Riot Games Plans to Reduce Chat Abuse and Toxicity in Their Games
The Player Dynamics team at Riot Games is taking steps to address the issue of chat abuse and in-game toxicity in their primary titles. By implementing new evaluation methods, Riot aims to tackle the problem, particularly among repeat offenders.
A Large Number of Reports
In an update on Player Dynamics, Riot revealed that they received about 240 million reports in 2021 across all their games. This amounts to a total of nearly 3 billion reports in a year. Considering these figures, if each Riot employee dedicated their time solely to reviewing reports, they would need to handle about six reports per minute.
The Challenge of Resolving Reports
Resolving reports is not a simple task. The Player Dynamics team focuses on improving player experiences and reducing harmful behavior. They receive reports for various behaviors, some of which may not be worthy of penalties, while others are deliberately malicious. Detecting behaviors like intentional feeding and trolling is more challenging compared to clear instances of text/voice chat abuse and AFK (Away From Keyboard) behavior.
However, Riot’s existing methods and penalty systems are effective. Less than 10 percent of players who received a report in 2021 received another report within the same year.
New Evaluation Methods
Riot has leveraged the data gathered over the past few years to develop innovative methods of evaluating player behavior. They plan to introduce automated voice evaluation in VALORANT later this year, which will enhance the detection of disruptive voice communications. Investments in machine learning and multi-language support will improve automatic text evaluation. Additionally, Riot is expanding their zero-tolerance word list.
The Player Dynamics team is also intensifying efforts to target repeat offenders, chat-based offenders, feeders, AFKers, and throwers. They are exploring real-time punishment options as well.
Many of these solutions and features will be rolled out over the next year.