Detecting Jackwagons in Online Games

griefing

Reddit Syndrome, The Eternal September, et al.

Counter Strike Global Offensive (CS:GO), a game I have been playing often since summer of last year (2013), is currently facing a dilemma that all online multiplayer games (and many social networks) face: as it grows in popularity, which is required to grow the monetary kick-back for developing and running the service as well as pushing the service’s features forward, the average level of player maturity decreases in proportion, to the point where older players who are used to playing with a more mature player-base will flee the game for some other outlet until this process takes over that one, and so on. It’s important to note that I’m not speaking about the skill of CS:GO players, since that is handled quite well by their Elo ranking system, but instead the maturity level, which means things like the level of racist voice and text chat, lack of statesmanship, etc.

Paul Graham’s Hacker News experiment is an attempt to solve this problem on the social news side of things. He writes about several of his reasons behind the choices he has made while running the site.

I wonder…

Is there a way to programmatically ensure that higher-maturity players do not intersect with lower-maturity players while not specifically removing the lower-maturity players from the player-base, since those lower-maturity players are required to keep the service growing?

My idea is that the service would have two or more pools of players, which would be kept secret from the player-base. My supposition is that lower-maturity players are “high-churn” in that they will likely not stick with the game for a great length of time and will instead switch their attention to some new game that arrives 3-6 months later. This “high-churn” player-base would essentially subsidize the higher-maturity players and game without the higher-maturity players ever having to intersect in game-play with them.

How do you detect an asshole, in code?

My guess is that this will have to be done in a similar way to detecting email spaminess: users will have a value between 1 and 100 for assholery. Being an asshole in online forums such as games is not binary (being either true or false) nor can any one action or decider change your state to true or false. So, it will have to be a collection of actions, over a given space of time, which will increase or decrease your assholery value.

Counter Strike Global Offensive offers a way for players to report users for griefing which offers one opportunity, though I’m not sure how much weight to put on it since it could be easily gamed directly by the assholes we’re trying to prevent.

A manual process is, at first glance, out of the question, since it’s not scalable. Thousands of games are on at any given point in a day. How could you possibly oversee them to identify assholes? Here, Counter Strike Global Offensive offers us a unique idea: Overwatch. As a developer, this solution smells bad because it feels like something we should be able to automate.

Perhaps a combination of encouraging users to not act this way combined with an Overwatch-for-Assholes system would reduce it.

I don’t have an answer

This problem is not going away and will only get worse as the gaming population grows.

Discuss on Hacker News.