Boards›Player Behavior

A look at Toxicity: The In-Game Chat and Why You Shouldn't Use It

Sunka·11/18/2017, 9:47:12 AM·2 votes·602 views

[if you want to reply to this discussion, please do so in an educated and respectful manner. thank you.]

Why do we have an in-game chat system? I believe the purpose is so you have the ability to communicate with your team verbally, rather than graphically. But therein lies an issue; written word is often misunderstood or taken out of context. This is the concern I raise.

[1] To what degree do people truly judge "toxicity" in a player's words? Sometimes it is obvious: 1) using hateful speech or racism, 2) repeating a phrase like "wtf" over and over every time a person dies, 3) using profanity, and 4) hateful threats directed at players like "kys". However, it is my belief that in most circumstances people do not mean to be hateful or rude. Some examples of that are: 1) asking a question like "are you ok?", or "why did you just do that?", 2) making a one time comment like "wtf" as it could apply to anything, 3) mentioning that you plan to report somebody because they did something to you in game/pre-game and then explaining it simply and not imputatively, 4) making a joke like "QQ" as it could easily be them upset at themselves, ETC.

[2] So to what degree does the chat system work? I would love Riot to explain this further, but I think I have a good idea. All it takes is one report to be flagged in the system. Then the system uses an algorithm to syphon through the chat log associated with ONLY the reported player and determines if any predesignated phrases match a database of so-called "toxicity." If this chat matches anything in the database (regardless of context), then the account is marked for penalty. Once a threshold has been met for marks, a penalty is issued. I think, if this is correct, that this system has a HUGE problem regarding the effectiveness of the system. That problem being that CONTEXT has not been accounted for in situations like the ones laid out in section 1.

[3] "Context" refers to the way in which people speak to one another, a basis for psychological understanding of the human psyche. It is a common problem in language that written dialogue is misunderstood for the true meaning because human speech is filled with inflections, tone, intensity, etc. Without context, phrases like "wow" can be interpreted to mean anger, disgust, excitement, confusion, bewilderment, hatred, and so many more. In-game chats have this same problem. When a player asks, "should we report her for feeding?," they could be instigating an argument or simply asking a real question because they genuinely care about the game. When a player says, "I'm not gonna group because I don't wanna die," they could be refusing to help the team outright or they could be seriously concerned that the team's decision is a bad one. If a player says, "that's retarded," it could imply that a comment was not good, an action was done poorly, a person did something perceivably stupid, or even that they are just being facetious. Context refers to a number of situations that determine the meanings behind written dialogue in in-game chats.

[4] So where do we go from here? It is important that a game-maker has a basic understanding of human psychology and the differences between written and spoken dialogue. A reporting system should not sacrifice context for ease of function and simplicity of results if it directly affects the players involved. There are many players who get reported because another player misunderstood them or they just don't like them. To combat this, a reporting system should have very strict rules regulating which written comments, and how many in terms of repetition, are used for instant feedback penalties. Rather than retroactively correcting computer decisions based on algorithms, the reporting system should open a ticket for the game moderators to review and render decisions based on best intentions. Riot could serve the community better by implementing this plan of judgement, therefore preventing a multitude of unjust restrictions based on other players' "feelings" and "perceptions."

[Conclusion] The chat report system is exactly that, a CHAT report system. You can do whatever you want in game, but if you say anything (good or bad), regardless of context (and somebody gets offended by it), you will get marked and possibly penalized. At this point it is better to simply say nothing and play the game (whether you contribute to the team or not), effectively making the chat system pointless.

9 Comments

Fresh Untilted11/18/2017, 11:36:35 AM7 votes

The English language in particular, is full of syntax that can be used without being offending to anyone. If it's such a minefield to you, maybe you don't have to use it. However, it's very easy to recognize your own offensive syntax before you post it. If you get punished, they point out exactly what in your chat was offensive, and 99.99999% of the time, it is correct.

The chat system isn't useless. It's a great tool in solo queue for those who are fast and well-written typers. If you can write, "lets take their inhib, then do baron", then it's a very powerful shotcall in many situations. Nobody gonna punish you for saying that. Likewise, nobody going to punish you for saying "I'll come bot after clearing blue, wolves, red". That's useful information that your botlane can or can not use to set up their lane so that you can gank.

There's no need to tell someone you will report them, or say "why did you do that". Say "that's not nice" and mute them. Reporting is a personal thing, no need to broadcast it. All it does is agitate whoever you say it to, which is bad in 100% of situations.

I hope you'll eventually learn what a powerful tool the written language can be in a game like League. Until then, happy rifting.

Kei14311/18/2017, 3:44:17 PM3 votes

Repost from my GD post, cos I hate the GD. We can continue the discussion here.

Point [2] and [3] is the probably the main bone I can pick within your argument. Only hate speech / encouraging others to harm themselves works on a "word database" system. Saying such things typically net a punishment regardless of context, which I've seen a handful of recent false positive. The system does go through context for the rest of the non-extreme toxicity.

A point you missed is how the system punishes based on consistency x severity. It is true that you get a mark on your behavioral history for each game you validly get reported in, but it's not just a mark. It's a mark that grades your severity as well. "GGEZ" is an example that is on the lower end of the toxicity scale, but to get punished for that, you'll need to consistently do it for MMMMMMMMMMAAAAAAAAAAAAAAAAAAANNNNNNNNNNNNNNNNYYYY games (not to mention you need to get reported for it as well).

Speaking of consistency and going back to the days when the Tribunal was still around, when people saw that there was only 1 game of toxicity with the context of how a player was clearly getting tilted by the harassment of their team mates, majority of the community would pardon that player's actions. However, when the tribunal gives you 12 games of a player's counter flaming behavior, the community would overwhelmingly vote to punish that individual, as it showed that he consistently showed lack of control.

The IFS takes the long behavioral history into account. Those are information we can not see, and because we don't see it, we pass judgement on the system thinking it is not capable enough.

Now and then, there are some people calling for the Tribunal to come back, thinking the community will turn around the IFS' judgement cos they were only shown 1 game in their reform card. But if players were shown full details of their behavioral history, such as; [% of validly reported games in the past 3 months]; [% of games with X level of severity]; [# of games played in the past 3 months] etc, I think many would not have the same opinion on whether their punishment was fair or not.

The chat system / IFS system is not the thing that is operating poorly here. It is the poor transparency of the system that gives people the misconception on how the system works. If we were to change anything, it should be based around transparency, clarity and the education of the behavioral system.