alt Hacker NewsI think from a purely technical viewpoint, cheaters will always have the advantage since they control the machine the game and anti-cheat is running on. Anti-cheat just has to keep the barrier high enough so regular players don't think the game is infested with cheaters.
Replies
I have never worked on AAA games, but I have developed software for 35 years and play many competitive online games regularly.
I have always wondered why more companies don't do trust based anti cheat management. Many cheats are obvious from anyone in the game, you see people jumping around like crazy, or a character will be able to shoot through walls, or something else that impossible for a non-cheater to do.
Each opponent in the game is getting the information from the cheating player's game that has it doing something impossible. I know it isn't as simple as having the game report another player automatically, because cheaters could report legitimate players... but what if each game reported cheaters, and then you wait for a pattern... if the same player is reported in every game, including against brand new players, then we would know the were a cheater.
Unless cheaters got to be a large percentage of the player population, they shouldn't be able to rig it.
> Anti-cheat just has to keep the barrier high enough so regular players don't think the game is infested with cheaters.
And even that's the (relatively) straightforward part. The hard part is doing this without injuring the kernel enough that the only sensible solution for the security conscious is a separate PC for gaming.
I agree, but that’s precisely the interesting ‘technical’ problem. Like bitcoins “proof of work” in 2011 (it took me few years to comprehend) was an eye opening moment for me. While I do believe that it firmly failed to achieve its lofty goals, the idea of “proof of work” was a really captivating and interring technical idea. Can a video game client have a similar zero-trust proof of their authenticity? I personally can’t think of one. I can’t think of a way to have remote random agents (authenticates or not) to proof they are not cheating in a “game”, and like you, I suspect it’s not really possible. But what does that mean?
I grew up with star trek and star wars wondering what a “I’ll transfer 20 units to you” meant. Bitcoin was an eye opener in the idea of “maybe this is possible” to me. But it shortly became true to me that it’s not the case. There is no way still for random agents to prove they are not malicious. It’s easier in a network within the confines of Bitcoin network. But maybe I’m not smart enough to come up with a more generalized concept. After all, I was one of the people who read the initial bitcoin white paper on HN and didn’t understand it back then and dismissed it.