[cowtools]

Okay, but you couldn't just have a linux computer procedurally generating that video, then exposing itself to the host computer as a webcam over usb gadget driver?

Not to mention that you would have to comb through all this footage to detect cheaters... It is honestly a laughable solution.

[charcircuit]

Yes, but writing a program to synthetically generate correct images will take a while to come out in which players can play without cheaters ruining games.

[cowtools]

Creating a system to automatically verify correct videos will take a while to come out, not to mention that it is extremely invasive for the end user and requires that they own all this extra equipment and bandwidth.

The cheater does not even really need to generate fake video like I've described. Aimbots can be as subtle as the cheater wants them to be, offering <5% precision adjustments which won't be visible on a webcam. Not to mention that half of cheating is just information assistance like wallhacks which this doesn't even cover.

[mentos]

What do you think is a better way to validate the input is legitimate?

[majewsky]

The point is that the cheating problem is unsolvable at scale against sufficiently motivated attackers.

It would be possible to solve if the incentives are not that strong to cheat, but the status game associated with multiplayer games takes care of the "sufficiently motivated part".

It would be possible to solve at a small scale (e.g. at a tournament) by manually vetting hardware and manually reviewing footage.

But preventing cheating at scale against motivated attackers is so expensive as to be uneconomical. The devs will probably try to install whatever malware they can get away with in order to demonstrate to their shareholders that they care about it, but I'm guessing even the devs will be relieved when Microsoft just blocks their kernel-level malware because they know the risks it entails.