Valve’s perennial problem ever since Counter Strike launched has been dealing with cheaters, who end up ruining games for skilled players. Counter Strike has been one of the top games when it came to the number of cheaters in it, and that just goes to show the popularity and interest in the franchise.
Ever since the days of CS 1.6 cheaters have been rampant, and even the latest installment of the game, Counter Stirke: Global Offensive, has seen no respite. It is one of the top competitive eSport games in the world with the second-highest prize pool. As much as it is an encouragement for gamers to play better, it is also a cue to hackers and cheaters as well.
Over the years, Valve has tried to make lives difficult for cheaters with measures like, Prime membership and Overwatch system. But, according to Valve programmer John McDonald, Overwatch has a very low percentage (15-30 percent) of successfully detecting a cheater. This just means that most people hacking are getting away because other players cannot judge whether it is a legitimate situation.
Valve’s latest attempt at making sure that cheaters is using a deep learning neural network that zeroes in on possible cheaters and ban them. Speaking at GDC, McDonald said that the new system has been extremely effective, and has been able to convict people sometimes in 99 percent of the cases.
According to McDonald, the success of this method has been crucial in its adaptation in other games as well. He says, “We think that it is really helping us get developers off of the treadmill without impacting our customers in any way. Our customers are seeing fewer cheaters today than they have been, and the conversation around cheating has died down tremendously compared to where it was before we started this work.”