Competitive multiplayer games see millions of players logging on every day. Making sure they're matchmade into the right sessions is as much an art as it is a science, one that Josh Menke is here to teach.
Josh chatted with GDC over a video call about his two-day GDC Masterclass, Engagement using Matchmaking and Ranking in Competitive Multiplayer Games, which is being held on December 9 and 10 from 9:00am to 1:00pm PT (12:00pm to 4:00pm ET). Josh shared how matchmaking has changed over the years, as well as the biggest mistakes developers make when building their own matchmaking and ranking systems.
Below is an edited, condensed version of our interview.
GDC Staff: You're teaching a Masterclass about matchmaking and ranking in competitive multiplayer games, a field that has exploded over the past 5-10 years. How has the art and practice of matchmaking players evolved since the early days of these competitive multiplayer games?
Josh Menke: There's a long history there, really. The old way you match-made was you just logged into a game and you tried to find a server or a room to join on your own. There was usually some server out there that had a master list, and then your game would ping that server and it would get back this big list and give you the big list. Then you would have to go scroll through that list to try to find a game that somehow looked good to you.
Then they started to automate that. The first approaches were, like, we'll just have the game automatically search through that list for you and then try to get into a game automatically. That's how some of the games have implemented it, even recently some games still do that. Your game asks some server, "What are the available games?" Then your own game tries to get into one of them before everybody else does.
The bad old days were just do-it-yourself. The benefits of that is no one blames the developers when the matches are bad, but the downside of that is it's a lot harder to get people into matches.
Over the years, we've evolved more into one central location that is negotiating everybody's desires to have a match and trying to get them into higher quality matches all at the same time. That's kind of how matchmaking has gone.
GDC Staff: What are the biggest changes you've seen with players, specifically? Especially kids who've grown up on these competitive multiplayer games, and they've grown up watching esports, Twitch and livestream?
Josh: It just depends on what game they grew up on. They tend to kind of keep that culture, those expectations. Although I do see when they go from game to game, their expectations are more adapted to that game—at least in my opinion. You'll have players who play Call of Duty, that will be like, I don't like skill-based matchmaking, but then they go play Valorant and it's fine.
While a lot of players who first experienced Call of Duty, that was back in 2004 and 2005—when Modern Warfare hit is when it really exploded. That game did have some skill-based matchmaking—all of them always have, it's just the math and the science has gotten better over the years and caught up. If you grew up on it back then, your expectations are very different than if you have it now.
The same thing happens in Fortnite, even today. When the game first started and kids were playing Fortnite, I believe they had very little skill-based matchmaking and then over the years–I don't work there, I don't know exactly what they're doing—I feel like they've experimented with various levels of skill-based matchmaking and various degrees of using bots and trying to find that sweet spot. But yeah, you definitely have people who started early on in Fortnite, who didn't like when they started adding the more skill-based matchmaking.
GDC Staff: What would you say is one of the most-common mistakes you see designers or developers doing when they're building their own matchmaking and ranking system in their game?
Josh: The whole area is really a minefield. On the matchmaking side, the most common mistake—and this is a mistake everybody makes, and I made over most of my career, and we only recently realized was a mistake—is the common way to matchmake is have this goal of, "I want to match really good," and then relax my constraints over time. Settle, basically. Wait for awhile, and then settle for a bad match. And that's provably the wrong way to matchmake.
The way it should be done is you should be gathering statistics, real-time stats on your population, and then using that information. Picking what those constraints should be real-time, and not expanding. Because expanding gives players this feel of, "Don't worry you're going to wait a little bit longer but the match will be much worse, it's okay." Which is not the message you want to send. You want to say, "You're going to wait longer because we know we can find you something better if you wait longer, so we're not going to expand it because we know it's coming."
GDC Staff: What drives you to share your knowledge and experience with others?
Josh: I really like this area a lot, and I've found a lot of fun things in it. I've found a lot of good solution to these problems. And I guess I'm the type of person that when I find something–it's like when you find a place that has a flavor of ice cream you like, you want to share it with people. And so when I find a new method that solves problems I know we all have in this area, I like to share it with people and say, "Hey, this is what we've found."
I don't want this to be an area where companies feel like they need to be secretive because of competitive advantages and things like that. I feel like the competitive advantage of a game should be in the gameplay, not in the matchmaking. Ideally, we'd all get out of each other's ways and everyone have good matches, so we could figure out which games are better.
Be sure to head to GDC Masterclass for more information on Josh's Engagement using Matchmaking and Ranking in Competitive Multiplayer Games course, which runs December 9 and 10. Keep in mind this course is happening virtually, not in person.