positive feedback and ideas

World in Flames is the computer version of Australian Design Group classic board game. World In Flames is a highly detailed game covering the both Europe and Pacific Theaters of Operations during World War II. If you want grand strategy this game is for you.

Moderator: Shannon V. OKeets

macgregor
Posts: 999
Joined: Tue Feb 10, 2004 6:44 pm

RE: food for thought

Post by macgregor »

Excuse me for not maintaining the proper level of decorum komrad coregames! For god's sake this is a forum not a pep rally! I'll save my opinion for my own threads if that's the case. It sounds like you only understand properly formulated code as well. Which is subjective by the way.
I'm confused. Is the person that posts under the name "Coregames" a developer for computer WiF or is he a potential customer like the rest of us?

Well I guess that's the 64,000 dollar question.
User avatar
Greyshaft
Posts: 1979
Joined: Mon Oct 27, 2003 1:59 am
Location: Sydney, Australia

Play nicely everyone

Post by Greyshaft »

AFAIK the only developers for MWiF are Robert Crandall with assistance from Chris Marraccini. Calling coregames a matrix developer with the implied slur that he's not presenting an impartial perspective is both incorrect and inappropriate. It also lowers the tone of the Forum by attacking the individual rather than his arguments.

Get a grip guys... Matrix has an obligation to Harry to produce the best dam' game of computer Wif that it can and sometimes it seems that your only criteria for the success of that game is that you want it NOW!!! Well I want it now also and I'm sure coregames is the same, but reality intrudes and we must all wait a while. Let's try and do it nicely.

/Greyshaft

PS I'm not a matrix developer either.

PPS I wish I was.
/Greyshaft
Shannon V. OKeets
Posts: 22135
Joined: Wed May 18, 2005 11:51 pm
Location: Honolulu, Hawaii
Contact:

RE: standardized strategies for AI

Post by Shannon V. OKeets »

There was a reference to the AI software for chess programs and its decision tree approach. In case you aren't aware, WiF is a much more complex game that chess. There are more pieces. There are more unit types. The playing surface is larger. Instead of moving one piece per turn you move dozens. Combat is probabilistic not deterministic. The terrain is diverse and changes with the season. Need I go on?

Decision trees are pretty hopeless as a solution to this level of complexity. You have to go back to basics. Assess the strength of your forces and those of the enemy. List your possible moves and the enemies' possible responses (only one level deep!). Refer to an operational goal and use that to determine tactical goals. Formulate a move (which involves multiple units) that might achieve a tactical goal, then assess the probability of success. Assess the effect of failure. Weigh the risk-benefit tradeoff and decide whether the move is still feasible. You now need to repeat for all the other tactical goals you might want to acheive. Note that you must avoid the trap of using all your best units to acheive the first tactical goal thereby leaving you with nothing for other tactical goals. How about units to defend a counterattack? Hey, there is still the second impulse to think about too!

Operational plans would require the development of a special language. The language, at a minimum would list objectives, a time table, a maximum level of troops commitment (e.g., no more than 6 armored nits and 10 infantry), maximum sustainable loss permitted. If all we are trying to take out is a resource point, then we would not want to lose too many units. However, if we are talking about the fall of France, then we are willing to risk many more. As a case in point, it is always difficult for the Commonwealth player to decide on how many land and air units to deploy in France. You want to be a good ally but you don't want to lose so many units that important parts of the Commonwealth are lost immediately after France falls.

In my experience the best way to model something is to follow in the steps of those who have (had) day to day exposure to the problem. The AI should 'think' in terms of army and air groups. Theaters of operation and 'fronts'. Take the allied attacks in Western Europe in 1944. Which army gets the oil? Who gets reinforcements? Are we using our air superiority for stratgeic or tactical attacks? Some units might be transferred from army group A to army group B for the short term or long term. Within this structure it is easier to evaluate what is feasible and what isn't. Yes it does limit the number of options and perhaps a brilliant use of diverse forces would produce dramatic results. And yet, to always be evaluating ALL the potential moves creates the risk of doing something goofy. Human players pick up on this stuff pretty fast. "Just leave that sacrifical lamb out there. The AI will go after it and leave the back door wide open." Instead, do what the Generals did, from Zhukov and Ike on down.
Steve

Perfection is an elusive goal.
User avatar
coregames
Posts: 470
Joined: Thu Aug 12, 2004 4:45 pm
Contact:

RE: food for thought

Post by coregames »

I am just an avid WiF player and an aspiring game designer. If I represented Matrix, I assure you there would be an admin icon, or at the very least, more than just two stars under my name. My motive in creating this thread and in participating in Matrix forums in general is to support and take part in their excellent online gaming community, and to encourage success at creating the best MWiF possible.
"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson
User avatar
coregames
Posts: 470
Joined: Thu Aug 12, 2004 4:45 pm
Contact:

RE: Play nicely everyone

Post by coregames »

thanks Greyshaft
"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson
User avatar
coregames
Posts: 470
Joined: Thu Aug 12, 2004 4:45 pm
Contact:

RE: standardized strategies for AI

Post by coregames »

ORIGINAL: Shannon V. OKeets

There was a reference to the AI software for chess programs and its decision tree approach. In case you aren't aware, WiF is a much more complex game that chess. There are more pieces. There are more unit types. The playing surface is larger. Instead of moving one piece per turn you move dozens. Combat is probabilistic not deterministic. The terrain is diverse and changes with the season. Need I go on?

If you can't assemble a grand strategy in WiF, you can't succeed. Obviously, an AI can't create a decision tree for every land, air and naval move possible for turns (or even impulses) in advance. My point was not to micro-control tactics with a decisions tree (chess was only an analogy). But something must connect strategic themes into a cohesive whole, or the AI's position will drift aimlessly into ruin. It is true that WiF uses a lot of luck, but if that was the biggest determining factor, I for one would not play it. There are enough instances of luck for the effects of it to be somewhat statistically averaged out (especially if you play many times).

Operational plans would require the development of a special language. The language, at a minimum would list objectives, a time table, a maximum level of troops commitment.

Am I to assume that the "special language" you propose for operational plans is one-dimensional, with no branching allowed? If branching does exist in the way the operational plans are formulated (as in having a Plan B), that sounds a lot like a decision tree to me (though admittedly not a tactical one). If no branching exists, in my opinion the AI will be weak, following a shortsighted and inflexible game plan. In the middle of a long scenario, operations are sometimes planned 6 turns or more in advance, and the game is so variable that standardized strategies past a certain point are of limited value.

I am not an advanced programmer, and I don't know what goes into a strong AI for a complex grand strategic computer wargame. I have played a lot of WiF, however. To model what strong WiF players do strategically, the AI will need to use some sort of lookahead, weighing various possible outcomes against eachother based on some kind of valuation of the position.

If you read my post, you will notice that my point was not about using brute-force lookahead, but about positional recognition:

Strong chess programs are increasingly dependent on positional parameters, still relying on brute-force look-ahead to see many variations to a great depth, but having a better idea of how to evaluate the position at the termination of each branch of their search. To play well, MWiF will need to have a strong meta-strategy component...

I go on to talk about using a high-level meta-strategy "tree" that is constructed by part of the AI, tasked with stringing together these "called plays" from a library of plans that represent the experience of strong players (Greyshaft mentions this in an earlier post in this thread). The result is not an attempt to use brute-force to plan an impulse or a turn, but rather, to prepare well in advance so that the most important operations have an improved chance of success.

I do understand that WiF is much more complex than chess, and that it is unpredictable. Everything I say may be way off-base, given my lack of experience developing AI. I am willing to be shown why a meta-strategic decision tree is a bad idea.
"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson
Shannon V. OKeets
Posts: 22135
Joined: Wed May 18, 2005 11:51 pm
Location: Honolulu, Hawaii
Contact:

RE: standardized strategies for AI

Post by Shannon V. OKeets »

I apologize if I offended. It was unintentional.

I would suggest that the AI in some sense follow the same decison making process that good players do. People rarely follow decision trees when making decisions. It is actually an artifical and non-intuitive approach to decison making. I speak from the experience of having to teach decision trees to college students, the majority of which found them completely alien. People primarily use If ... Then ... rules to make decisions. When the conditional statements are met, then they execute the action portion of the rule. For example, what clothes to where, where to go for lunch, who to develop as a friend, etc..

Rules work pretty well for AI opponents. They permit taking into consideration weather, the probability of success, the relative strength of the forces involved, and the other concerns that human players worry about. Rules do not have to be absolute go or no go. It is easy to add in a random element that controls whether the action is actually executed. They can be employed to connect strategic plans to operational plans to tactical plans and in the reverse direction as well. What I have used in the past started at the strategic level and worked its wasy down to the tactical. At that point the program assessed the viability of the tactics and fed the evaluations back up the chain to the strategic level. You can iterate this a couple of times but have to have a way of definitively cutting it off so that the program actually makes a move instead of studying its navel indefinitely.

There are a lot of papers available in the literature on both the theoretical and practical applications of rule based systems, going back to the EMycin program for internal medicine in the 1970's. This is not to say that rules can be used exclusively for developing the AI opponent. The determination of which units to use in an attack will have to be hand crafted to optimize efficent use of units based on the WiF combat system. Many of the other decisions will also need specialized code. Rules look to be a viable choice for deciding whether to use preplanned responses though.

Bear in mind that the thoughts I am puttng forth here are only 10 or 20 seconds old and they may not age gracefully. I write them in the spirit of "ideas for the AI in CWiF". My ideas haven't been alive long enough for me to grow attached to them. If they provoke a thought or two in the reader, then I would be happy.
Steve

Perfection is an elusive goal.
SeaMonkey
Posts: 796
Joined: Sun Feb 15, 2004 3:18 am

RE: standardized strategies for AI

Post by SeaMonkey »

Bear in mind that the thoughts I am puttng forth here are only 10 or 20 seconds old and they may not age gracefully. I write them in the spirit of "ideas for the AI in CWiF". My ideas haven't been alive long enough for me to grow attached to them. If they provoke a thought or two in the reader, then I would be happy

Excellent philosophy in the context of game development......a "think tank" environment conducive to the evolution of game play structure. I like your approach Shannon, logically analytical to examine the possibilities.

Still, I think that the developers need to offer a definitive path of creation for WiF as others have suggested, so that the ideas can be elaborated on in an atmosphere of finite discussion. Certain answers to questions about the "faithful recreation" of WiF need to be specified, like are we going to see phases with player interaction ala TCP/IP/LAN connection? Is a PBEM sequence viable for a "faithful recreation"? How much deviation from the traditional WiF will be tolerated. Will this be a totally new game? It seems many compromises will have to made to make this a viable economic project......does that ruin the concept of "faithful recreation"? There are many unanswered questions and so far I have only heard ambiguous responses, so I witness the hesitation of the posters to contribute ideas that are maybe,.. inconsequential, ...insignificant...a waste of everyones time.

If a path of development could be a little more focused then perhaps it would incite a profusion of ideas...I actually have some myself...even having never played WiF, I'm still interested in this apparition of a legendary WW2 simulation for the computer.

But alas, my ideas...your ideas...are like the fleeting thoughts Shannon has described, put forth with no assemblage into a vast cauldron of bubbling thought. Eventually there will be so many that the soup will mask all the characteristics of the added herbs and so will have no distinct flavor, and we will have to rehash them all again.

My time is better served playing WaW, HttR or perhaps and old scenario of SC.
User avatar
coregames
Posts: 470
Joined: Thu Aug 12, 2004 4:45 pm
Contact:

input after-the-fact

Post by coregames »

ORIGINAL: SeaMonkey

But alas, my ideas...your ideas...are like the fleeting thoughts Shannon has described, put forth with no assemblage into a vast cauldron of bubbling thought. Eventually there will be so many that the soup will mask all the characteristics of the added herbs and so will have no distinct flavor, and we will have to rehash them all again.

My time is better served playing WaW, HttR or perhaps and old scenario of SC.

I understand your point SeaMonkey, but I disagree. It seems to me that the best time to put your suggestions forth is before the direction of the design is set in stone. This is not to say that Matrix will necessarily listen to anyone in particular about how to approach this project, but to propose ideas about how to proceed after the design documents are in existence seems to be somewhat futile. Of course, ideas will still be proposed in this forum once some big decisions have been made about the form the game will take, but these ideas will be regarding smaller technical issues. For many of us who have played the game for years, the larger issues are important enough to comment on in advance, in the hope of affecting Matrix's design decisions.

As far as flavors in the soup of this forum, there are many as you suggest, but two in particular are clearly identifiable: the flavor of a faithful adaptation, and the flavor of PBEM suitability. These issues in particular are worthy topics for discussion in this forum. By waiting for a firm commitment from Matrix on the direction of MWiF before commenting, WiF fans runs the risk of having to settle for a game they might not prefer. Of course, regardless of the comments posted in this forum, Matrix may go their own way, but at least this way they have some feedback before hand to get a sense of what gamers in general (and WiFers in particular) want.
"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson
User avatar
Greyshaft
Posts: 1979
Joined: Mon Oct 27, 2003 1:59 am
Location: Sydney, Australia

ambiguous responses?

Post by Greyshaft »

ORIGINAL: SeaMonkey
...There are many unanswered questions and so far I have only heard ambiguous responses ...

It is perhaps relevant to note that the ambiguous responses have all come from non-Matrix posters (such as myself perhaps).

Official Matrix types have consistantly said the same thing... "No news yet guys. We'll tell you when the situation changes."

While I dislike that situation I cannot accuse Matrix of inconsistancy or ambiguous responses.
/Greyshaft
User avatar
Mziln
Posts: 667
Joined: Mon Feb 09, 2004 5:36 pm
Location: Tulsa Oklahoma

RE: ambiguous responses?

Post by Mziln »

RE: My POV - 5/19/2005 9:45:06 PM


Erik Rutins
Administrator





Posts: 4640
Joined: 3/28/2000
From: Vermont, USA
Status: online This project is not vaporware. We also did not promise an update after E3, but after Origins, so you can expect to hear something by mid-July. We will be meeting in person with Rob Crandall at that point and going over the release plans. Until that time, all I can say is it's not vaporware, we are not trying to keep it from the community and these forums most certainly exist for two-way communications.

Regards,

- Erik

_____________________________

Erik Rutins
Director of Product Development and Business Relations

It's omly about 40 days left untill origins. Please note it says some time in mid-july.

SeaMonkey
Posts: 796
Joined: Sun Feb 15, 2004 3:18 am

RE: ambiguous responses?

Post by SeaMonkey »

OK...OK, perhaps "ambiguous" was an improper use of vocabulary. But to a person that is not familiar with WiF, other than the observations available on line, the term "Faithful Adaptation" has a lot of maneuver room, especially in the context of PBEM and a semi-successful AI.

My point, which I know many of you see, is that a distinction needs to be made. Either it will be phased/impulse player interactions as I understand the board game to possess(not sure of the beta WiF) or it will be programmed optionally selective AI interaction for the nonphasing player.

Now I'm not completely convinced in my own mind that there is not another possibility, but in the realm of CPU wargame design, it is very hard for me to grasp another concept of implementation in light of a PBEM and AI inclusion.

I apologize to Matrix and my fellow posters for my simplistic outlook, but it seems irrelevant to offer ideas until this anomaly is solved.
User avatar
coregames
Posts: 470
Joined: Thu Aug 12, 2004 4:45 pm
Contact:

RE: ambiguous responses?

Post by coregames »

ORIGINAL: SeaMonkey
Either it will be phased/impulse player interactions as I understand the board game to possess(not sure of the beta WiF) or it will be programmed optionally selective AI interaction for the nonphasing player.

Chris Marinacci's beta was in all respects a direct port of WiFFE, with the sole exception that the map scales were integrated worldwide, to the European scale used in the tabletop version. AI and PBEM were not included in any way. Only three scenarios (two training scenarios and Global War) had been implemented at the time the project was discontinued; still, it was an amazing accomplishment by Chris.

Interactions in the boardgame are much more than on a phase-by-phase or impulse-by-impulse basis. For example, during the groundstrikes phase, air combats are possible over multiple targets, and during each air combat, multiple rounds are possible (even likely if many aircraft are involved). During a round of air combat, both sides may be called upon to implement results, including which aircraft to clear, abort or shoot down, and then at the end of each round, both sides get to decide whether or not to abort the mission. Naval combat can also involve multiple rounds for each engagement, with decisions to be made by both sides as combat progresses. By far the simplest case of interaction is land combat, where table choice, casualties and possible retreats are the extent of the required decisions, and where only one round of combat is possible during an impulse.
"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson
StrictlyRockers
Posts: 11
Joined: Thu May 26, 2005 4:31 am

RE: ambiguous responses?

Post by StrictlyRockers »

I played with Chris Marinucci's version about four years ago and sent it to some friends, too. It was still in a rough state when I saw it and I had problems with updating the map and keeping it clean. Moving units sometimes left "trails". I was still quite impressed with what he had accomplished and I am happy knowing that Matrix has taken over development of CWiF. This has been a dream for decades. I have been an avid WiF fan for over 20 years. I belong to a good group that has been playing the board game consistently for that long.

I met Harry Rowland and Greg Pinder at Origins in the 80's and discussed the game at length with them and participated in some games there on a map that was blown up and mounted vertically and with magnetized units. It was great fun. Some in the group have given suggestions for improvements over the years and one even recieved playtesting credits on Days of Decision.

This game as a loyal fanbase of fanatical players like me. There are few games that compare with WiF for historical representation and simulation of grand strategical conflict. The game has gotten better and better over the years with all of the expansions. CWiF has been a dream, a chimera, an illusion that dances just out of your grasp for twenty years.

I sent some emails back and forth with Chris on tips to improve his program. I am eagerly awaiting news on CWiF and I am willing to do whatever I can to help with production of a game that comes as close to replicating the board game as possible.

Programming the AI for a game like this sounds like a bear from hell. I know how difficult it is to code the algorythms for the AI of a good wargame. It's very hard to get it right. I suggest programming three or four "plans of action" for the AI to select from on a grand scale and to then tailor decsions to suit the grand strategy. For instance, a German AI could select from grand strategies such as early Babarossa, late Babarrossa, Sealion or Mediterranean/Middle East strategies, depending on circumstances and luck received.

If anyone can actually present a workable, challenging AI model for WiF that includes the same level of decision making as the original, I will personally buy two copies of the game and also send fifty bucks to each one of the designers. I don't think it's a very easy thing to do.

When it comes to playing by email or TCP/IP, the game would have to be modified somewhat to streamline it a bit. When you are talking about needing to decide which air units to intercept with for each air strike or whether to intercept each ship movement depending on which sea zone and sea-box it enters, I can see a reason to want an option for pre-set instructions or automated orders. When you have to decide which airplane to abort for every air combat result or implementing each sea zone interception battle, the game can bog down a bit if playing by email.

It's the AI for the naval stuff that I think could be the most difficult to get right. It's hard enough for a human player to know how to move fleets around in heavily occupied territory with the sea zone and sea box system.

I for one am a huge WiF fan, and I will follow this one closely. I am happy that Matrix games has it. I like this company. There is real dearth of good grand scale stragegic simulations on the market. Some of us have no interest in FP shooters or RTS. We want the real thing. It's been more than 20 years since Clash of Steel, which is the one that came closest to recreating the right feel and system for Eastern Front and earlly WW II warfare. COS was partly based on WiF. Recent efforts by Gary Grigsby and Battlefront's Stragegic Command do not satisfy the palate. I haven't tried Heart's of Iron yet, but there really isn't much out there to choose from is my point. I think the market for a good quality version of CWiF will be larger than expected. The market is ripe for this game and there is NOTHING out there to compare with it. I know that there is an interesting updated version of Third Reich coming out soon called A World at War. Hopefully that will tide us over until CWiF comes out.

Christo
"SR"
User avatar
coregames
Posts: 470
Joined: Thu Aug 12, 2004 4:45 pm
Contact:

WiF or computer gamer?

Post by coregames »

Sorry I haven't responded before this StrictlyRockers, but two other threads have consumed my attention. I'm glad you appreciate WiF as much as so many of us "civilians" in here. I envy you having met those two (only Steve Jackson and perhaps Greg Kostykian are as high in my pantheon). They have been our version of Lennon and McCartney for 25 years*.

When it comes to playing by email or TCP/IP, the game would have to be modified somewhat to streamline it a bit.

I don't agree with you about TCP/IP. In that mode, the swap files would be tiny given the fact that WiF is, after all, "only a board game". Any multi-media content would be stored at either end, being only eye and ear candy to make the game more interesting to younger (and more computer-oriented) players. Play could very closely resemble the board game in turn sequence, even if the players did not have broadband, and still move relatively quickly, provided everyone was playing concurrently.

PBEM does require adustment to the turn sequence to be viable in less time than the real war. As we've been discussing in here, that could mean scripting contingencies, delegating decisions to the AI, consolidating and/or simplifying the sequence, or some combination of these things. This seems challenging, but probably not as much so as the AI.

Many of us agree with you about retaining authenticity in MWiF. The debate for months has revolved around that very issue (PBEM vs. faithful adaptation). I am curious: what is your feeling about the other (long-quiet) debate? I'm referring to unified scale vs. variable scale. At first, I lobbied for Matrix to include the option of playing the game as a direct port of the board game. Obviously, strictly computer players would want the unified scale, but the game becomes a much better training tool for the board game if the variable scale is retained. I have long since given that debate up, preferring to rally behind the synchronous play banner (gotta pick your fights).

By the way, most of the issues you talked about had been resolved in Chris's version. It's not perfect, but still farther along than when you last checked it, based on your post. I haven't made it out of 1940 yet, but Patrice DeForno says he has finished a Global War scenario.

* (Gygax and Arneson as Jagger and Richards?)
"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson
StrictlyRockers
Posts: 11
Joined: Thu May 26, 2005 4:31 am

RE: WiF or computer gamer?

Post by StrictlyRockers »

ORIGINAL: coregames

Many of us agree with you about retaining authenticity in MWiF. The debate for months has revolved around that very issue (PBEM vs. faithful adaptation). I am curious: what is your feeling about the other (long-quiet) debate? I'm referring to unified scale vs. variable scale. At first, I lobbied for Matrix to include the option of playing the game as a direct port of the board game. Obviously, strictly computer players would want the unified scale, but the game becomes a much better training tool for the board game if the variable scale is retained. I have long since given that debate up, preferring to rally behind the synchronous play banner (gotta pick your fights).

By the way, most of the issues you talked about had been resolved in Chris's version. It's not perfect, but still farther along than when you last checked it, based on your post. I haven't made it out of 1940 yet, but Patrice DeForno says he has finished a Global War scenario.
* (Gygax and Arneson as Jagger and Richards?)

Ok, I am definitely a purist on this one. Keep the map scales the same, with Asian maps having a hex represent twice the area that the European maps do. I don't think it will be a travesty if they switch it to a unified scale, but I would argue for keeping it the same.

The China front has it's own dynamic going on. And the Pacific Ocean covers huge, huge areas, so you are talking about taking on an additional level of complexity to convert the Pacific maps to European scale. If they can do it, and still maintain the game's integrity, then ok. But it's going to be plenty hard enough already to get this right and do this game justice.

I am serious; if I see a working model of this game with a working AI I will give you one of these [&o] and also mail each designer a personal check as a reward for doing this. I don't think it can be done. I am hopeful, but not expectant.

SR
User avatar
coregames
Posts: 470
Joined: Thu Aug 12, 2004 4:45 pm
Contact:

RE: WiF or computer gamer?

Post by coregames »

ORIGINAL: StrictlyRockers

I am serious; if I see a working model of this game with a working AI I will give you one of these [&o] and also mail each designer a personal check as a reward for doing this. I don't think it can be done. I am hopeful, but not expectant.

I may only speak for myself StrictlyRockers, but I would rather you not let us know about your pessimism. It could hinder morale, which has been high in here lately.

As to scale, my hope is that they include the option to play at the ADG scales, as well the option for a unified scale.
"The creative combination lays bare the presumption of a lie." -- Lasker

Keith Henderson
Post Reply

Return to “World in Flames”