Steam games will need to disclose kernel-level anti-cheat on store pages • Hacker News Board

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#581

Post by jakebasile » Thu, Oct 31, 2024, 4:53 PM UTC

xvinci wrote: post #575

jakebasile wrote: post #534

Earlier quoted context omitted.

We had the best alternative decades ago. Let the community run the servers and ban cheaters while allowing individual servers to form their own culture and community. The obsession with matchmaking and games as a service (requiring publisher run servers) is what painted the industry into this corner. Note that I like matchmaking, specifically skill based matchmaking, in some games and at some times but completely end…

I was pretty good at CS:S, semi-professional level. I nevertheless still enjoyed hanging out on public servers with friends. I cannot tell you how often I have just been banned. Another example: I was an anti-cheat admin in a major league about 20 years ago. I am quite confident a lower double digit percentage of banned players were innocent - it was simply too hard to get enough competent people for doing manual che…

As I said, offering both is the ideal. I'd rather have some false positives if it means a way forward without kernel rootkits becoming the new normal.

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#582

Post by Kapura » Thu, Oct 31, 2024, 4:57 PM UTC

I know there are large multiplayer shooters that have or are going to be moving off of Linux soon. Anecdotally, the percentage of linux users running cheat software is significant, north of 50% in some cases.

Ultimately, I sympathize with game developers trying to create a good, _consistent_ experience for players across multiplayer titles. The reason players accept anticheat software in large mp games is because the alternative is worse.

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#583

Post by Retr0id » Thu, Oct 31, 2024, 4:59 PM UTC

nathants wrote: post #236

kernel level anticheat is not enough. client inputs have to be trusted, and there is no provenance. the kernel has no visibility of inputs. i’m shipping a 100 player matchmaking game now. clients tick at 360hz, server ticks at 120hz. fair up to 60 ping, which covers entire continents. servers are metal, not vms. epyc 4244p with 2Gbps egress, 1 server per 15 minute game. mitigations=off and nosmt on all clients and th…

I'm booting your OS in a stealthy hypervisor and operating my cheat logic from there.

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#584

Post by BlueTemplar » Thu, Oct 31, 2024, 5:00 PM UTC

dbrueck wrote: post #310

Gamemaster1379 wrote: post #257

Earlier quoted context omitted.

> game companies are highly incentivized to work with (or at least not antagonize) the elite players Actually, this is generally untrue. Companies BELIEVE this but often times, these players are a vocal minority put on pedastal and they often end up making the game worse for the general player base.

Sorry for not being more clear, I was referring to the advertising or promotion that comes via the elite players. Take Valorant, for example. Riot Games leveraged their League of Legends user base and gave early access to high-end players and that apparently played a big part in helping its popularity take off. Now it has a robust presence in eSports, again helped by the high-end players. It's not uncommon now for po…

Are they popular because they are the best, or because they are entertaining ?

I wouldn't discount those mediocre (or even outright bad) at the game, but moving huge audiences...

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#585

Post by LelouBil » Thu, Oct 31, 2024, 5:00 PM UTC

Kapura wrote: post #582

I know there are large multiplayer shooters that have or are going to be moving off of Linux soon. Anecdotally, the percentage of linux users running cheat software is significant, north of 50% in some cases. Ultimately, I sympathize with game developers trying to create a good, _consistent_ experience for players across multiplayer titles. The reason players accept anticheat software in large mp games is because the…

I think I saw on HN a while ago someone suggesting that Valve could make a "clean kernel" build that anti cheats could whitelist for Linux/Proton games.

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#586

Post by BoorishBears » Thu, Oct 31, 2024, 5:03 PM UTC

Ajedi32 wrote: post #500

BoorishBears wrote: post #485

Earlier quoted context omitted.

Are you seriously asking me how getting killed by someone cheating to know exactly where you are feels different than just being outplayed?

Actually yes. If you've never been falsely accused of cheating in an online game by people who can't tell the difference between that and getting outplayed, chances are you haven't played any multiplayer games long enough to get any good at them.

You'd end up pretty awful at games if you treated getting killed by something that has no counterplay the same as getting killed by something that does...

A person who's good enough to know where you are for in-game reasons is something you supposed to learn to respond to: They aren't cheating so they're using heuristics and information that you're supposed to learn to counter play.

A person who has omniscience at the press of a button does not play similarly at all. Often times that's how they're even caught: since they're not bound by any sort of in game heuristics no counterplays work against them.

-

In some games that even makes cheaters a liability when they're on your team. With any ranked game mode "soft cheating" will result in them climbing to a rank where non-cheaters have all sorts of great heuristics on how to play that they lack. So without ragehacking they're actually at a disadvantage and either need to get increasingly bold with cheats or lose.

In some games that's transformed into another way cheaters ruin modern competitive games: intentionally de-ranking so they'll play against less mechanically sound players who are easier to cheat against.

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#587

Post by steveklabnik » Thu, Oct 31, 2024, 5:04 PM UTC

bunderbunder wrote: post #559

So, I'm not really into online multiplayer, but from a distance I keep wondering if this could be more easily mitigated by altering the incentive structure? What if online games track how well people do and sort them into tiers based on skill level? And then put people who are roughly evenly matched together. I am guessing that cheaters will naturally end up clumping together with each other, and maybe a smattering o…

This is called "Skill based matchmaking" and it is absolutely something that many games already do. Studies have shown that it increases player retention significantly. But also, there's a chunk of people who absolutely hate on it at every opportunity.

Cheating happens regardless.

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#588

Post by bigstrat2003 » Thu, Oct 31, 2024, 5:04 PM UTC

davrosthedalek wrote: post #303

froggit wrote: post #295

Earlier quoted context omitted.

It's quite literally kit running as root...

It's quite literally not. Root is technically a user with extra rights (including modifying the kernel, but there is still an API the root user has to go through). This is running as part of the kernel. It's not running in userland "as root". A rootkit is something that gives other users the power of root.

That's taking it too far. By that logic, rootkits can't exist for Windows because the super user is called "Administrator", not "root".

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#589

Post by consteval » Thu, Oct 31, 2024, 5:04 PM UTC

ethbr1 wrote: post #322

sunshowers wrote: post #217

Earlier quoted context omitted.

While all of what you're saying is true, I think it is worth noting that historically a large chunk of this problem was solved by communities hosting servers. I agree that in the matchmaking era, remote attestation via kernel-level anticheat is the inevitable solution that you converge to after a few iterations. And yes, servers would often kick out people who were too outside of the general skill level, even if they…

> historically a large chunk of this problem was solved by communities hosting servers Yes and no. I lived through that era too, and there are serious scaling problems: at some point, trying to banhammer griefers with rotating IPs becomes a full time job, and then the public servers turn into a dumpster fire.

Solution: submit them for an account ban.

The games that have the most cheating either:

1. don't do account bans

2. don't limit account creation

You can trivially limit account creation by just charging money for the games.

Re: Steam games will need to disclose kernel-level anti-cheat on store pages

#590

Post by 0x457 » Thu, Oct 31, 2024, 5:05 PM UTC

screcth wrote: post #242

0x457 wrote: post #228

Earlier quoted context omitted.

Where did I say they are the same? We have a kernel-space thing (anti-cheat or gpu driver) and a user-space thing ((a game actually talks to both) that talks to a kernel-space thing.

I understood that you were making an analogy between installing a piece of hardware and its associated kernel driver with installing a game and its associated kernel anticheat. When you install a hardware device you are trusting the manufacturer with full access to your machine, so installing a driver does not give them any more powers. You have already "unlocked the door". When you install a game that runs on user s…

> When you install a game that runs on user space you are not trusting the vendor nearly as much as you are trusting a hardware manufacturer.

I'm not sure where this trust comes from. I absolutely do not trust any hardware vendor. I just have no choice here.