I don't think you understand what I was saying (though I'm not going to categorically deny that there may be small inconsistencies in what I say - my explanation below should be helpful).

I am very curious though . . . in real life - outside this forum - when someone says something that strikes you as unappealing or incorrect, do you go with the admittedly sophisticated "circle-jerk" criticism, or the more classical "yo momma's fat" approach? I personally prefer the latter style, as it's far more likely to wind up the opponent and thus prevent him from returning a cogent response.


You must acknowledge the distinction between real life, and an environment simulated in a computer. In real life, fish suffer from all sorts of sensory errors and tiny environmental influences, all of which together take the appearance of random "noise" in their behavior. Perhaps in principle, these factors are knowable, but in practice they are not. To approximate the influence of these factors in a simulation model, a stochastic variable is incorporated into the behavior of individual fish. The variable, usually drawn from a Normal distribution, is meant to represent that "noise" present in the real world. Thus, to extend the example in my previous post:

Preferred orientation of fish A = weighted average of (orientation dictated by information about resource + orientation dictated by position of other fish + orientation dictated by presence of predators) + random value from from a Normal distribution with mean 0 and SD=1.

To me this seems pretty straight-forward . . . random variables are incorporated into models as a short-hand for the tiny real-life influences that effectively add a stochastic element to individual behavior.

So, please tell me if there's still a contradiction.