Gaming News

An alternative matchmaking system for skill-based games

Gamingtodaynews1f - An alternative matchmaking system for skill-based games

This idea is something I have been thinking about based on my experience playing lots of different games. It’s not doing anything useful sitting in my head, so I thought I’d post it here instead.


In most skill-based games, the matchmaking system is based around the same central idea. The idea that the best game experiences happen between two players of identical skill.

The idea being that this creates the most exciting types of moments in games. Those “Yes, I made it!”, “Shucks! If only I had…” and “I can’t believe I lost!” moments.

While these are indeed some of the most memorable moments in gaming, I believe that focusing solely on this goal leads to many other enjoyable areas of the gaming experience suffering negatively, such that the value of the overall experience is reduced.

Some of those areas are…

Lack of variety:

Once a player’s skill level has been accurately measured, every match is aimed to give exactly the same level of challenge.
Some games may be able to provide variety in style of play. Even then, the intention is that the difficulty is always exactly the same.
There will never be a feeling of “Phew, this should be an easy game” or a “Oh God, I’m going to get crushed!”. Always the same feeling of “I might have a chance here”, over and over and over.

Generating a kind of emotionally monotonous experience for the player.

To be fair, most players have probably had an experience of being matched against an opponent of significantly higher or lower rank in a game. However, when this happens it is because the matchmaking system has failed in it's goal, the goal to provide what it considers a good match (usually due to low player availability). As such, it often happens in a way that detracts from the player experience rather than enriching it.

Difficulty building community:

Searching the global pool of available players, to get the most identical match of skills for each game, makes it very unlikely that any two players will ever meet enough times to even recognize each other.

This often has the result of effectively reducing a player’s interaction with a human opponent into one almost indistinguishable from an interaction with an AI opponent.

It inadvertently creates an environment that discourages people from acting in social and friendly ways. Being nice takes effort. It is only worth it because it might lead to the reward of making friends. When that reward is removed, then there is no more value in putting in that effort.

Of course, many game designers are aware of this. And they put in functions to mitigate it in different ways. E.g. the ability to request to replay the same opponent (or replay with the same team), the option to add people we've played with to our friends list etc.

This model tries to include those community building elements at the root of the matchmaking system, hopefully enabling game designers to go much further in encouraging and fostering the type of community they want.

Lack of feedback on performance improvement:

There is a kind of intuitive experience curve when learning a new skill.

It starts out being very difficult, with every inch of progress a struggle. Then as we get better, what used to be hard starts to become easier. Reassuring us that we are doing better.

At a certain point, we move on to a harder task (or, in the case of a game, a higher level of difficulty) and the cycle of challenge starts again.

When every game is matched to your current level, this cycle never happens. The player never gets that reward of easier games because their skill level has improved.

Instead, the players only feedback is their ranking number slowly ticking upward, downward or staying static. A much less visceral experience.

Another point related to this;
Playing against someone who is better than us is one of the funnest ways to learn.

For this to happen though, we have to be able to play against people of higher skill than us. Where the skill gap is big enough that, not only can we see that they are more skilled than us, we can also see why they are more skilled than us. And we can also see a path from our skill level to their skill level.

If the skill gap is too big, e.g. a beginner playing a pro, that last part (seeing a path from our skill level to their skill level) won't happen and it will simply leave us frustrated for losing so easily.

Constant pressure:

This is related to the first point of lack of variety.

Sometimes, a player may come to a game looking for a serious challenge, other times they may be looking for a more relaxing experience.

In a game with skill-based matchmaking, even un-ranked games tend to be skill matched. The only option for a relaxing game is then becomes letting yourself lose. And I don’t know any way of getting a more serious challenge.


The suggested alternative:

I think a lot of these weaknesses can be addressed by using a matchmaking model that takes ideas from the 90s neighbourhood arcade experience.

The basic concept is:

Rather than searching a huge number of players to find those most identically matched in skill level for each game, we split the payer base into smaller, more stable groups (neighbourhoods) but with a much more lenient skill range in each group.

The target size for each neighbourhood is hopefully large enough that searching for a match doesn’t take too long, but small enough that regular players can begin to get to know each other after a short time of playing.

In an ideal world, these neighbourhoods will be organised by geographical area, as a subtle gesture to encourage the social connections built in the game to spill out into the real world.

The neighbourhoods will then be divided into leagues based on skill, in a pyramid structure.

E.g. 1000 state leagues, 100 national leagues, 10 continental leagues and 1 global league. (In practice, I believe more than 4 league levels will be needed though. 6 or above should work better).

This should immediately address many of the issues raised with the current system (A more varied experience, across different skill levels, with a more static group of fellow players encouraging better communal behaviour).

However, we can still make some more adjustments to this base model to accommodate players of different temperaments and improve the experience in various ways.

  • Increasing or reducing rank:
    Some players will be looking to rank up quickly while others may be looking for a more relaxed game.
    To accommodate this, this system will give each player access to games across three league levels, in their current neighbourhood.
    Their own league level, i.e. the one where their rank is currently has them placed, as well as one league level above and one league level below their current level.

When playing a level above their current league, we can assume the player either wants to rank up quickly or is just looking for a more difficult challenge. To facilitate this, we can have players rank up twice as fast when they win when they win a game a rank above their own (with increased ranking gains for win streaks). If they lose a game however, they still lose rank at close to the normal rate. Effectively, the downside of losing stays the same, but the upside of winning is drastically improved.

When a player is playing a league below their current rank, we can assume that they are seeking a more relaxed game. As such, when it comes to ranking, there will be no upside for winning and a greatly reduced (but not zero) downside for losing.

  • Protecting beginners:
    When starting any new game, even small differences in experience can greatly impact the gaming experience. So we want to keep beginners in a separate matchmaking pool while they learn. Leaving that pool can be framed as an early achievement. Eg. “Win five games to access your neighbourhood league”.

  • Smoothing out shocks:
    One of the ideal goals of this system is encouraging a more varied experience.
    Instead of the player constantly getting matched against people at an identical skill level, we are aiming for the player (when playing in their own league) to have a more varied (but hopefully still roughly balanced) experience of playing against people of both higher and lower skill .
    However, this creates a problem for people near the top or near the bottom of their league. For a player reaching the top of their league, they will necessarily continue to be placed against players who are mostly of lower skill than they are. Then, as they manage to break into a higher league, they will suddenly be in the bottom half of the league. In an environment where everyone they are matched against is of a higher skill than they are.
    This also means that players in the top half of a league are likely to win more games and keep increasing in rank faster, while those in the bottom are likely to lose more games and lose rank faster.
    To deal will this, we can have a promotion/relegation zone between leagues where matchmaking acts differently. Once a player gets close to the top or bottom of their league, they will enter promotion/relegation zone.
    Once in this zone, the player is treated as if they are effectively in both leagues. So, they get matched against players from both the league they are in and the league they are approaching. If they choose to play one league up, then they will be matched in a league above the higher league. If they choose to play one league down, then that would be a league below the lower league. Rankings and rewards adjust accordingly.

  • Choice:
    A big part of the goal of this model, is to create groups that are small and stable. To encourage / allow communities to form.
    However, there is always a possibility that, for one reason or another, a player might not enjoy the social experience in their original neighbourhood (e.g. they might log into the game at an unusual time, when others in their geographical neighbourhood are asleep. Or they may want to play with a friend who is in a different neighbourhood ). If that is the case, we don’t want them to feel trapped in their assigned neighbourhood.
    However, at the same time, we do want to “lock in” the neighbourhoods a bit, to give the community a chance to form and self-regulate.
    My suggestion would be to give each player access to one or two guest neighbourhoods (The equivalent of visiting an out of town arcade).
    This will be different neighbourhood, at the player’s league level, which the player can switch to for matchmaking purposes. Games played in guest neighbourhoods won’t affect a players rating.
    Also, players will only be able to change their home neighbourhood to another one at specific time intervals (Perhaps once a month or once a season) changing guest neighbourhoods can be more lenient.

Well, that’s the basic structure of the model.

I do think this model would then lend itself to various other design elements that could further enhance the gaming experience.

Some examples:

When a player is playing in a league lower than their own (looking for a more relaxed game), they could be loudly tagged as “Visiting Champion” to the opponents in that league.

This might inspire terror in the lower ranked players, or even (in certain types of games) even inspire cooperation or alliances among the lower ranked players to take the champion down. Penalties for losing to a “Visiting Champion” should be reduced (most ranking systems automatically take this into account). The game could then give more exciting awards to lower league players who score points against them.

Conversely, when a player is playing a league above, they can be tagged as “impetuous upstart” or something similarly cheeky.

This might encourage supportive behaviour from some friendly higher ranked players. This is a good result. Helping build the community.
However, it could also encourage an aggressive “Who do you think you are?” response from some others. Which could lead those higher ranked players to try even harder to punish the interloper. This is also a good result!!
Because the “upstart” player has specifically chosen to have a more difficult experience! People of higher rank, going out of their way to crush them, is exactly what the player is looking for.

These small stable neighbourhoods can allow more stable rivalries to form.
A player could gain a “nemesis” over multiple games. This would be a player of the same or lower skill level in your neighbourhood that you have consistently lost to.

The matchmaking service could allow players to ‘queue’ for a game with any specific player in their neighbourhood.

Even more open neighbourhood game lobbies, where all members of the neighbourhood can chat amongst themselves while waiting for a game or doing other stuff in the menu.


What do you guys think about this system?
Can you think of any ways to improve it?
Can you think of any games that might benefit from a system like this?

Source: Original link

© Post "An alternative matchmaking system for skill-based games" for game Gaming News.

Top 10 Most Anticipated Video Games of 2020

2020 will have something to satisfy classic and modern gamers alike. To be eligible for the list, the game must be confirmed for 2020, or there should be good reason to expect its release in that year. Therefore, upcoming games with a mere announcement and no discernible release date will not be included.

Top 15 NEW Games of 2020 [FIRST HALF]

2020 has a ton to look forward the video gaming world. Here are fifteen games we're looking forward to in the first half of 2020.

You Might Also Like

Leave a Reply

Your email address will not be published. Required fields are marked *