I'm currently auditing a graduate/undergraduate seminar on the Fundamentals of Complex Adaptive Systems (CAS) that is cross-listed between about 6 different course prefixes. For the most part, the seminar is based around presentations by people conducting research on CAS. Today, however, the presenter had to reschedule and--instead--we spent the hour or so discussing what we've learned so far.
While there were some general questions, most of the time was spent discussing modeling and simulation. The biggest question was "What does it mean?", in a somewhat paraphrased manner. I also got the impression that most people in the seminar had a general idea of the concept of modeling and simulation, yet lacked an understanding of how to do one or the other.
To be honest, my attempts at mathematical modeling are rather simple and algebraic. I understand what they can do, I just have a basic level of calculus so the transformation from concept to creating the parameters in the modeling is a little beyond me for some forms. (One thing I intend to do--assuming time and resources--if I have a gap year between undergrad and graduate is to refresh and upgrade my math and/or statistics background.) So, for the most part, I prefer working on agent-based models in NetLogo, primarily because I'm interested in social behaviors.
I seem to grasp the concepts and just what the basic utility of modeling and simulation on an intuitive level--partially because of the amount of coursework I've had in the sorts of systems they use for examples. Another factor, I think, is a teenage and young adult history of playing war games like "Steel Panthers" and "Panzer Blitz".
What that form of war game has in common--whether played against a computer or another player--is the combination of hidden elements, heterogeneity of terrain, and interactive constraints for all the elements. While most people can cognitively say these factors are present, after a while most (successful) players begin to make predictive assessments based on the interaction between the factors in a rewarded (win) or punished manner (loss). The mental "conditioning" usually moves beyond the agent(simulation elements)/unit(game elements) toward sensing out how the interactions play out.
Doubly so with a human opponent.
It's unlike a conventional game like chess or Risk or even other computer games like first person shooters in the fact successful players integrate the partial knowledge they can know from the system/scenario with the knowledge underlying processes are at work unseen and are limited by the constraints of the model/scenario in such a way as to produce what they see. Additionally, the understanding of path dependence (whether by terrain on the map or the parameters set such as opponent's nationality and date) becomes a somewhat default process.
Additionally, for anyone playing a number of such war games, the concept of scalability and aggregative processes is also quite clear. What works at a "tactical" level doesn't necessarily show up at an "operational" level within such games. Instead, you learn to focus on the processes at that level and identify what traits do carry over when you rescale as well as what might have a changed effect.
(For WWII era gamers, German Tigers are loads of fun at a tactical level, but as soon as you start having to manage fuel and repairs in a game system you start seeing how the Allies won with primarily Shermans...)
Another use for the history of war gaming, is the simple idea of competitive situations. It may be a bias of mine, but many of the other students I talk to are more emotionally invested in the ideals of cooperation and being nice than they are in rationally trying to understand systems and/or people. This means they often are unwilling to see or insert conflict into their understanding of different systems. Additionally, when they do, it comes across as simple and "reluctant"--more as a "token" effort than an attempt at making a good approximation of an adaptive system.
For war gamers, the idea of conflict in systems is why you usually play the game. It doesn't help understand the "whys" beyond a basic level, but it helps you get in the mindset of "how" an opponent might game the system and to expect novelty in tactics used by human opponents*.
In general, it makes gamers look for options--given the system and parameters--in creative ways that intentionally run counter to the expectations of even the initially programmed intention. For the historically minded gamer, it also presents them with a better guess about path dependent effects as they ask "Why didn't/don't they do this in real life?".
That's what I was thinking about for most of the seminar...
*- For anyone who's played Steel Panthers 3, one of my favorite scenarios was a Russia vs. Germany battle set in 1999 against a human opponent. He brought Leopard 2's and a lot of infantry. I brought an insane amount to rocket artillery, light infantry as spotters, and engineers to lay mines. He essentially kept searching for my tanks until all of his units were bled dry.
After that, I expected more novelty on his part after my example and--thankfully--got it.
No comments:
Post a Comment