aka_tom_w Posted March 4, 2002 Share Posted March 4, 2002 Originally posted by John D Salt: </font><blockquote>quote:</font><hr />Originally posted by Puff the Magic Dragon: An idea that rised in another threat: The problem of borg spotting is, each unit can receive orders or give reports, doesn't matter where on the map they are. Just something to think about : would it be possible to build a 'line of communication' or 'command'? [snips] Well now, this bids fair to be a horrifically interesting thread. Much of what has already been posted is interesting, but I have a couple of handfuls of worms to add to the can. The "realism vs. playability" chestnut is not really one I wish to pursue -- it's not clear to me what "realism" can sensibly mean in the wholly unreal world of a simulation, and almost everyone who discusses the matter confuses "realism" with "detail", which doesn't help. The "Commander's Boots" problem has been mentioned -- does the player represent Major Carstairs, OC of "B" Company the Borsetshire Light Infantry, or is the player a disembodied spirit, one of the household gods of the Company, looking down on all its members wherever they are? The question here is really "Do you want to play a role, or do you want to play a game?". Both are reasonable things to want to do; I would be happy to participate in a MUD set in WW2, but it would render competitive gaming problematic, if not impossible. The ultimate solution to the "Borg spotting problem" will require more effort than merely modelling communications channels and the orders and reports that flow along them. The comms infrastructure (whether using radio, telephone, flags, lights, runners or carrier pigeons) is merely the "plumbing" that permits messages to be transmitted from one place to another. What matters, from the point of view of tactical behaviour, is how the people who receive those orders or reports act on them. This, in turn, depends on how they integrate the new information into the picture of the tactical situation they currently have in their heads (as mentioned in van Creveld's superb "Command in War"). This is what is known as "situational awareness" (a web search on the name Micah Endersley might be a good start for people wanting to know more). Can this sort of thing be modelled in a computer simulation? Yes. The technique is known as "agent-based computing". ... snip Anyone else on the board going to the second AISB agent-based computing symposium at Imperial in April? All the best, John.</font> Link to comment Share on other sites More sharing options...
Redwolf Posted March 4, 2002 Share Posted March 4, 2002 Interesting discussion. I think what John D describes is really a different thing than CMBO. We are talking about a useful AI. An AI module that can play different levels, from platoon to the overall battlefield. The player can then choose "his" position. The real difference to current play is that if you have thoughtfully acting "TacAI" that can follow the orders and SOP you gave in advance or give during the game, then you can move away from the knows-everything, control-almost-everyone directly. To make this really work, you must not have a sharp seperation between TacAI/SOP and programmed opponent (global AI), you need stackable agent-commanders which know how to deal with either terrain or subordinates. Once you are there, the player can choose any position he wants, even a position where he gets commands by higherlevel agent-commanders. And you get cooperative multiplay, a bunch of people playing together against the AI. I am harrassing game developers to open the TacAI/SOP and the global AI (programmed opponent) modules to customers, so that you can write your own and plug it into the game (like a DLL or a Java class), but so far without luck. And it will probably be too restricted, as I said if you have seperate AIs for Tac and global, it already goes downhill. Time for an OpenSource wargame, it seems. P.S. Tacops 4.x will have multiple players per side, with FOW between them. And I find it to be a very realistic game, some of you might want to look into it. Link to comment Share on other sites More sharing options...
Dschugaschwili Posted March 5, 2002 Share Posted March 5, 2002 My opinion regarding the borg spotting: As long as CM lets the player control every unit directly (by giving orders to it), a part of the problem is unavoidable: the player's knowledge is the sum of his units' knowledge. This has to be the case because he must have the ability to take a single unit's role and issue orders based on this unit's full knowledge. The second part of the problem can be (at least partially) "fixed": Currently, if the player knows about an enemy unit, every friendly unit within LOS of this unit also know about it, not necessarily immediately (if an enemy unit is spotted during a turn, not every friendly unit will fire at it immediately; especially buttoned tanks often need a while to notice), but at the start of the next orders phase. To get rid of the problem, several things are required: </font>Each unit needs its own list of known enemy units (preferably with the possibility of mis-identifications independant of the other friendly units). </font>A communication system (like discussed in previous posts) to simulate units telling each other what they just spotted. </font>Restrictions for the use of the area fire command. </font>As for restricting are fire, this is of course difficult to balance. Only allowing it near enemy units known to the area firing unit may be a possibility (at least for guns/tanks/mortars). Dschugaschwili [ March 05, 2002, 05:44 AM: Message edited by: Dschugaschwili ] Link to comment Share on other sites More sharing options...
Recommended Posts