Counter-Strike: Global Offensive, like many other online games with competitive modes, is filled with cheating players.
Cheaters pose a problem for other players and developers alike, as legitimate players will lose interest in a game where playing fairly often leads to a disadvantage, and a dwindling number of users is not what any developer would want for its game.
For Counter-Strike: GO developer Valve, it seems that it has found a new ally in the fight against cheaters in the form of artificial intelligence.
The Current State Of Anti-Cheat Systems
Most of the current anti-cheat systems in games rely on players reporting cases of cheating, which is labor-intensive for developers as a team would need to sift through the reports to determine who are cheaters and who are simply skilled. In addition, if cheat detectors would be implemented based on the reports, cheat developers will be able to circumvent the detectors to have the cheating tools slip through until the next detector is set up.
This is the so-called "arms race" between Valve and cheat developers, and it never ends as cheat developers continue working on their tools. This was brought up in a post on the official Counter-Strike: GO subreddit in a discussion on spinbots.
Valve Taps AI In Fight Against 'Counter-Strike: GO' Cheaters
A spinbot is a cheating tool that spins characters really fast to give players a 360-degree field of view in Counter-Strike: GO. The cheating players look like they are glitching, which makes it very difficult to land a shot on them, and when a spinbot is paired with an aimbot that makes all fired shots hit onscreen targets, these cheaters essentially become unbeatable.
Valve, in the Reddit post, claims that the best approach for the detection of spinbots is through machine learning, which will allow systems to receive continuous training to detect the difference between cheaters and skilled players.
"The process of parsing, training, and classifying player data places serious demands on hardware, which means you want a machine other than the server doing the work," the official account of Valve Anti-Cheat posted on the Reddit thread. The post added that with the need to monitor matches from the perspectives of all 10 players in a match, and with more than a million Counter-Strike: GO matches played daily, there is a need for a separate data center powered by "thousands of CPU cores" to do the work.
Valve, however, notes that it has already started doing the work to implement machine learning in anti-cheating systems, with an early version already deployed and already submitting cases to Overwatch, Valve's anti-cheat system and not to be confused with Blizzard's popular multiplayer shooter.
The developer says that the results of tapping artificial intelligence to detect cheaters in Counter-Strike GO have been promising, and that work will continue to improve and expand the technology.
Valve In The News
Valve has been making the headlines recently, with some of the news centered on its move to finally fix a 10-year-old bug in Team Fortress 2 and its current lack of interest in making games for consoles.
Valve president and cofounder Gabe Newell also recently revealed that some company employees have themselves been releasing hoaxes regarding Half-Life 3.
