I would put a guess in the 70% range. From what I've seen, and discussed with some other Rioters, it appears to be extremely successful in punishing those who deserve the punishment. A lot of thought and analysis goes in the the decision making for bans based off of report. Sure, there's a system that sorts the cases into "Yes- this needs to go further" or "No- this is just a troll report", but the actual ban decisions is made by the specific Riot team themselves based upon past games and experiences that the player had.
I'll throw in a quick quote from Phreak here about nonsensical reports / game anomalies here.
One thing I really gotta tell you is to not worry about frivolous reports. There's a few things to keep in mind:
- People who abuse reports, their reports become meaningless. So anyone who's abusing the report system is basically not abusing the report system. So you truly don't have to worry about abusers putting you in the tribunal, or up for restrictions. If this is your primary concern, do whatever you can to push that out of your mind. I've looked up my report history and had a bunch of bogus reports for things like "Didn't build a Trinity Force" and "Is impersonating a Rioter" and I have never ended up in the Tribunal, with Chat Restrictions, etc. And I know I lose my temper sometimes, so I have real reports in there, too.
But again, all I can do is just keep restating point #1. Fake reports aren't going to get you.
So there's a pretty damn high chance that you won't get permabanned randomly for just having one or two bad games... but if it is a consistent trend, then you sure as hell deserve it.
Lyte also throws in a quick bit about previous Tribunal stats in one of his responses:
For the majority of players (74%), the first warning in the Tribunal was enough for them to improve their behaviors. In fact, with warnings + chat restrictions, that number goes up to 75-77% in many servers.
More % of players improve their behaviors after the next tier ban, but at the end of the day, we see about 1-3% of players refuse to reform regardless of the consequences. Systems like the one here try to identify those players, and immediately escalate their bans to the highest tiers. We've seen that players who tend to be consistently severe in their behaviors tend to never reform, which is why we built a machine learning approach to try to identify these patterns.
In general, we'll always focus on reform. It works for the vast majority of players that ever see a punishment system; however, by giving the worst 1-3% so many chances, we're actually letting them ruin a lot more games and players' experiences and that's something we want to try to reduce.
So the previous system from Riot had to have been doing SOMETHING right, so I assume this one will be pretty effective. Sort of a trial-and-error thing, Morgan would quite proud I think! However, there's always the chance that people won't change and in these cases, the permas should be handed out.
One last point to make here, just gonna throw in my last quote from Lyte:
No systems are perfect so obviously there will be false positives. In fact, even when every single case is reviewed by a human being, there are still errors sometimes. As people get tired they might miss something, or they might click the wrong buttons in the interface. However, every time we see a request on social media, we'll commit to being transparent.
So hopefully, it will have a positive overall impact to the game by removing the negative side of the community and keeping the community in check! Though it isn't a permanent feature at the moment, I strongly believe that the results of the test will be enough to implement it as a permanent feature. However, the function should be better and more universally announce, as well as have an easier route of access for players to see the chat logs / reasons that caused them to receive the ban.
Riot has a long way to go with this system, but I believe it will only improve the game that League of Legends has become.
