Most of the other kernel anticheat rootkits are more polite and don't run until a game starts them, at least.

In general having a bunch of random stuff from video games running with kernel permissions is spooky. You never know whether someone will find an exploit in one of them, or if it will cause performance, stability or data integrity problems.

Honestly, if I was responsible for building an anti-cheat system this is the way I'd go about it. Anything that can run that you don't know about (e.g. because it ran before you started) is a potential risk.

Aren't there some security benefits from trying to confine apps to user land?

Some developers aren't even denying that it's nasty, but excuse it with "no one cares anyway" or "we already could snoop on you, so don't complain if we make it even worse" which is disgusting.

Example: https://na.leagueoflegends.com/en-us/news/dev/dev-null-anti-...

Server side enforcement is not only heavy, but has much of the same problems as the client side. It's running on a layer even higher than the client side anti-cheat. How can you tell if someone is running an aimbot or just has good aim? Sure if they do something outright impossible you could ban them, but heuristic approaches are going to have problems with false positives as well as false negatives.

> How can you tell if someone is running an aimbot or just has good aim?

That's the point. I don't care. As long as the game is enjoyable. And I'm sure AI can be gradually trained to detect cheating in more sophisticated manner do differentiate non human from human patterns. Not a trivial issue, but something they should invest money in as above, instead of rootkits.

  14:22:01 *Shmerl has joined the game
  14:22:02 [DefinitelyNotAnAimbot] headshots Shmerl
  14:22:05 [AnotherTotallyHumanPlayer] headshots Shmerl
  14:22:07 [DefinitelyNotAnAimbot] headshots Shmerl
  14:22:07 DefinitelyNotAnAimbot is on a killing spree!
  14:22:10 *Shmerl has left the server

multiplayer game have the choices: a) ship game with malware/trojan to try to prevent cheats, or b) open up for cheats (even if client side only, like see trhu walls)

Most popular games chose option A (and still fail preventing both client side and remote hacks)

here is another idea: proper ranking of player abilities! crazy right? something so simple should be a day-one feature right? that would put all the hackers together and they could even compete in their own hacking league... 100% of clients satisfied!

now think about game companies failing to put that single feature to work correctly, what changes do you think a AI that detect human behaviour have or working well in that space? :(

I have too many games in my library to worry about not being able to play a couple more, I guess. But, if I could, then I would. So I just stick with whatever can run with proton.

Keep in mind that many people are interested in single player or local multiplayer that do not require anti-cheat solutions. They will hardly ever run into problems on that front.

I suspect the largest problem with Proton is that support will be biased towards recent popular titles based upon common game engines. Replicating all of Windows is not practical, Valve themselves will be most keen on supporting games they can generate revenue from, and community contributions will come from people who have an interest in getting particular titles working as well as the technical skills to contribute.

Can’t say the same for many other anti-cheat games though.