The day this number reaches 100% will be on of those before/after moments. Considering the only problem with remaining 8 titles not running is DRM/anti-cheat, it is possible that we'll see this day arrive.
I usually make fun of "linux on the desktop" prophets by saying "Yes, 90% is great but problem is that everybody uses a different 90%". But 100% of the 50 most popular games is a different thing. It is a level of compatibility that you may not even achieve on windows depending on your hardware. Of course, performance and a few bugs will exist, but compatibility will be good enough for almost everybody.
I wonder if this coupled with steam deck will finally make supporting linux sustainable. I mean, a point where supporting is so cheap and the number of users is big enough that it becomes profitable to do so. If that happens, most excuses to not support linux will finally vanish away.
> Considering the only problem with remaining 8 titles not running is DRM/anti-cheat, it is possible that we'll see this day arrive.
Valve has confirmed they are working with Epic (and another vendor) to support EAC on at least Steam OS if not every distro but I'd imagine Steam itself would contain everything that'd be needed to run EAC. So if you can run Steam on your favourite distro, I think EAC should work as well.
It won't stop anyone who has the skill to author cheats themselves (i.e. someone with basic reversing and programming experience) but it does do a good job of detecting and banning players who are using widely distributed "commercial" cheats.
I look at EAC as being very analogous to a virus scanner, very easy to bypass if you know what you're doing, but good at catching the common and soon-to-be common threats. It also puts up enough roadblocks to make some novice attempts at cheating inconvenient at least.
Source: I deal with all the anti-cheat stuff (including EAC) on a semi-popular multi-player game. I can verify that it does make a very tangible difference to our player base in terms of the number of cheaters they are exposed to on a per match or per session basis.
EAC is actually pretty terrible against commercial cheats, it's only use is against free cheats pretty much. If you're willing to pay, you'll find many for EAC games.
Many major titles are using EAC currently, Fortnite, Apex Legends, PUBG, etc. There's a pretty large market for cheats in any of those titles. Just dropping "<game> aimbot" or "<game> cheats" into google is likely to turn up a multitude of commercial cheat developers, many of which are legitimate and will bypass the current anti-cheat.
Generally the legitimate commercial cheat developers offer status pages detailing any of their cheats that are currently detected and offer additional tools to do things like bypass hardware ID detection if you did get banned in the past.
Maybe I'm clueless but as I understand it some of these shooters (e.g. PUBG) are games that you pay for. If one of these commercial cheats gets detected by EAC isn't the result a permaban of your paid for account? And then you have to buy the game again and make a new one, if that's possible at all? Seems like that would be a serious deterrent, although obviously it won't stop everyone. (Encountering a cheater once out of every ten matches is probably acceptable. Encountering one in every other match probably isn't.)
If you are into cheating, new licenses are just a cost of business. Even a new, full price AAA game is about the cost of a round of golf. Furthermore, with free to play and microtransactions/ongoing ways to get revenue from players, there's a growing incentive to either give the game away (Fortnite et all) or offer the game at a fairly low price (eg rainbow 6 siege, can be had for as little as $5).
If you are willing to pay for a cheat then rebuying the game every few months when there is a ban wave is not that much effort. Especially if you buy it from some shady CD-Key reseller.
Do these anti cheat systems detect popular products of popular dev tools to create cheats typically? Or is the cheat world way beyond that phase and everyone just codes to the native OS APIs directly, making that hard to detect?
Is there still cheating in console players? Maybe with jail broken devices? Do these companies really have to think about cheating with consoles? Do you think the eventual future of these kind of games is some sort of locked down console-like system, which is what apple seems to slowly be going to with their hardware?
My warframe account was banned a while ago for having cheat engine running in the background. I was using cheat engine with total war, a single player series and never hooked it into warframe. So some systems at least just looks for running processes to create cheat tools.
In EU phone verification should basicaly prevent cheating as phone:person has almost 1:1 correspondence, so once you ban (or shadowban) there's not easy way to bypass.
No anti-cheat will ever completely stop cheaters. There are plenty of cheat tools that are basically impossible to prevent. The question is whether EAC is good enough? I don't know, I play very little multiplayer these days.
In the end, games need to be built better by sending the least information needed to clients, ie. zero-trust. This is a big part of things like CS:GO, Valorant, and RTS games like starcraft/League/Dota since they can implement a 'fog of war' and only network information if their client actually needs it[0].
The only cheats that still plague Dota are scripts that perform tasks automatically, such as disabling an opponent as soon as they're visible (beyond human ability), and those are taken care of by heuristics[1] and a community-voting-based 'overwatch' system in CS:GO and Dota 2[2].
The problem with zero trust and only sending what the player can actually see is lag. For example, sure, you could eliminate wall hacks by only sending positions of players that you have line of sight of, but that essentially eliminates any client-side movement prediction or other lag compensation and means you have to send each frames data quickly enough. This may work for some games that are less lag sensitive and as internet speeds improve, but its not currently a solution for most players.
Also most fps games use the sound of a player to indicate generally where a player is. You'd need the player xyz on the client regardless of line of sight of they are in aural proximity.
>games need to be built better by sending the least information needed to clients, ie. zero-trust
That isn't possible in real world for every type of competitive gameplay; trade-offs need to be made. Even games you list are not "zero trust" - things like looking direction are left to the client because of the latency.
Moreover, it's not enough. What's also needed is a hardware chain of trust for input devices. On-device cryptographic signing of mouse events would be sufficient and would guarantee that input comes from a fair play-compliant device.
SGX is a super-privileged, encrypted, and isolated enclave in the main CPU memory that can be used for anything. Signed mouse input is much more benign. But sure, I see your point - this is DRM, and it can be used to control the access to your own mouse, even outside of online gaming.
Still, it's the most logical step, and it would probably happen in several years, after Microsoft started demanding TPM 2.0. Valorant already requires TPM to be enabled on Windows 11 to run.
By the way, A4Tech has on-device DRM for more than a decade, they are using it to stop people pirating their software. Which is, ironically, designed for cheating.
> SGX is a super-privileged, encrypted, and isolated enclave in the main CPU memory that can be used for anything.
My position is s/"a super-privileged, encrypted, and isolated enclave in the main CPU memory that can be used for anything"/"a failed experiment in computer pseudoscience".
And the nightmare scenario I imagine is a more lubricated path to framing political dissidents by forging attestations that they were searching for child abuse imagery or some similar scenario.
Heuristics are great in theory but if they could be used reliably then you wouldn't need an overwatch system. Also the overwatch system depends on people being accurate judges of whether people are cheating which we know from pro players going undetected for so long (KQLY) and from pro players being accused for so long and being clean (flusha) that they are not.
On Steam Deck they will be using signed Proton builds to make EAC work. Which means you wont be able to use EAC on just any Linux distro unless you use the same signed builds.
What will a signed proton build do when a cheater can run as root or load a kernel module? I suspect they'll go much further and demand secure boot with vendor keys (no custom keys) and a locked down kernel.
Yes. You will likely need to be running a signed Steam install plus a signed version of the game. I’m 100% confident that it will be figured out in a repo like the aur, but it may be some work for Ubuntu folks and whatnot.
Archlinux steam user here: Steam auto-updates and Steam's default Proton install is pulled down by Steam and stored within it's paths.
Mostly this would affect anything that needs an unsupported build of Proton to work properly. Which is mostly lots of games that use the insane plethora of media presentation layers every big company has added to Windows over the decades.
From what I've seen of games on Twitch, EAC style games tend to not rely on FMV and are generally 'all live' so there shouldn't be many problems.
Check the chart again - ARK is native. If you run it in proton (many mods have issues with the linux-native version of the game, so many people do run it in proton), then you can't join battleye-enabled servers.
>I wonder if this coupled with steam deck will finally make supporting linux sustainable
I'd settle for devs trying to make games whose win32 api calls are within the bounds of that which Proton can handle. Why bother natively supporting Linux when you can architect something that'll run in Windows and Proton? Elsewhere in the thread I saw comments about how most indie stuff Just Works even now; I wonder if that's the subset of indie stuff that's, like, fairly paint-by-numbers projects in Unreal or Unity that aren't doing anything really unusual and thus are handily translated by Proton.
As a game developer who used to run Linux exclusively it's really frustrating to work full time on games knowing they don't work on Linux just because of DRM.
> Considering the only problem with remaining 8 titles not running is DRM/anti-cheat
This is going to be a huge problem. These corporations are going to be installing borderline malware into our computers. People will be forced to accept it if they want to play the games they paid money for. Invasive proprietary kernel modules designed to monitor your activities isn't something we should be supporting.
> There's no way I'm installing kernel-level proprietary monitoring software.
Yeah, me neither. It's gonna suck if they start requiring it on Linux though. I hope smarter people than me will find workarounds that allow us to pretend their shitty modules are running.
Just in case anyone is wondering why this is a big deal:
Cheaters aren't a strong enough reason to ship malware to users to monitor them. This sort of thing ought to be unacceptable.
This is 100% on the games industry. They engaged in an arms race with cheaters and now their ineffective solutions are worse than the problem they're attempting to solve.
Who decides that though? I would bet the vast majority of players would be very happy to be monitored if it meant less cheaters. Look at Valorant, that anti cheat is way beyond what HN would deem acceptable and the game is very popular.
We do. We used to have choice. Online games back then had the option to search for servers with no requirements for any of this bullshit. Now we're lucky if we even get dedicated servers.
> I would bet the vast majority of players would be very happy to be monitored if it meant less cheaters.
Honest players letting themselves be monitored does nothing to help matters. Cheaters will work around the system and evade monitoring.
> Valorant
> they scan all of the software that has loaded up on the individual’s computer
> Vanguard operates from system start-up, constantly operating on your PC unless you forcefully close it or uninstall it from your computer.
> Vanguard also operates a Kernel level driver, which has access to the entire computer system
And now it requires TPM and secure boot. It's disgusting really. It's like all that stuff RMS warned us about is becoming reality. All for what, a video game? No video game is worth this.
From what I'm reading it's not just me either. People are concerned about the kernel module:
> Can people trust a piece of software that has so much access?
> it appeared as if they were just brushing off anyone’s concerns about it
> The concerns people had weren’t exclusively from people scared of Riot Games or their parent company Tencent from performing “spying” operations on a computer and harvesting endless amounts of personal data.
> A noted subsection was worried that if the Vanguard kernel component was compromised, it could be accessed by people with less pure intentions than Riot Games simply wishing to prevent cheating in their title.
This has happened before with capcom.sys, it was literally a backdoor into the kernel that any user space program could use to execute functions in ring zero:
Play with trusted friends instead. Modding the game in these circumstances can even be fun. The online gaming model where you play with internet randoms is broken.
Haha, go tell that to the esport community :) honestly I feel like your comment is out of touch. You're basically telling me: don't play competitive games or games online with strangers. Some people actually enjoy these things believe it or not.
His reply does make sense. In the age of self hosted game servers you could ping an admin on irc/xfire and have him decide if someone is following you through walls etc etc. Sure he couldn't possibly catch every player being smart about aimbots but it was a remarkably good system because unlike now, there was a human element in all stages of the conversation.
If you were wrongly accused you could just move over to a different server with little downtime instead of pleading customer support for weeks that Process Explorer is NOT a game hacking tool.
Minimal anticheat, human support from your friendly neighbourhood admin is all you need.
> pleading customer support for weeks that Process Explorer is NOT a game hacking tool
Yeah, it's such a humiliating experience. They'll never believe whatever you say anyway, your only option is to sue them. One person actually did that in my country. Can't imagine it ever happening in corporation-friendly USA.
Pay hard earned money for a game only to get banned when their idiotic systems flag our software developer tools as cheats. Looks like you have a debugger running, citizen. We permaban people for that here. What's that? A virtual machine? You naughty user.
It wouldn't work these days though, cheats are much more subtle and harder to detect by eye now. If the player just shoots the head of every opponent on screen then a human being involved doesn't always help, they won't be able to distinguish between a great player and a cheat sometimes.
You act like you can have your cake and eat it too. The alternative is: No one buys the game (except a small segment of the population who are privacy-conscious enough to accept that the lobbies will be full of cheaters), and the game company declines to update the game on Linux, since no one is buying it anyway.
I have previously. I'm currently not seriously involved in a competitive game. Actual serious gaming competitions (the kind with prize money) take place under circumstances where cheating is immediately obvious and gets you disqualified in a public setting. I believe the "understanding" gap is on your end; ranking the integrity of systems above the integrity of gaming is a common sense stance before you realize you don't need to compromise one for the other.
At this point it's not a matter of game compatibility, it's a matter of anti-cheat. As I've spoken about many times before, Linux has a kernel module problem. Anti-cheat wants a closed-source kernel module that hooks into everything. Linux wants to disallow access to internal kernel state unless your module solemnly swears it's GPL.
Valve has said that they're working on this problem, but the workarounds one can undertake are pretty grim, and I doubt other companies (Blizzard, Ubisoft, Sony, EA, etc.) are all going to go down that sketchy rabbithole.
I'm not so sure. Steam is nice but what about other online stores like the Microsoft Store? I'm worried that Microsoft will push more integration to break compatibility on Linux.
With Epic having their own store, Apple, Google, Sony, Nintendo and the Microsoft budgets, the key question is how many game studio is Valve able to afford buying.
I usually make fun of "linux on the desktop" prophets by saying "Yes, 90% is great but problem is that everybody uses a different 90%". But 100% of the 50 most popular games is a different thing. It is a level of compatibility that you may not even achieve on windows depending on your hardware. Of course, performance and a few bugs will exist, but compatibility will be good enough for almost everybody.
I wonder if this coupled with steam deck will finally make supporting linux sustainable. I mean, a point where supporting is so cheap and the number of users is big enough that it becomes profitable to do so. If that happens, most excuses to not support linux will finally vanish away.