You’re talking about running everything other than graphics rendering (not graphics compute mind) and UI input server side, including physics?
Apart from the fact that this would be inherently expensive as you mentioned, it would also require extremely fast internet with extremely low ping or jitter.
It would also not prevent aim botting, bots in general, or input related exploits like macros and “turbo mode”.
I’ll just reiterate what they said. There is no such thing as a perfect anticheat. It is impossible.
Nah you just do heavy statistical analysis on data that already gets streamed from the client.
Here’s a better explanation than I could provide:
https://www.i3d.net/ban-or-not-comparing-server-client-side-anti-cheat-solutions/
What they could do is set a player surveillance system that tracks game by game averages on hundreds of different metrics like critical hit accuracy and prehit mouse acceleration and compares them to a baseline and any time any player stat moves past the average the system will increase their scrutiny level and perform more advanced analysis on them.
Another thing that could happen is the server could submit ghost data to suspicious clients and honeypot the cheat software into reacting to it.
You could also train an ML model on your game to watch highly suspicious players.
There are already hardware solutions that can bypass those. Utilize video output from a PC, interrupt user input to react, add in humanesque delay. There’s no reason cheat software has to be as blatant as Minecraft auto fight cheats.
Can they be caught? Sure. Human pushes the limits too far, manual checks to see if you react like a human, etc. but you’re talking about a 100% effective anticheat, and I’m telling you there is no such thing.
It’s like claiming you can make an internet connected device that is “unhackable”.
