: give me ONE chat log where someone got punished for a single instance of "shut up" no one ever get punished for just "shut up" whether by the IFS or Tribunal. Usually there are much more flaming, passive-aggressive, condescension, negativity, argument, rank/kda/etc. shaming, report harassment, etc. etc. involved ON TOP of that one "shut up"
Dude, you’re wrong. I have a support ticket outlining “shutup” as a phrase deemed breaking the summoners code. I can ss for you, its your word versus riots.
rujitra (NA)
: You’ve said the Tribunal was better because the bot doesn’t review context. The tribunal displayed that no matter the context, players would vote the same on the same chat behaviors.
That was only because there was only one right answer? Of course no matter the context the players would vote the same...if they didn’t they’d have their accuracy knocked down. **I’m not arguing whether or not the Tribunal is BETTER than the AI, I’m arguing for its coexistence**
: You were the one that mentioned other games. CSGO is the only one of them I'm somewhat familiar with; which is why I mentioned the difference of what they handled. Since they focused on different things, doesn't seem right to really compare them imo. Runescape, the only thing that comes to mind are P-Mods that can mute people, but they have to be around for that to happen.
The point I’m making is they’re all **forms of community enforcement**
rujitra (NA)
: It doesn't matter that others were misbehaving - the Tribunal had a reputation for punishing players who were toxic or abusive **regardless** of what else was going on in game. You are free to disagree with that, but the Tribunal consistently demonstrated that across similar chat logs where the only difference was the game or what others were doing, the results were the same.
I don’t think I’m arguing the contrary here.
: I participated actively during the Tribunal era. If you want to go by my accuracy rates, circumstantial evidence is meaningless when the person in question did flame. My accuracy rating took several hits when I voted to pardon someone who got flamed by their team and toward the end, flamed back. The majority decision is still punished. If you vote punish blindly without reading in Tribunal, you would get about 90% accuracy rating....circumstantial evidence or not
I believe flaming and a solitary “shut up” are different. Currently, because it is an AI system, you will be punished for saying such even if only a couple times throughout a 30 minute game. (In correlation with my previous hypothetical scenario) I’d argue that isn’t toxic and it shouldnt be punishable. That was the case previously too, as far as I recall that wouldn’t be enough to garner a mute. Because this system uses keywords (which as I previously stated is useful for non tolerable remarks (racist, homophobic, etc.) but ends up being borderline comical for other things like “shut up” or “boner”. Like am I hallucinating in thinking that that should be fine? Game might as well be rated E at this point.
: People are not punished for saying "shut up". Unless this "shut up" is one of 30+ entries in chat log over 20 mins game.
That’s actually not true, telling someone to shut up is in itself a punishable offense. I have a support ticket from Riot stating this.
Kei143 (NA)
: Sure they could write some program to automatically pull games or chats and put them on a website for the community to vote, but for what purpose and at what cost? The verbal system is already 99.9% accurate, and the 0.1% of "error" includes player support lowering punishment tiers for first time ZT offenders. So a tribunal for the verbal punishments will just contribute to the community feeling they are contributing, but not actually contributing. On the gameplay side it'll probably be 50/50 where the community will get it right. At the end, the Rioter will have manually review the game and dish out a pardon / punishment based on his point of view, and not the community's. Instead of waiting for the community to vote, speed of getting to the punishment is much more of importance to remove true trolls and griefers compared to letting the roam free and letting the community feel good. There's also the time and resource needed to build the code that can drag the chat and gameplay to a website. Engineering is the most bottlenecked human resource Riot has. I'd rather the engineers use their time to squish spaghetti code bugs, code SR map to change terrains (as seen in pre-season 2020 videos), or even creating new games for different player experiences.
Well articulated and fair enough. I wasn’t under the impression they’d need to engineer an entirely new system again. EDIT: Riot is subjectively as big as the other companies I listed with implemented community enforcement. They made it work and I’d shift the focus then to community enforcement (in general) and it’s lack of existence in LoL alongside other popular multiplayer games.
: Tribunal handled verbal cases. Things with chat logs and was inefficient in comparison to the IFS as it took months for some cases to even be reviewed. Overwatch in CSGO focuses more on gameplay and seems to have actual clear lines for what is trolling and what isn't.
I'd argue that CS:GOs overwatch system isn't super comparable to Leagues Tribunal other than they were both community based enforcement. CS:GOs overwatch was used to punish hackers and griefers whereas the Tribunal was more catered to verbal abuse and chat. I'll agree with you in the sense that intentional team griefing is much more concrete in that game though.
Kei143 (NA)
: Let's start with why it was removed; https://nexus.leagueoflegends.com/en-us/2018/08/ask-riot-will-tribunal-return/ > The tl;dr is Tribunal isn’t coming back any time soon. > > We learned a lot from Tribunal when it was around, and it actually became one of the bases of our current models for abusive chats. A player-voted service like Tribunal comes with two huge drawbacks that are really tough to get around, though: > > 1. It’s slow and inefficient. Manual reviews require pulling those chat logs out to the website, waiting for responses from enough players, and deciding on a penalty from there, while the current automated systems can send out penalties in less than 15 minutes. > 2. It’s sometimes wildly inaccurate (especially when there was an IP reward per “successful” penalty, leading to a super innate bias in the system). > > There’s a couple of things Tribunal brought that we know we haven’t tackled in a big way, like more focus on trolling and intentional feeding detection (though we are continuing to iterate on how we detect these) and more agency to make it feel like you’re really having an impact on unsportsmanlike behaviors. > > We’re making a lot of progress with dodgy chat content and the new name filtering service with our automated models, but we do still need you to hit the report button when you see bad shit to help us gather information on what you don’t want to see in League. It really does make a difference. You are correct it was inefficient. A Rioter had to manually pull the logs, manually post it on the website. After the voting was done, the Rioter had to manually check whether the vote was correct. The thing the tribunal was good for was to provide some visibility to the community and make them feel they were contributing.
Which is why I proposed their co-existence. If the tribunal was good for visibility to the community and made them feel like they were contributing, why can it not exist with the AI system that negates the inefficiencies of the Tribunal?
rujitra (NA)
: I'm intrigued why you think the AI system, which is faster, more consistent, and more fair than the old Tribunal is "inferior".
I'd like to reiterate that I believe neither of these systems are amazing by themselves but I'd like to point out the inaccuracies of any AI system and the circumstances revolving around inaccuracies. You can submit a support ticket if you feel wrongly punished sure, but take a hypothetical case like this for instance; >You're being flamed by a blatant racist for the majority of the game, you mute him but the rest of your teammates are still engaging with this idiot. >You tell this guy dude please shut up and play the game >You're bestowed a mute or temporary ban because shut up is a key word for the AI system AI doesn't account for circumstantial evidence and I'd have a hard time justifying any sort of punishment on the guy that said shut up. But hey, maybe I'm misinterpreting the kind of environment riot is trying to create
: There are not enough people who will inspect games for free to judge people If you add a reward for that, there will be people who will just spam judgements for rewards, and that would be even worse Current system is not perfect and it misses almost everything but hard int and chat toxicity, but at least it is consistent, it works the same way for everyone and everyone knows what rules they shouldn't break. And thats how it should work, to be honest. There should be strict rules thta you can be puished for, not some shitty people deciding "well, i think building full ap on bard when he is already 0-3 is trolling, lets ban him"
That's why I initially suggested they should be co-existing. It gets rid of the inefficiency of the Tribunal through compensation of AI, and then the AI could stop having such tight constrictions on the parameters of what's "toxic". Clearly 0 tolerance policies should exist in the realms of racism, sexism, homophobia, etc. but things like "shut up" are honestly nonsensical in terms of punishability.
Comentários de Rioters

Łęe Sin

Nível 43 (NA)
Total de votos positivos
Criar uma discussão