What’s so funny? The whole point of machine learning is to go “here is a bunch of examples of A, and here is a bunch of examples of B. Given sufficient training, a machine can look at a new piece of inputted data and decide whether it is A or B with a high degree of accuracy”. In this case, “A” is botting and “B” is legitimate farming.
It’s my understanding that even without machine learning, this is accomplished to some degree in games like Runescape.
TheMohawkNinja:
there are 20 clients running from this one public IP
If you run an Internet cafe, then you probably aren’t using VMs, which would make it easy to differentiate you from a bot farm. As a result, you could apply for what would be a “commercial whitelist” of sorts that ignores your NICs HIDs when counting how many clients are going through your public IP. The bonus being it would act as a source of money for the game company.
LAN parties would be similar, however instead of a permanent HID based solution, you would have a temporary whitelist that basically says “you can have up to X clients coming from this public IP for the next Y hours”.
Admittedly, I did not know that. However the aforementioned solutions should solve this.
here’s the thing: anti-cheat design needs to err on the side of lenient with as few false positives as possible. The false positives from Vmware, spice based input, and vbox seem acceptable to the devs, because these technologies aren’t 3d accelerated, so the only use case is probably for farming automation or botting.
Checking for virtualization is far less complicated than analyzing each client’s traffic or in-game behavior, and also has a much lower computational overhead.
mainly the resource overhead making it completely unjustifiable over simpler means, and the likelihood of false positive shitstorms once the meta changes or the process is left unsupervised too long.
a) you’ve answered your own question, and b) not necessarily, see below
asian game cafes run massive LANs, some with deployments in the thousands, or using game streaming on virtualized instances to reduce hardware costs. This would be seen as draconian, as well as bad business over there, and china/japan/taiwan/SE asia are some of the largest gaming markets.
spawning and correctly assigning these would be uneccesarily complex, and again, add additonal overhead to each client.
Yeah I suppose that is the biggest nail in the coffin here.
I’m just annoyed (if you couldn’t tell lol), as I had this wonderful idea to be able to play videogames on Win10 while using Arch Linux as my main overall OS without the downtime that a dual-boot system would create.
RDP is wholly unusable due to compression, synergy would work from a “i get to use Linux and play video games on one mouse and keyboard”, but has the akwardness of not being able to move things around easily, and while I initially thought LG was the holy grail I was looking for, it turns out modern anti-cheat bans them.
I mean, Wendel did a great job of showing that gaming directly on Linux is MUCH farther along than I previously understood it to be, but even he noted that it’s still only mostly there, and I’d rather not have to deal with “mostly”.
TheMohawkNinja: