..vs Elmer only section..

..vs Elmer only section.. (Full Version)

All Forums >> [Current Games From Matrix.] >> [World War II] >> Norm Koger's The Operational Art Of War III

Message

a white rabbit -> ..vs Elmer only section.. (5/13/2008 6:23:46 PM)
..come on, i still say we don't have the foggiest idea how to program Elmer, not well anyway.. ..so can we have a place just to discuss vs-Elmer design ? .. frankly, i think we don't really know how to do pbem, but that's another area, already covered..

cdvalenta -> RE: ..vs Elmer only section.. (5/14/2008 2:32:30 AM)
I was just about to make a post regarding what people thought were the scenarios with the best programmed opponent and I found this thread right at the top. We just welcomed a new addition to the family and the time I have to play TOAW is very sporadic right now. I know people like to play PBEM, but what scenarios do you guys feel have the strongest PO?

Silvanski -> RE: ..vs Elmer only section.. (5/14/2008 4:19:18 AM)
Operation Mars by D.McBride is designed to be played vs the Soviet PO only Race to the Sea is quite entertaining

a white rabbit -> RE: ..vs Elmer only section.. (5/14/2008 5:25:12 PM)
..i know how cdvalenta feels.. ..and there are some fun scens vs Elmer but mostly any PO programming is done just to satisfy the check-list... ..ok, i started on this because of limited internet time, so a pbem is out of the window, but as we've now got electricity i can at least have a computer, and after dark, when the baby's asleep, and the farm slumbers i can get on with something useful, so back to Malaya, and a Japanese attack on the Slim river. Fine so far but then that's the historical, how do i get the PO to do interesting things, there's still only 3 objective tracks and they're global in effect, so not to be wasted on one combat, but if Elmer always attacks in the same way, at the same point, and that regardless of the situation, where's the challenge ?.. ..i suppose what i asking for is some sort of Critical Path Analysis tool, as i reckon the triggers, enemy attacks/occupies with a turn-linked activation that lead to a formation orders change, even guerrilla effect/withdraw unit/ EvEd new unit on a g-occupied hex etc are already there.. ..and as a side-product of this, it strikes me that vs-Elmer design is a thing apart, very different to pbem design, and so it needs its own discussion area..

Curtis Lemay -> RE: ..vs Elmer only section.. (5/14/2008 7:36:40 PM)
I'm pretty sure I've posted about this before, elsewhere, but the PO works best when you have sufficient formations that you can fill the map with objective paths. The denser the objective paths, the better. Redundancy helps as well. Where I've done that (Cambrai 1917, France 1944, Okinawa 1945, & Germany 1945) the PO tends to give a pretty good account of itself. But it still can't play nearly as well as a human, and I find all the effort to achieve those dense paths don't return well on the investment. Generally, the PO is a black hole in which an enormous amount of design effort can be poured without much result.

Veers -> RE: ..vs Elmer only section.. (5/15/2008 12:59:38 AM)
quote: ORIGINAL: cdvalenta I was just about to make a post regarding what people thought were the scenarios with the best programmed opponent and I found this thread right at the top. We just welcomed a new addition to the family and the time I have to play TOAW is very sporadic right now. I know people like to play PBEM, but what scenarios do you guys feel have the strongest PO? Anything modded by Silvinski. :D

rhinobones -> Objective Paths (5/15/2008 3:39:34 AM)
Think the biggest problem with the PO that may be readily fixed, is that formations will not necessarily follow the path prescribed if intermediate objective hexes are under friendly control before the formation advances to the intermediate objective. I had noted this in an earlier thread and was thoroughly shot down for the solution. It works like this, if formation A has objectives 1, 2 and 3, and objective 2 is already under friendly control, formation A will skip objective 2 and find its own path to objective 3. Quite often this will not present a problem, but imagine a case where the designer’s intent is that formation A goes to objective 2 makes a hard left turn and flanks the enemy all the way to objective 3. With the current system, when objective 2 is in friendly hands, formation A will instead head straight to objective 3 and the flanking attack will not occur as planned. The effect would be even more pronounced when there are more than three formation A objectives. A second, and equally troubling problem is the path that formation A reinforcements are forced to take. Who knows how a PO reinforcement unit will find its way from objective 1 to the parent formation if it doesn't follow the programmed path. Think we have all seen reinforcement take some curious paths to their parent formation. The solution, as I see it, is to program formations that have Attack, Independent and Advance orders to follow the objective path as programmed by the scenario designer. This means that formations will advance to objectives in numerical sequence regardless of objective ownership. This includes the path taken by reinforcements and reconstituted units. Formations with other orders could probably stay with the objective paths currently programmed in TOAW. Hopefully this is does not present a big problem for Ralph to incorporate. Regards, RhinoBones

a white rabbit -> RE: ..vs Elmer only section.. (5/15/2008 6:07:04 PM)
quote: ORIGINAL: Curtis Lemay I'm pretty sure I've posted about this before, elsewhere, but the PO works best when you have sufficient formations that you can fill the map with objective paths. The denser the objective paths, the better. Redundancy helps as well. Where I've done that (Cambrai 1917, France 1944, Okinawa 1945, & Germany 1945) the PO tends to give a pretty good account of itself. But it still can't play nearly as well as a human, and I find all the effort to achieve those dense paths don't return well on the investment. Generally, the PO is a black hole in which an enormous amount of design effort can be poured without much result. ..can you explain "redundancy" in more detail..xs units or xs objectives ?

a white rabbit -> RE: Objective Paths (5/15/2008 6:12:22 PM)
quote: ORIGINAL: rhinobones Think the biggest problem with the PO that may be readily fixed, is that formations will not necessarily follow the path prescribed if intermediate objective hexes are under friendly control before the formation advances to the intermediate objective. I had noted this in an earlier thread and was thoroughly shot down for the solution. It works like this, if formation A has objectives 1, 2 and 3, and objective 2 is already under friendly control, formation A will skip objective 2 and find its own path to objective 3. Quite often this will not present a problem, but imagine a case where the designer’s intent is that formation A goes to objective 2 makes a hard left turn and flanks the enemy all the way to objective 3. With the current system, when objective 2 is in friendly hands, formation A will instead head straight to objective 3 and the flanking attack will not occur as planned. The effect would be even more pronounced when there are more than three formation A objectives. A second, and equally troubling problem is the path that formation A reinforcements are forced to take. Who knows how a PO reinforcement unit will find its way from objective 1 to the parent formation if it doesn't follow the programmed path. Think we have all seen reinforcement take some curious paths to their parent formation. The solution, as I see it, is to program formations that have Attack, Independent and Advance orders to follow the objective path as programmed by the scenario designer. This means that formations will advance to objectives in numerical sequence regardless of objective ownership. This includes the path taken by reinforcements and reconstituted units. Formations with other orders could probably stay with the objective paths currently programmed in TOAW. Hopefully this is does not present a big problem for Ralph to incorporate. Regards, RhinoBones ..i didn't know that, thanks.. ..is there any common factor(s) in the paths chosen to un-owned objectives ? like they always use roads or some such ?

rhinobones -> RE: Objective Paths (5/21/2008 6:54:57 AM)
quote: ORIGINAL: a white rabbit ..is there any common factor(s) in the paths chosen to un-owned objectives ? like they always use roads or some such ? In plain English, your question is what . . . ? Regards, RhinoBones

rhinobones -> RE: ..vs Elmer only section.. (5/22/2008 2:18:00 AM)
quote: ORIGINAL: a white rabbit ..come on, i still say we don't have the foggiest idea how to program Elmer*, not well anyway.. Think that we will never be able to make OWPI (Opponent with Programmed Intelligence, that’s my version of Elmer) a thinking man’s equivalent, but as scenario designers we can make OWPI unpredictable and give it objectives that increase OWPI’s probability of victory. Think there are three things to consider if the scenario is intended to be played solitaire against the computer. Track Programming. I like to program the tracks so that OWPI has multiple possibilities for attack and defense. To do this I program the tracks and then use events to program their possibility of happening. Nesting events within events is important. As an example, I might program Track 1 as defensive, Track 2 as a double envelopment and Track 3 as a counter attack up the center. The events would be set up as such: 60% Defensive . . 50% Defensive (Event Range 3 thru 7) . . 30% Counter Attack Up The Middle (Event Range 3 thru 7) . . . . 70% Counter Attack Up The Middle (Event Range 8 thru 15) . . . . 30% Double Envelopment (Event Range 8 thru 15) . . 20% Double Envelopment (Event Range 3 thru 7) etc . . . 30% Counter Attack Up The Middle . . 60% Counter Attack Up The Middle(Event Range 3 thru 7) . . 20% Double Envelopment(Event Range 3 thru 7) . . 20% Defensive (Event Range 3 thru 7) 10% Double Envelopment etc . . . Etc, etc, etc . . . the nesting of events could be endless. The short story is that OWPI might attack, might remain defensive and might change his mind during the course of the scenario. The object is to make OQPI unpredictable and that can be done using nested events to screw around with the three available objective tracks. No Objective Victory Points. I have always thought that assigning victory points to specific hexes is nothing more than signaling exactly where the attack is directed. To me this eliminates a large part of the maneuver and tactical/operational aspects of the game. Better to measure success by the kill ratio. However, if objective points must be used, try making the kill of specific units (such as HQs) or capture of hexes scored by events (increase/decrease of supply and/or reinforcements) as the determinant for victory. I think there are better alternatives than the designation of specific hexes as worth “points”. No history that I have ever read has stated that a victory at hex XYZ was worth victory points. Played from One Side. My experience has been that if you try to build a scenario in which either side can be played by OWPI, the result is disappointing. I suggest that the design be for a Cheyenne player as Force 1 and OWPI as Force 2. This way the tracks for OWPI can be maximized without any distraction from the requirements of Force 1. As an example of these ideas you can check out the scenario Tulane. Regards, RhinoBones

a white rabbit -> RE: Objective Paths (5/25/2008 11:34:57 AM)
quote: ORIGINAL: rhinobones quote: ORIGINAL: a white rabbit ..is there any common factor(s) in the paths chosen to un-owned objectives ? like they always use roads or some such ? In plain English, your question is what . . . ? Regards, RhinoBones ..do the units that don't follow the objective paths, because of ownership, follow any identifable paths with common factors, ie always roads, or open land or whatever ?

a white rabbit -> RE: ..vs Elmer only section.. (5/25/2008 11:49:10 AM)
..thanks, some i'd got to, other points not. Above all i agree that a vs PO should be one side only,.. ..i was also looking at following possibilities: ..1/ variable re-enforcement arrival points tied to the game situation, ie formation 1 is available in different colours, it's arrival point triggered by the situation, which also removes all other variants from the schedual. each formation variant has its own unique objective paths ..2/ tieing some of the PO events to player ThOp choices..

Page: [1]

Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.7304688