3 ms·

Still, I'm not an expert and this might be entirely unfeasible but I can totally imagine some kind of trusted execution environment that cross checks specific i

Still, I'm not an expert and this might be entirely unfeasible but I can totally imagine some kind of trusted execution environment that cross checks specific instructions at the API level when the client communicates with the server using some kind of trusted operation (backed by cryptography). There are already VMs that are used to run trusted instructions on a target (read: client) machine and regardless of you knowing how the VM works and what the code does, if you don't "play by the rules" (as is the world of cryptography), your data will not be accepted and you will get kicked/banned/disconnected. Is it computationally expensive? Probably. Is it feasible? I don't know. Is it theoretically possible? I'd say so. >Increasing the time and skill needed to reverse engineer your anticheat increases the time it takes for someone to develop a cheat. This is the good ol' security by obscurity fallacy. Having it open increases the amount of people who have eyes on the code and contribute to the code and finds exploits to patch, etc. It's two sides of the same coin.

charcircuit 4y ago
A cheat would just look at the detection code and be sure it doesn't violate what the anticheat is checking for. >Having it open increases the amount of people who have eyes on the code and contribute to the code and finds exploits to patch, etc. If your detection is to collect the names of the open windows on a computer to check for a cheat's name and then send the anticheat server the hash of the binary that opened that window how is making that information public making the anticheat stronger. The more eyes that look at that code the more cheaters who know to name their cheat windows as "Notepad." To avoid being found. If you find an exploit there is no patch.
- hypertele-Xii 4y ago
  Nothing you said has anything to do with security nor preventing cheating. If that's the sorry state of modern anti-cheat then it makes sense why TF2 is still plagued by incredibly obvious sniper bots. Detecting cheating by window names? Who is getting paid writing this crap?
  - charcircuit 4y ago
    >Who is getting paid writing this crap? I'm not sure what you are expecting. Either you try and locate cheats on the machine / in memory via some heuristic or look for software that is acting suspicious like trying to write into your game's memory.
    - hypertele-Xii 4y ago
      No, you analyze behavior on the server. Look at what the player is doing. Track their movements. Graph their performance. Look for hyperconsistencies (inhumanly consistent success).

charcircuit 4y ago

A cheat would just look at the detection code and be sure it doesn't violate what the anticheat is checking for. >Having it open increases the amount of people who have eyes on the code and contribute to the code and finds exploits to patch, etc. If your detection is to collect the names of the open windows on a computer to check for a cheat's name and then send the anticheat server the hash of the binary that opened that window how is making that information public making the anticheat stronger. The more eyes that look at that code the more cheaters who know to name their cheat windows as "Notepad." To avoid being found. If you find an exploit there is no patch.

hypertele-Xii 4y ago
Nothing you said has anything to do with security nor preventing cheating. If that's the sorry state of modern anti-cheat then it makes sense why TF2 is still plagued by incredibly obvious sniper bots. Detecting cheating by window names? Who is getting paid writing this crap?
- charcircuit 4y ago
  >Who is getting paid writing this crap? I'm not sure what you are expecting. Either you try and locate cheats on the machine / in memory via some heuristic or look for software that is acting suspicious like trying to write into your game's memory.
  - hypertele-Xii 4y ago
    No, you analyze behavior on the server. Look at what the player is doing. Track their movements. Graph their performance. Look for hyperconsistencies (inhumanly consistent success).