The Matchmaking Process
To get an understanding of the problem, we first need to understand how matchmaking works. However, the way it works differs from developer to developer, and game to game. So let’s take a look at some case studies to see how different games handle matchmaking. We’ve introduced Elo, which a lot of games use, but here’s some alternative examples. First, let’s look at TrueSkill, the matchmaking system for games like Halo. According to one of the creators, “[the] rating is a Gaussian distribution which starts from N(25, ). μ is the average skill of a player, and σ measures how likely this is true. [The] real skill of a player is between μ±2σ with 95% confidence.” (Lee, 2012, para. 4). These are applied through the multiple algorithms that manipulate the rating depending on the type of match, determining how players would do against each other. Now, this is definitely a workable system. It’s been the backbone of Halo matchmaking since its inception, and has made its way into countless other games. The issue is that the data for it is locked to a set of systems, and that leads to some matches that have some bad experiences.
Another challenge with this approach is that it has difficulty balancing a predefined level gap which defines match balance and the amount of players of an appropriate level of skill. This has caused a rift in the developers, common players, and some content creators and professionals, who think that their skill rating doesn’t match with their actual skill. One example of a pro player’s perspective is the player Eric “Snip3down” Wrona, who stated in an interview recorded by Ethan Davidson that “I’m one of the best players in this game and I’m losing 70 percent of my games, how is this possible?” (Davison, 2022, para. 9, citing Wrona). This was a complaint to 343 Industries about how matchmaking in the game “Halo 5: Guardians” caused him to feel as if he was set up not to enjoy the content. The mindset seen in content creators like him is one where, rather than wanting to be matched up with people’s skill level or thereabout, they instead desire matches in which they can generate popular content by winning matches, fighting against casual players. But they aren’t a majority.
One thing that Thore Graepel and Ralf Herbrich state to the Game Developer Magazine of Microsoft emphasizes the importance of “the purpose of the game and the behavior of the rating system [being] aligned: people striving for high ratings should be forced to play in accordance with the spirit of the game. Taking the margin into account by which a game was won can be very misleading.” (Graepel & Herbrich, 2006, p. 3). To them, the clean sweeps that Wrona and streamers like him desire are detrimental to the other players, some of them wanting to genuinely improve, not get bullied by pro players for views and revenue. Still, Wrona’s view persists, and even some developers are agreeing. One example is Max Hoberman, designer of ranking systems in Halo 2 & 3, who stated that “perfectly balanced games…were often the most stressful.” (Davison, 2022, para. 27). With this big divide, a solution seems unclear, though the game Farmville may hold a temporary solution to this issue.