Comments on: Anti-Cheat Systems (Some Basics And Ideas…)

By: Vyross

Vyross — Sat, 18 Apr 2009 21:53:32 +0000

@ Ezequiel:

that’s exactly what they do in Left 4 Dead. a cheating/flaming/hacking/otherwise undesirable person can be banned based on vote of team members. you might think “lets keep this guy in, because he’s winning for us” but on the other hand “he’s hacking which makes us all look bad” counter balances that.

it works very well actually

By: Mar

Mar — Sat, 18 Apr 2009 05:14:06 +0000

Someone I know was banned under VAC, but did not cheat. The complicated part is that Valve will not address the issue with the consumer. They will only cut and paste a form letter saying that you are guilty. You cannot defend yourself at all. They will not admit that their system is not perfect.

If you wish to implement an anti-cheating system, prepare the customer service end of things and be prepared to be on the customer’s side sometimes. Allow an investigation to happen. Communicate openly and transparently, and above all, not in form letters.

By: Ezequiel

Ezequiel — Fri, 17 Apr 2009 18:59:12 +0000

Why don’t you give some in-game ways for the players to handle this by themselves?

Instead of them relying on you (the game creator/game admins) to handle the issue you could give them tools to ban/kick people for cheating. Maybe a player proposes a ban, and once some do the same a poll comes up on all nearby players to vote for the ban. And if the ban is issued, the user gets banned.

That’s just an example.

By: Chris Kugler

Chris Kugler — Fri, 17 Apr 2009 17:41:52 +0000

I used a process at a former job where we did a CRC check to get the hash values of every file we distributed and stored them on a server. Then, when the client is launching it performs the same CRC check and makes a webservice call to send the hash values to the server. The client cant proceed until the server responds that the codes match.

By: Sargon

Sargon — Fri, 17 Apr 2009 17:09:23 +0000

As I understand, the job of the client is only to visualize the data the server sends him, and send the input by the user.
That is, you need to represent the visual aspect of the game as data.
So the server will only tell the client, you see a human at position X,Y.
If the user is far in the desert, and in the enemy base there are enemy troops, the user does not need to know about them, because he doesn’t see them visualy.
So a sophisticated system could simply not give the information about characters unseen by the user player, like a character behind a wall.
The problem might begin when the enemy character is only partially hidden.
Or when the client extrapolates, he might be in a position he should see the enemy already, but he didn’t get that information yet.

By: Juuso Hietalahti

Juuso Hietalahti — Fri, 17 Apr 2009 13:35:50 +0000

But… think about somesort of really fancy hacks that make walls invisible or something (since it’s rendered by the client’s computer it’s technically possible)… in theory, there’s always some way to cheat even with fancy AC systems :)

By: Sargon

Sargon — Fri, 17 Apr 2009 08:36:02 +0000

Hmm, I think some of the cheatings can be completly prevented, depenging on the game.
For a multiplayer WW2 FPS with a client\server scheme, you can make the server limit “super human” abilities.
You can make the server limit players movement to only 1 meter per second?
Or you can limit the crosshair rotation speed, like, you can’t rotate 180 degrees in less than half a second.
Things like that.
Other than that, there is punk buster, that I think checks that the client software isn’t hacked, somehow.