Ranked Based Matchmaking
Ranked based matchmaking is a system that matches players of similar skill levels together based on a player’s rank, defined by their performance in matches, ideally creating an competitive environment where players feel challenged but not overwhelmed. This system is used in many competitive games: Valorant, League of Legends, Overwatch, and Counter-Strike.
The objective of these ranking algorithms is to take all the data points from the game such as match outcome, round differencial, individual performance and create a score that represents the player’s skill level. Whilst these algorithms are usually kept secret to prevent exploitation, they are developed by multi-million dollar corporations which are incentivized to keep players hooked to the game. Therefore, they have a strong incentive to make sure that each game is optimized to have a 50-50 probability for win or lose, creating a sense of competitiveness and challenge for the player.
Furthermore, because most ranked-based competitive games are team-based games requiring coordination and communication, the win-rate of a player is not only dependent on their individual skill level but also the performance of their teammates. Therefore, increasing the variance in win-rate.
Additionally, (for the average player) most of the teammates will be strangers who have different emotional states. For example, if a player is on a losing streak, they may feel frustrated and demotivated, affecting not only their performance but also the performance of their teammates, creating a negative feedback loop. On the other hand, with a large enough playerbase, players may also have experiences where they are on a winning streak, leading to a sense of euphoria and overconfidence, which could rub off on their teammates, creating a positive feedback loop.
Whilst psychology is more difficult to predict when compared to the concrete data points available based on performance, the idea of variance in experience is important to keep in mind in the following section discussing dopamine and anticipation.
Anticipation and Dopamine
Dopamine is released when an individual expects or anticipates a rewarding or pleasurable outcome. It is important to note that this expectation can happen even before the reward is received.
Reward prediction error explains how dopamine release is based on the difference between expected and actual outcomes. When an event or outcome is better than expected, dopamine release increases, reinforcing the behavior. Conversely, if the outcome is worse than expected, dopamine release decreases (Similar to the loss function in backpropagation).
In the context of competitive games, the anticipation of winning can trigger dopamine release. This creates a pleasurable experience for the individual to continue playing (or gambling), even if they are not winning. Because the brain is sensitive to the uncertainty in reward prediction, the anticipation of the win can be more rewarding than the win itself, leading to even greater dopamine release.
Because it is human nature to be competitive, the anticipation of winning can be a powerful motivator to keep playing, even if the player is on a losing streak. This creates a cycle of anticipation, dopamine release, and prolonged gaming, even if the player is not winning. Additionally, because the ranked matchmaking system is designed to have a 50-50 probability of win or lose, players are more likely to experience the anticipation of winning.
Food for Thought
- Should game developers be more transparent about their ranking algorithms to prevent exploitation and addiction?
- What other algorithmic systems are naturally exploitative and addictive that should be more transparent? (Social Media?, Dating Apps?, etc.)
- How can players be more aware of the psychological effects of anticipation and dopamine release in competitive games?