Hey players, it’s Hana “TimTamMonster” Dinh, your product lead for Behavioral Systems. Today, I want to dig a little into some of the results we've been seeing from the recent changes and features we've added that tackle some of the biggest player behavior pain points you've been feeling in League.
As is always the case for Behavioral Systems, we are constantly trying new strategies and approaches to improve our effectiveness in this space. We know that in the past you have been unsatisfied with the effectiveness of our systems to protect your League experience, so we’ve been applying your feedback and a large amount of research in the space to try new kinds of solutions to your most immediate problems.
And the result has been promising: You’ve been telling us that our recent changes have made significant positive differences to how players are acting in your games. We know there’s still so much more to do, but these changes are giving us a lot more confidence that we’re moving in the right direction, and we’re looking forward to making even more impact on preventing disruptive behavior from ruining your experience in games.
Some of the big changes we’ve introduced that you may have noticed are:
As a result of the above changes we made in late 2022, we heard that this improved your experience of disruptive behavior significantly.
We’ve seen increases in the following areas:
The full integration with our game agnostic text evaluation service (GATES) machine learning models improved disruptive text detection by over 15x.
Some of the key learnings we’re taking away from these results are: we've reduced the time between the offense and player penalties and ensured our features reach the maximum number of players, and proactively protected players before they are harmed by disruptive behavior.
We’ve released some other features this year including a party chat, which allows you to only chat with your premade teammates. We know that it can often be a safer and more protected environment to communicate in your premade team rather than the entire lobby. We created new commands for muting yourself and deafening all players so other players know you are disengaging with chat. And most recently, we’ve finally released in-game reporting, allowing you to report disruptive behavior right when it happens! This unlocks us to take additional action in real time as we help improve behavior in your games.
We’ve also been continuing to improve our existing solutions for game ruining behaviors. We recently extended our credibility models (using probabilistic math to determine disruptive behavior based on number of reports) across multiple report categories. We added penalties for players any time they leave/AFK a game.
In the next few months we will be restricting Ranked play for players who have been punished by our systems. This means that if you are punished for disruptive behavior including intentional feeding, multiple AFKs, and severe chat abuse, you will be restricted from accessing the Ranked queue and will need to play in other queues to re-unlock the Ranked queue.
While there may be concerns around disruptive players moving to other queues, we recognize that the environment of the Ranked queue results in considerably more disruptive behavior. These restrictions will also be in addition to all existing punishments, so for players who are consistently disruptive, they will receive account bans where they will be unable to play any games no matter what queue for a specific period of time. We will be monitoring disruptive behavior in other queues to ensure it doesn’t measurably increase after the release.
Finally, as you know, we have two splits for this year’s Ranked season. There won’t be a separate reward for Honor at the end of the first split, and instead it will still be calculated on an annual basis. That means End of Season rewards will be awarded near the end of the calendar year.
While the majority of what we’ve covered here is more text or social based interactions we’re also looking at some more in-game behavior impacting situations and will have some updates on that front soon.
We know how critical this space is so we’re continuing to learn and improve our approaches over time. Through some of these promising results, we’re also excited to begin to experiment more significantly with some of our approaches. With that said, thank you for your feedback and we’re looking forward to integrating even more features to improve your experience real-time in game in the future.