RE: CV and Combat Modeling (Full Version)

All Forums >> [New Releases from Matrix Games] >> Gary Grigsby's War in the East Series



Message

moses -> RE: CV and Combat Modeling (12/7/2010 5:06:26 PM)

A greater problem is getting people to agree to the very large number a formula that would be required.

For the specific example he gives (German Mech Bn vs. Sov. Inf co. with T-26 co.) Does anyone have actual data? If we had a whole lot of data under various conditions for this specific match-up, we might be able to settle this one conclusively. Otherwise we are speculating and people would surely debate any results which we assigned.

There are probably millions of possible tactical combinations. Some of these would include elements which never faced off against each other IRL and for which no combat data exists.

The designers do the best they can and we can agree or disagree. But we should probably cut them some slack as the problem is ultimatly unsolvable.



Flaviusx -> RE: CV and Combat Modeling (12/7/2010 5:36:55 PM)

I don't have access to the in game combat formulas, and they are rather complicated, but the game doesn't just take aggregate strengths on each side, compare them to one another, and generate a result. There's a fire procedure involved, and leadership, morale, experience and terrain effects get calculated at the unit element level for each stage of the combat procedure, which doesn't involve all the elements at once. Unit elements can indeed be disrupted and damaged before their point in the firing procedure arrives. Failed leadership checks can result in significant variances in combat results, too.

Pavel can explain this better than I can (or Gary for that matter) but I'm not sure how much of it can be discussed publicly, it is proprietary coding information I suspect.

There's been some pretty fierce discussions about the particulars of this in the dev boards. (Hi Bob!) And the engine has been tweaked repeatedly (and will probably continue to be tweaked post release) to get more accurate results.







willgamer -> RE: CV and Combat Modeling (12/7/2010 5:45:15 PM)

quote:

ORIGINAL: Ascended

It might seem like a lot of work for nothing, but as pointed out above, in operations historically these are by far the largest kind of influences on the success or failure of operational plans. Not supply (more strategic), not experience levels (often overwhelmed by numbers, quickly), not numerical superiority (tactical hangups 'clog' numerical odds OR neutralize them -- see 6 TC, first two days of Op Mars.), not technology... (terrain mitigates -- again, tactically.)


Not sure I buy this rather sweeping generalization.

But even if the point is granted, there's are larger problem that puts simulating at Ascended's level for a game scoped at the level of WitE (or any game where the scope of the simulated units is orders of magnitude greater than the desired scope of the simulation variables). This problem cannot be solved by more computing horsepower. The problem is DATA; for starters: detailed, quantative, historically accurate data about every piece of terrain that affects cover and concealment at the tactical level for each particular encounter.

If you have to randomly generate this data, then you have only added vast appearance of detail, without adding accuracy in the least.

The scope of the data needs to be appropriate to the scope of the simulation. Then combat resolution algorithms can incorporate Murphy's Law to represent unusual events at lower levels of combat within overall statistically valid results. It is impossible (due to the passage of time alone) the gather vast quanties of accurate low level data for a high level simulation.



kirkgregerson -> RE: CV and Combat Modeling (12/7/2010 6:19:03 PM)

I hear what Ascended is asking for and agree it would be adding more realism to the game. However, it's asking for almost the impossible. Here's some reasons why:

- I agree with the importance of tactical decisions after reading enough books about the many battles fought on the eastern front. Some of the outcomes are directly or indirectly related to some bizarre leadership decisions or just bad luck. So if the data existed it would still be impossible (at this point in time) to code up a database with all the leaders from Army down to squad level with their personalities, leadership strengths and weaknesses. All of this would have to be thrown into a very complex algorithm to determine battles outcomes.

- We haven't even begun to discuss all the communication issues that plagued the battlefields in WWII. How many times did units receive the wrong orders or just misinterpret orders. So wouldn't it make sense to factor in some factor based on leadership, weather, and maybe terrain as to units possibly not doing exactly what we have clicked on them to do? We take it for granted as armchair generals w/computer games that our wonderfully laid out plans to move units here and there, attack here and there, all happen as we wanted. Well that wasn't reality on the eastern front and with the communication technology at that period in history.

I actually remember an old SSI classic, Kursk, in which there were 2-4 phases (inf 2, armor 4) that you could plan your turns. After each side planned their turns they would be executed simultaneously. That actually gave a little flavor as to how your operational plans could go completely in a different directions. Not says that WitE should be modeled that way, but rather just something to think about.

Ok can I buy this game now? My credit card just jumped out of my wallet again and is humping my leg for WitE like some dog in heat. [:'(]



Ascended -> RE: CV and Combat Modeling (12/7/2010 9:32:02 PM)

It's not just a question of technical means, I think there is a "philosophic" edge to it about whether operational factors dominate tactical ones or vice versa.

I have no interest in the tactical level being highly detailed in an operational game per se, but I do want operational decisions to have realistic results. I want the player to be punished for making bad operational decisions the same way the commanders were historically. And I propose that current operational level games don't do that so well. They tend to fulfill expectations (even perhaps stereotypes, but that is debatable) rather than challenge them.

And yeah, Cray computers would be nice. But who wants to program on that level. ><



Page: <<   < prev  1 [2]

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
0.875