[Kigen]

Well, it was purposely built not to be that random from the clients. Basically, if client A fired it would take a few milliseconds for that command to reach the server, but client A's operator expects that done instantly, so the client program shows a bullet hole, or blood animation instantly.

Basically, the server and clients share the same seed for their random number generator.

[The Unknown]

oooh... the good ol' tribes days

best game ever made (No joke.)

[bug]

Originally Posted by Kigen Its probably similar to "no spread" type cheats. HL randomizer isn't really all that random to clients, since it uses it for legit prediction reasons. (Bullet hole placement, Crit annoucements, etc) I haven't disassembled TF2 or in any way looked at the code but that's how I guess its possible to have 100% crits. (delay firing until a predicted crit comes up)

Yes, now you got it! It is random, and if you fire at the precise same time as this random generator thingy says "you got a crit!" then you got a crit. What hacks can do is synchronising the shots with this random generator because they can read it. Because of the lag reduction it’s needed because otherwise you would not even know that what you shot was a crit until later when the server says so and it will appear laggy, but it would make cheating crits impossible if it was all server side (which it is in practice, unless hacks are used). The (supposedly) random crits has nothing to do with how well you do. There could be extreme cases of luck when you time the crits more than usual, but if you keep track of it should even out unless hacks are used.

Here is an example. http://www.youtube.com/watch?v=t7Oig-OTMs0


Now those cheats really isn’t that bad (for gameplay they are of course) because they are easy to spot if you are looking. The thing is that they will or already have created cheats that will be a lot smarter and only uses the crit trick maybe 10% (that is two times the amount of crits other players will get) of he time or only when a key is pressed or maybe even for melee only since melee crits so often, making it undetectable unless you keep track of players amount of crits for maybe a few rounds. Crits are completely random and occur 5% of the time with the exception of some new weapons that should not be counted and the short stream of crits due to damage to the other team.

[Kigen]

Yeah, its possible to break such cheats by obviously making the server not pass the correct seed. Though I'd only suggest doing that in LAN or low-ping environments where the predication isn't so heavily relied on.

I have some alternate ideas, but for the time being I'll keep them to myself.

[dirtyjob]

hows it goin on KAC v2 ?

[Kigen]

Its going, but it of course takes time.

[dann]

after a few hours of the server being full I seem to start getting roughly 5 second freezing lag spikes when players join when I have this plugin installed with fake clients.

If I remove this plugin I can run for days without any problem.

thanks

[Derek]

Don't use fake clients.

[Derek]

I found the problem when the map turns dark.

The problem is that most custom maps set mat_fullbright to 1. The problem here is that this is a cheat.

When you execute config files that contain sv_cheats 0 in them, it will turn the map dark as it has set mat_fullbright to 0. This is why when executing some config files or your anti cheat (ones that didn't have sv_cheats 0 in them) it had no effect.

The fix for this is to remove "sv_cheats 0" from any config files that you may want to run during gameplay. It is probably also safe to remove it from server.cfg also as I would assume the default value would be 0.

We will also need a fix for your anti cheat to work with custom maps that use the variable.

[Kigen]

Well, sv_cheats is restricted from being changed from 0 by KAC since this version relies on CVARs to do its work.