much smaller then. Once you accounted for rank, region, and the ridiculous hours I was playing, the actual pool of potential teammates and opponents became remarkably small. And it was fantastic.
I recognized people. I remembered their nicknames, their preferred heroes, how they approached the game, and sometimes how unpleasant they were to play with. I knew who was aggressive, who liked farming, who could be relied upon to make a play, and who had a tendency to do something completely idiotic. They recognized me, too.
There was something resembling a community. We hadn't organized one. Nobody had invited us to a Discord server or assigned us to a competitive league. We simply kept meeting each other because we happened to be playing the same game, at the same level, at the same hours. Somewhere along the way, those repeated encounters created familiarity, expectations, rivalries, and an unspoken social contract.
The games weren't necessarily fairer, more competitive, or even strategically better than they are today. But they had something that I increasingly struggle to find. The matchmaking system had accidentally produced something far more interesting than a balanced match: a social environment. And I think modern competitive games are remarkably bad at recognizing the value of that.
The obsession with balance
Modern matchmaking seems increasingly preoccupied with a deceptively simple objective: assembling teams with approximately equal probabilities of winning. Give each player a skill rating, calculate the expected outcome, and try to produce a match as close to 50/50 as possible. Systems such as Glicko, developed by statistician Mark Glickman, and Microsoft's TrueSkill offer sophisticated ways of estimating competitive ability. Glicko introduces uncertainty and, in its later version, rating volatility. TrueSkill extends skill estimation to multiplayer competition.
These are useful tools. I don't dispute that. But a rating system estimates competitive ability. It doesn't tell us what makes a game enjoyable, what makes an opponent memorable, or why we sometimes spend thousands of hours pursuing a higher rank. And the distinction matters.
Consider ordinary sports. One team might be substantially stronger than another. Everybody knows it. The weaker team knows it, the spectators know it, and the bookmakers certainly know it. Yet those matches can be extraordinarily compelling. Sometimes the favorite dominates, as expected. Sometimes the underdog puts up a remarkable fight. Sometimes the favorite chokes. Sometimes something completely unexpected happens. The imbalance itself can be part of the story.
Of course, professional sports operate within leagues and tournaments that have their own structures. They're not directly comparable to anonymous ranked matchmaking. Nor am I suggesting that developers should routinely throw beginners against professional players.
My point is simpler. Competitive equality is one possible ingredient of an enjoyable match. It is not the objective of competition itself. A perfectly balanced game can be miserable. An uneven game can be memorable. And a system that cannot distinguish between those outcomes is missing something fundamental.
Call of Duty and the disappearing hierarchy
One of the most persistent complaints about modern Call of Duty concerns skill-based matchmaking, or SBMM. Players describe public lobbies as exhausting, repetitive, and relentlessly competitive. The complaint is usually summarized as something like: "I don't want to sweat every single game."
The obvious counterargument is that these players simply want to dominate weaker opponents. They enjoyed being the strongest player in the lobby, and now they're upset that the game makes them compete against their equals. There's probably some truth to that. But I think the explanation is incomplete.
Consider someone who spends 2,000 hours becoming exceptionally good at Call of Duty. They've developed their aim, movement, positioning, map knowledge, and mechanical consistency. They're substantially better than the average player.
In an environment with a broad distribution of skill, that improvement is visible. They frequently outperform other players. People notice. They might admire them, complain about them, accuse them of cheating, or challenge them to another match. Whatever the reaction, their achievement has social consequences.
Now imagine that matchmaking responds to every improvement by placing them against progressively stronger opponents. Their absolute skill increases, but their relative performance remains approximately the same. They were an average player in their lobbies when they started, and after 2,000 hours of practice, they're still an average player in their lobbies. Their improvement hasn't disappeared. The matchmaking system has recognized it perfectly well. But it's made that improvement considerably less visible.
The system acknowledges your progress by ensuring that you rarely get to experience its consequences.
This isn't merely my interpretation of the controversy. In a March 2024 Reddit discussion about Call of Duty matchmaking, players explicitly questioned what incentive remained to improve if every improvement simply resulted in more demanding opposition. Others asked why their competitive achievements were effectively invisible without a meaningful indication of the skill level of their opponents. A 2019 Modern Warfare discussion made a related argument: if matchmaking was going to create skill-based lobbies, players at least wanted to see the skill level of the people they were competing against.
And that's revealing. Because the problem isn't necessarily that players are losing. It's that they no longer understand what their performance means relative to the wider competitive population.
But I think there's an even deeper explanation. Skill isn't merely a competitive attribute. It's a form of social capital. And social capital requires an audience that recognizes it.
In traditional competitive environments, people develop reputations. Some become known as exceptional players. Others become respected specialists, reliable teammates, notorious troublemakers, or familiar rivals. A hierarchy emerges, not necessarily through any formal ranking system, but through repeated observation and interaction.
Modern matchmaking can interfere with that hierarchy in two ways. First, tightly matching skill can make differences between players less apparent within individual games. Second, constantly assembling new lobbies can prevent reputations from developing in the first place. The first makes it harder to demonstrate superiority. The second makes it harder for anyone to remember it.
And perhaps what people miss isn't simply easier matches. Perhaps they miss the social environment in which becoming good at something actually meant something to the people around them.
What Call of Duty players were really losing
The controversy surrounding Call of Duty matchmaking becomes considerably more interesting when we look beyond complaints about difficult opponents. In an October 2024 discussion on the Black Ops 6 subreddit, players complained about disbanding lobbies. They remembered being able to stay with the same group of strangers for hours, developing rivalries, learning each other's playstyles, and occasionally making friends. A July 2024 Call of Duty discussion expressed similar frustration. Players missed encountering familiar opponents, comparing themselves against them, and building relationships over multiple matches.
This is something entirely different from wanting easy opponents. It's a complaint about losing continuity.
Imagine you're playing against someone who keeps beating you. Over several games, you begin recognizing their habits. You learn how they approach engagements, which positions they favor, and what mistakes they tend to make. Eventually, you adapt and beat them. That victory has meaning because of everything that came before it.
Now imagine that the lobby automatically dissolves after every match. You lose, the game ends, and the players disappear. The matchmaker finds another collection of strangers, and everything starts over. You might receive an equally competitive match, but you've lost the opportunity to develop a relationship with your opponent.
There's no history to build upon. No reputation to challenge. No familiar rival to finally defeat. The matchmaker has produced another statistically satisfactory game while destroying the conditions that could have made the previous one memorable.
Interestingly, Call of Duty's developers have themselves experimented with reversing this trend. For Black Ops 7 in 2025, Treyarch announced that multiplayer would launch with substantially looser skill considerations and persistent lobbies. The studio explicitly recognized the community's complaints about lobby disbanding.
The debate wasn't settled. In August 2026, the Modern Warfare 4 beta's matchmaking plans revived criticism of stricter skill matching and disbanding lobbies. Activision defended its approach using player retention, quit rates, and even player-reported match enjoyment, while critics argued that persistent lobbies provided social value that those measurements did not fully capture.
That back-and-forth is instructive. These aren't simply arguments about whether the next match should be easier or harder. They're arguments about what kind of social environment multiplayer games should create. A lobby is not necessarily a temporary collection of players. Given the opportunity, it can become a community. And perhaps developers have been too quick to dissolve those communities in pursuit of the next mathematically optimal match.
What I actually wanted from high-ranked Deadlock
This brings me back to my own experience. I've spent an absurd number of hours playing competitive games. I'm in my thirties now, and I don't necessarily want to exercise every ounce of my mechanical ability every evening. Sometimes I want to concentrate intensely. Sometimes I want to relax. Sometimes I want to play something mechanically simple while still making strategically interesting decisions.
I also have a tendency to gravitate toward characters that impose ridiculous problems on other people. That's one reason I enjoy Graves in Deadlock. I can create enormous inconvenience for my opponents with relatively little mechanical effort. I send my zombies down a lane, and suddenly somebody else has a problem to solve.
They wanted to play a shooter. I've made them play Plants vs. Zombies. I find that hilarious.
I enjoyed Axe in Dota 2 for similar reasons. Blink, taunt, and for a brief period the other player no longer gets to decide what their character is doing.
One of my favorite examples is Wascot from Super Monday Night Combat, an absurdly entertaining character in a game that combined third-person shooting with MOBA-like objectives. Wascot could launch fake coins resembling the money players collected during matches. Try picking up the wrong coin, and you were punished for your greed. He could grapple opponents, leap toward them using his Crook Hook, and counter their attempts to grapple him with Shifty Shuffle. He even had an ability called Party Pooper that interfered with an opponent's offensive effectiveness. The entire character was a practical joke embedded in a competitive game. And it was magnificent.
These characters aren't necessarily connected by mechanical difficulty or raw competitive strength. They're connected by something much more interesting: the ability to impose an absurd, asymmetric problem on another player. I don't necessarily want to outshoot you. I want to make your ability to outshoot me irrelevant.
Now, imagine repeatedly encountering the same players while using these characters. Perhaps someone begins recognizing me as that annoying Graves player who constantly pushes lanes. They learn my habits and develop strategies specifically to counter me. Perhaps we develop a rivalry. Suddenly, the game isn't just about twelve strangers trying to destroy each other's bases. It's about familiar personalities pursuing conflicting objectives.
Maybe somebody sees my nickname and immediately knows what kind of bullshit they're going to have to deal with. Maybe somebody else remembers how I played last time and develops an elaborate counterstrategy. And perhaps, after several matches, they finally get their revenge.
That's what I find so compelling about repeated encounters. Our preferences, personalities, and reputations become part of the competitive environment itself. It's not merely the mechanics of the game that create depth. It's the accumulated knowledge we develop about the people playing it.
And that's precisely what I enjoyed about high-ranked Deadlock. I wasn't necessarily pursuing a higher rank because I wanted a more prestigious badge or because I believed everybody beneath me was unworthy of playing alongside me. I wanted to enter a smaller, more stable population of players. I wanted games where people recognized one another, understood one another's tendencies, and developed expectations about how the match might unfold.
The rank was valuable partly because it gave me access to that social environment. That's something a numerical rating cannot adequately describe.
Competition needs an audience
There's a broader principle here that extends far beyond video games. People rarely pursue mastery exclusively for the internal satisfaction of becoming better. Recognition, status, belonging, rivalry, and social comparison often play important roles. We want our achievements to mean something.
A talented musician wants people to appreciate their performance. An athlete wants to compete against respected opponents. A chess player might enjoy defeating a familiar rival more than defeating an anonymous opponent with a slightly higher rating. The quality of the achievement and the social meaning attached to it are not the same thing.
In competitive games, the social dimension is particularly important because other players are simultaneously your opponents, teammates, spectators, and judges. They experience your skill directly. They recognize your habits. They remember your victories and failures. And over time, they develop opinions about you. This produces something that an anonymous ranking system struggles to replicate: an informal social hierarchy grounded in shared experience.
That hierarchy isn't always pleasant. It can produce arrogance, exclusion, hostility, and the sort of insufferable egos that make ranked matchmaking miserable. I've experienced that in Deadlock, and I won't pretend I've never contributed to the problem myself. There's also an obvious conflict: being recognized as superior shouldn't require permanently subjecting weaker players to hopeless matches. A social hierarchy can become destructive if the same people are always expected to lose for somebody else's enjoyment.
But this doesn't mean hierarchy itself is undesirable. Recognition from a familiar community of strong players may be more satisfying than repeatedly dominating anonymous weaker opponents.
And that recognition requires continuity. If every match consists of twelve strangers who will never meet again, there is little opportunity to establish a lasting reputation. You become a rating, a nickname, and a collection of statistics. Then the match ends, and you're forgotten.
In a way, players are already trying to create that continuity themselves. They "mark" people by muting them, sometimes even when they're on the enemy team, so that a future encounter comes with a little history attached: "I've met this player before. He wasn't to my taste." The mute isn't just about silencing someone in the current match. It becomes a crude memory aid, a way of carrying a judgment about another player into the next encounter.
Could matchmaking understand people instead of just rating them?
This raises a provocative possibility. What if matchmaking could understand the psychological and social dimensions of competition rather than merely estimating the probability of winning?
I don't mean asking players what kind of games they would prefer. Questionnaires are an unreliable foundation for something like this. People don't necessarily know what they enjoy. They rationalize their choices, misremember their experiences, and sometimes tell you what they believe they ought to want rather than what actually entertains them. And once people understand that particular answers produce desirable results, they'll begin optimizing their answers accordingly.
Instead, consider what we could learn from naturally occurring behavior. People discuss games constantly. They complain about matchmaking on Reddit, discuss competitive strategies on public Discord servers, argue about hero balance, celebrate memorable victories, and explain what makes them frustrated. They also produce enormous amounts of in-game data. Now imagine an AI-assisted research system capable of connecting those two worlds.
A player publicly complains that matchmaking has become miserable. Their match history reveals that their win rate has actually improved. Perhaps the problem isn't losing. Perhaps their matches have become longer, their preferred strategies have become less viable, or they've stopped encountering the familiar people who previously made the game enjoyable. The statistical data alone doesn't explain their dissatisfaction. Their public comments provide context.
Likewise, someone who complains constantly about oppressive heroes might actually derive enormous satisfaction from learning to counter them. A behavioral model that combines gameplay with broader evidence could potentially recognize that contradiction. This resembles open-source intelligence, or biały wywiad in Polish: gathering and analyzing information from accessible sources to understand relationships that aren't apparent from individual observations. Modern AI makes the scale of such analysis genuinely interesting.
But there's an important distinction between researching these patterns and secretly constructing unrestricted psychological dossiers on every player. Public posts are still personal data, identities can be linked incorrectly, and automatically analyzing public conversations creates serious questions about consent, privacy, and how the resulting inferences might be used. Nor would a comprehensive profile necessarily reveal what someone enjoys. Human preferences are contextual, contradictory, and sometimes difficult even for the individual to understand.
Still, the underlying proposition is worth investigating. What if we could identify not just which players have similar abilities, but which kinds of social and competitive interactions are meaningful to them?
Perhaps matchmaking could learn when repeated encounters are productive, when familiar rivals create enjoyable competition, and when certain combinations of players predictably produce hostility. It might recognize that some people enjoy the challenge of defeating oppressive strategies while others find those interactions tedious. It might identify productive teammate relationships worth preserving, rather than treating every player as an interchangeable numerical quantity.
Such a system would need to be tested against meaningful outcomes rather than simply trusted because it possessed more information. It would also need strong protections against harassment, discriminatory profiling, and the manipulation of individual players. But it suggests a different philosophy. Instead of treating each match as an isolated event, we could begin treating multiplayer games as evolving social environments.
Engagement is not enjoyment
There's another reason I find conventional matchmaking objectives unsatisfying: the games industry already has considerable experience optimizing measurable player behavior without necessarily improving the player's experience. Consider battle passes. A battle pass creates a progression structure around limited seasons, rewards, challenges, and recurring opportunities to earn experience. There's nothing inherently wrong with this. Progression can be satisfying, and rewards can give players meaningful goals.
But battle passes can also introduce deadlines, recurring obligations, and fear of missing out. A reward is available now but might disappear later. A challenge resets tomorrow. You've already invested time into unlocking something, so abandoning it feels like wasting your progress. Suddenly, the game isn't merely inviting you to play. It's giving you reasons to feel that you should play.
Research on daily quests and engagement rewards has documented this tension. Players may experience these systems as motivating and enjoyable, but also as chores, obligations, or sources of FOMO.
And this creates a fundamental measurement problem. A player returning every evening might be having the time of their life. They might also be exhausted, frustrated, and playing because they don't want to miss an exclusive reward. Both players appear engaged. Only one is necessarily enjoying themselves.
Now consider the research paper EOMM: An Engagement Optimized Matchmaking Framework, published in 2017 by researchers including collaborators from Electronic Arts. The paper challenges the assumption that fair matches necessarily maximize player engagement. It proposes a matchmaking framework that attempts to reduce disengagement by predicting how players will respond to different match outcomes. The research is interesting because it demonstrates that competitive fairness and continued participation aren't interchangeable objectives. It is not evidence that Electronic Arts or another publisher secretly deployed such a system in production.
But it also raises a question. Should continued participation actually be the thing we optimize?
Suppose a matchmaking system discovers that certain frustrating experiences make a player more likely to queue again. Perhaps losing narrowly creates a feeling that victory is just one match away. Perhaps a particularly satisfying victory after several defeats encourages another hour of playing. An engagement-optimizing system could potentially discover these relationships without understanding whether the player is actually having a worthwhile experience.
And that's the danger of optimizing what we can measure. Win probability is measurable. Retention is measurable. Match duration is measurable. But the satisfaction of seeing a familiar rival on the opposing team is much harder to measure. The pleasure of finally defeating someone who has repeatedly beaten you is harder to quantify. The gradual emergence of friendship, rivalry, social status, and community is even more complicated.
Yet these things may be precisely what makes multiplayer games worth playing. And unlike a battle pass that gives us another reason to return tomorrow, a genuine social relationship gives us somebody we actually want to return for.
Balance is not the same as good design
The obsession with balance doesn't end with matchmaking. It extends into how developers approach the games themselves. Deadlock is still developing, and substantial changes are inevitable. Balance patches are necessary. Sometimes mechanics are genuinely unhealthy, certain heroes become excessively dominant, and strategies emerge that leave too little room for meaningful responses.
But I wonder whether developers sometimes undervalue stability. A relatively stable metagame allows players to develop a shared understanding of the game. They learn what certain heroes are capable of, develop counterstrategies, establish conventions, and discover which approaches work against particular opponents. Eventually, those mechanics become part of the game's identity. And perhaps more importantly, they become part of the community's shared culture.
Consider Wascot's deceptive coins. The mechanic wasn't interesting because it provided a beautifully symmetrical combat interaction. It was interesting because it introduced a ridiculous problem into the game. The opponent had to recognize the deception, understand the threat, and adjust their behavior.
Axe's taunt in Dota 2 operates on a similar principle. Once caught, the opponent temporarily loses control of their character. The meaningful decisions often happen beforehand: positioning, anticipating initiation, managing vision, and preparing counterplay. There's strategic depth in responding to an unpleasant situation. And there is enormous enjoyment in being the person creating that situation.
A frustrating interaction isn't necessarily a badly designed interaction. Sometimes making the opponent uncomfortable is the entire strategic purpose of an ability.
Of course, not every frustrating mechanic deserves preservation. An oppressive strategy that offers no meaningful counterplay can become tedious, and an excessively dominant hero can crowd out variety. But the relevant question shouldn't simply be whether the mechanic is statistically balanced. It should be whether it creates interesting decisions, memorable encounters, and worthwhile forms of competition.
Likewise, changing the metagame is not automatically an improvement. Disruption can be healthy. Discovering new strategies is fun. Nobody wants a completely solved game that never evolves.
But stability also has value. Players invest time mastering mechanics, developing approaches, and establishing a common competitive language. Frequent changes can erode that accumulated knowledge. There's a tradeoff between novelty and continuity. And continuity deserves to be treated as something valuable, rather than merely an obstacle to keeping the game fresh.
The social contract we never explicitly agreed to
I keep returning to my experience playing Deadlock late at night. The small player population meant that I frequently encountered familiar names. Over time, we developed expectations about one another. We learned which players were reliable, which were aggressive, which preferred farming, and which were likely to make unusual strategic decisions. That knowledge influenced how we played.
Sometimes I would make strategic calls. I enjoy coordinating teams, discussing objectives, identifying which hero should receive more resources, and deciding when we should fight or push. In a game as strategically complex as Deadlock, these decisions matter enormously. But coordinating strangers is difficult. They don't know whether you're competent, whether your proposed strategy makes sense, or whether you're simply another arrogant player convinced that everybody else is an idiot.
Familiarity changes that. If people have played with you repeatedly, they have evidence about your judgment. They know whether your calls tend to work, whether you understand their preferred heroes, and whether you're capable of adapting when a strategy fails. Trust doesn't have to be manufactured by an algorithm. It can emerge naturally through repeated interaction.
And this is precisely what I mean by an unspoken social contract. Nobody formally agreed that we would all play at the same hours, recognize one another, and develop shared expectations. It simply happened. A community emerged from the structure of the matchmaking environment. And because it emerged naturally, its relationships had a certain authenticity that would be difficult to recreate through an explicit system of social preferences.
This also explains why climbing the ranks can be attractive for reasons that have very little to do with numerical progression. At higher ranks, particularly in smaller populations, you begin encountering the same people repeatedly. You enter a more recognizable competitive environment. Your reputation matters. Your decisions have continuity. Your victories and failures become part of a shared history.
And your rank becomes more than a number. It becomes a way of belonging somewhere.
What should matchmaking actually try to accomplish?
I don't think the answer is to abandon competitive ratings, remove skill restrictions, or deliberately create unfair matches. Those approaches would introduce obvious problems. Some players would be repeatedly overwhelmed, while others would enjoy the benefits of competition without facing meaningful resistance. But neither do I think the solution is to continue pursuing ever more precise skill matching as though achieving a 50/50 predicted outcome were the ultimate objective.
Ranking and matchmaking are different problems. A ranking system attempts to estimate how successful someone is relative to the competitive population. A matchmaking system decides which competitive encounter that person should participate in next.
The first is primarily a statistical problem. The second is also a social problem. It involves relationships, familiarity, personalities, expectations, and the possibility of creating experiences that extend beyond a single match.
Perhaps a better system would deliberately preserve productive groups of players even when another combination could produce a marginally closer predicted outcome. Perhaps it would recognize that repeated encounters between familiar opponents can be valuable, provided those encounters remain competitive and don't become abusive. Perhaps it would tolerate a wider range of skill differences when doing so creates worthwhile competition without making anyone consistently helpless. And perhaps it would use our increasing ability to understand player behavior to study what actually makes multiplayer games meaningful, rather than merely predicting who will win or who will queue again.
None of this guarantees enjoyable games. Nor does it imply that an algorithm can manufacture friendship, rivalry, or community on demand. In fact, that might be the most important limitation to acknowledge.
The objective shouldn't be to manufacture social relationships. It should be to stop destroying the conditions in which they can develop naturally.
Sometimes that might mean keeping a lobby together. Sometimes it might mean making room for more varied competitive encounters. Sometimes it might mean preserving a stable metagame long enough for a shared culture to emerge. And sometimes it might simply mean recognizing that a familiar opponent is worth more to someone's experience than another anonymous player with a marginally more suitable rating.
The match was balanced. So why wasn't it fun?
We've become remarkably good at describing multiplayer games through numbers. We measure skill, uncertainty, win rates, retention, engagement, match duration, and thousands of other variables. And those measurements are valuable.
But the things that make competitive games memorable are often much harder to quantify. The rival you've been trying to beat for weeks. The teammate whose judgment you've learned to trust. The obnoxious player whose nickname immediately tells you what kind of evening you're about to have. The satisfaction of becoming genuinely good at something and having the people around you recognize it. The small community that forms simply because the same people keep showing up.
These aren't incidental features of multiplayer competition. For many of us, they're the entire reason we play.
And perhaps that's what lies underneath so many complaints about balanced lobbies, sweaty matchmaking, constant metagame changes, and the disappearance of persistent servers. Not simply a desire to win more. Not merely a dislike of difficult opponents. But a longing for competitive games to feel like social environments again.
Places where people recognize one another. Where reputations matter. Where becoming better has visible consequences. Where rivalries develop, friendships form, and the next match feels like a continuation of something rather than another isolated transaction.
I don't necessarily want the fairest possible match. I want to recognize the people I'm playing with. I want to remember what happened last time. I want to know who I'm up against, what they're capable of, and why beating them would mean something. I want my games to have history.
A perfectly balanced match is a mathematical accomplishment. A match that becomes a story worth remembering is something else entirely. And I wish competitive game developers spent more time thinking about the difference.