Yes, this is the expected behavior, but I think in general we need to implement a paralysis cap of around 2 minutes. When counting down paralysis, if it's above 3600 steps it would set itself to 3600, etc.
how much extra processing would it take for dimishing paralysis returns? a simple example would be paralisis added =x*(m*x/(m*x+n*P) were x is paralis time for the attack and P is how long the target is currently paralized for, and m and n are wieghting factors. So if both m an n are 1 then geting hit by two 30 sec stuns would provide 45 seconds of stun. or 15 simultanous emp mines would cause 2 minutes and 17 seconds of stun. If you increase m then the returns will dimish slower, if you increase n they will dimish faster. and for somethin which you would want unchanged from the current (for example emp missiles) then set n to zero
Well, look at all those multiplications and divisions right there. That's an answer for you right there, if there is a lot of paralysis going on its fairly notable. Everything happens in aggregate in AI War, so it's not just that simple formula, but that 1000+ times per second (potentially) in big battles. Even that's not going to be killer, but the main thing is that it doesn't add enough to the simulation to be worth the extra cost that it does bring. There are hundreds of little calculations in the AI War simulation, from everything to protection area calculations, to collisions, to targeting, to movement, to all the other little randoms stuff that has to happen for all the specialized units (engineers, paralyzers, etc, etc). The way that I keep the game running so smoothly overall (or running at all, in general) is by figuring out where to shave the corners off. This is one of those classic cases where doing something simpler gives pretty much the same general effect unless players are really looking hard at it, and so to me it's an clear case for going the simpler route.
