Reply to thread

Message

<blockquote data-quote="AbdulAlhazred" data-source="post: 9027156" data-attributes="member: 82106">Right, and I think that, whether via some actually codified, or uncodified, rules a sufficiently expert person CAN make determinations as to something like &quot;If 50 infantrymen rush through a 36 meter long field of corn stubble at dusk under enfilading fire of an MG42, then 37 of them will be casualties (give or take, add some dice if you wish)&quot; I'm a bit divided as to the degree to which this is a simulation, but lets imagine we can combine it with many other similar observations, and generalize the parameters such that we can predict the probable casualties from most similar situations. Lets assume these models are tested under real world conditions and produce reasonable results. I think we can call a resulting system, overall, a &quot;system for simulating small unit combat&quot; or something like that. This is very immediate cause/effect stuff, so its certainly doable. You can also construct logistics models that will tell you, based on past experience and some general parameters like &quot;how actively am I attacking along the front?&quot; and get out of that &quot;I need to supply 42 train loads of ammo to my forces every day.&quot; The remarkable thing though, people still cannot predict, even at coinflip levels of accuracy, the outcomes of wars! As for how that relates to rules, they will serve well in certain domains, but yet overall, when dealing with greater complexity, simply prove inadequate, or impossible to formulate in a way what is concrete enough to matter.</blockquote>

[QUOTE="AbdulAlhazred, post: 9027156, member: 82106"] Right, and I think that, whether via some actually codified, or uncodified, rules a sufficiently expert person CAN make determinations as to something like "If 50 infantrymen rush through a 36 meter long field of corn stubble at dusk under enfilading fire of an MG42, then 37 of them will be casualties (give or take, add some dice if you wish)" I'm a bit divided as to the degree to which this is a simulation, but lets imagine we can combine it with many other similar observations, and generalize the parameters such that we can predict the probable casualties from most similar situations. Lets assume these models are tested under real world conditions and produce reasonable results. I think we can call a resulting system, overall, a "system for simulating small unit combat" or something like that. This is very immediate cause/effect stuff, so its certainly doable. You can also construct logistics models that will tell you, based on past experience and some general parameters like "how actively am I attacking along the front?" and get out of that "I need to supply 42 train loads of ammo to my forces every day." The remarkable thing though, people still cannot predict, even at coinflip levels of accuracy, the outcomes of wars! As for how that relates to rules, they will serve well in certain domains, but yet overall, when dealing with greater complexity, simply prove inadequate, or impossible to formulate in a way what is concrete enough to matter. [/QUOTE]

Verification