[bf1942] Anti cheat tool

Scratch Monkey ScratchMonkey at SewingWitch.com
Wed Nov 19 15:09:17 EST 2003


--On Wednesday, November 19, 2003 9:24 AM +0100 Andreas Fredriksson
<andreas.fredriksson at dice.se> wrote:

> Excluding data from the client simulation on the server would be a bad
> design decision for BF because of our complex vehicle support; if the
> client and server were not in sync when an enemy vehicle entered the
> client frustum it wouldn't look the same on all connected clients, and
> the physics code would kick in from there, much as if you come with in
> a certain distance from an AI bot (they LOD).
> 
> Also, some vehicles are very fast compared to foot soldiers, and doing
> special cases for these would become cumbersome and wildly inaccurate
> because we can't know how the vehicles modders put in would behave in
> these conditions.

Is this just a velocity/acceleration issue? If so, the calculation to only
send data to the client for objects that will enter their fog sphere before
the next network update shouldn't be too complex. (Why "frustum"? Nice word (I
looked it up at m-w.com) but I don't see how that fits here.) Integrate max
accel with current velocity and current position and assume the vector points
at the client's position (with similar correction) to handle the worst case of
the client and object accelerating directly towards each other. No need to be
fancier than that.

I could maybe see a problem with projectiles, but you can always send data on
those no matter the distance. That would put stationary shooters at a
disadvantage.



More information about the Bf1942 mailing list