RE: When?

RE: When? (Full Version)

All Forums >> [New Releases from Matrix Games] >> World in Flames

Message

TemujinOz -> RE: When? (6/5/2012 2:27:48 AM)
quote: ORIGINAL: GordianKnot As a developer, I'm going to guess that one of the biggest roadblocks to getting this done quickly has been of all things the game options. Having so many options creates a huge development and testing challenge. In fact this game has so many options its probably impossible to test them all. If you have 50 different options and they are all binary (true or false) that's 2^50 or 1,125,899,906,842,624 possible combinations! This is a key point in any development. I've been working on a new patch for WiE-win lately, and in the forums for that game there has been a lot of discussion on possible changes, additions and corrections. Many times there's a request for an "option" to be added, to allow players to choose how they want to approach a given issue. I've had to point out more than once that although 'options' seem like a great way to solve any discussion/dispute (or to add new features), there are really two costs associated with adding an option - first, the immediate cost in time required to design and write the code; second, the future cost in increased complexity and testing that each option brings. In simplistic terms, as GordianKnot says, every binary option doubles the possible execution paths through the code. It doesn't take many options to quickly reach a point where "full testing of all combinations" is not possible. "Options breed bugs"...

RJL5188 -> RE: When? (6/5/2012 5:01:22 AM)
how about the first computer i ever got was in 1983...a radio shack trs-80 model 1...4k ram...upgraded to 16k later...wow...im old

ioticus -> RE: When? (6/5/2012 5:30:19 AM)
quote: ORIGINAL: Shannon V. OKeets I also wrote the AIO for two war games in the mid-1980's for the Atari (in assembler - yuck!). My programmer experience goes back to 1968. What were those two war games?

winky51 -> RE: When? (6/5/2012 6:46:28 AM)
I said it when the project first started, A.I. for WIF on a large scale is at this time impossible. The best you can hope for is tiny scenarios with limited pieces, no production, to teach naval, land, or air combat in a fixed environment. Once computers become self aware then an A.I. for WIF will be possible and it will be written by a computer, not a human. I would have released WIF a long time ago without A.I. or as noted above. Right now WIF on Vassal is free and easy. Its going to be hard to convice a player to cough up a lot of money for this game.

Shannon V. OKeets -> RE: When? (6/5/2012 8:51:48 AM)
quote: ORIGINAL: ioticus quote: ORIGINAL: Shannon V. OKeets I also wrote the AIO for two war games in the mid-1980's for the Atari (in assembler - yuck!). My programmer experience goes back to 1968. What were those two war games? GDW's The Battle of Chickamauga and Rommel: Battles for Tobruk.

pzgndr -> RE: When? (6/5/2012 1:15:23 PM)
quote: I just note that AH was unable to create a decent AI for its Third Reich pc game in ten years... Thus I think that the same problems found for computer 3R exist also in Computer WiF, but multiplied by 1000... Mey. I've successfully scripted decent and challenging AI for my A3R mod in the Strategic Command series. It's possible. I believe Steve can also develop a decent and challenging computer opponent for MWiF. It may take a few iterations but it's possible. Will it be perfect, will it challenge every experienced veteran? Maybe not, but it should be good enough for most casual players (aka customers). [8D]

ecz -> RE: When? (6/5/2012 5:11:51 PM)
[/quote] Mey. I've successfully scripted decent and challenging AI for my A3R mod in the Strategic Command series. It's possible. I believe Steve can also develop a decent and challenging computer opponent for MWiF. It may take a few iterations but it's possible. Will it be perfect, will it challenge every experienced veteran? Maybe not, but it should be good enough for most casual players (aka customers). [8D] [/quote] what? someone here thinks that an "average" AI would be satisfactory for the expert WiF players around waiting this game from decades? [&:] If the aim is to have an average/moderate AI good just for casual WiF players (BTW how is possible be casual WiF Players?) much better quit and try something else. Such a game cannot have "casual players" by definition IMO. But I'm sure Steve is instead working on a AI challenging expert/veteran players, the only AI level that could interest the potential audience for this game. The problem is that is a very hard task. And I'm using an euphemism here... If it would be less than a very hard task translate an existing game in a close PC version we would already have the best wargames of our life dancing in our hard drives and not only their VASSAL image.

Shannon V. OKeets -> RE: When? (6/5/2012 7:21:22 PM)
quote: ORIGINAL: ecz quote: Mey. I've successfully scripted decent and challenging AI for my A3R mod in the Strategic Command series. It's possible. I believe Steve can also develop a decent and challenging computer opponent for MWiF. It may take a few iterations but it's possible. Will it be perfect, will it challenge every experienced veteran? Maybe not, but it should be good enough for most casual players (aka customers). [8D] what? someone here thinks that an "average" AI would be satisfactory for the expert WiF players around waiting this game from decades? [&:] If the aim is to have an average/moderate AI good just for casual WiF players (BTW how is possible be casual WiF Players?) much better quit and try something else. Such a game cannot have "casual players" by definition IMO. But I'm sure Steve is instead working on a AI challenging expert/veteran players, the only AI level that could interest the potential audience for this game. The problem is that is a very hard task. And I'm using an euphemism here... If it would be less than a very hard task translate an existing game in a close PC version we would already have the best wargames of our life dancing in our hard drives and not only their VASSAL image. Well, I think almost every WIF player has 'holes' in their playing ability. They forget to do things that they know they should do, or planned on doing things but didn't when the time came, or failed to set up all the prerequisites for, and therefore couldn't do. Many other events arise during a game that cause players to fail to live up to their own opinion of their playing ability/skill. Even the best professional ball player screws up from time to time. And into those gaps in the human player's execution the AI Opponent can shove long sharp objects. If the AIO causes you to restore a game that was saved earlier, does that make it a good AIO? Or must it be able to defeat you despite your dozens of "oh, wait a minute, I need to go back and ..."? While many forum members are of the opinion that WIF is too difficult for the AIO to play well, they may be forgetting that it is also very difficult for human players to play well.

ioticus -> RE: When? (6/5/2012 7:38:02 PM)
"While many forum members are of the opinion that WIF is too difficult for the AIO to play well, they may be forgetting that it is also very difficult for human players to play well." Exactly. I've never played the board game but always wanted to. If the AIO gives me a good game by capitalizing on my beginner mistakes, I'll consider it money well spent.

ecz -> RE: When? (6/5/2012 8:19:54 PM)
Steve first of all let me say once more I really appreciate your work for WiF and respect your extra-long curriculum that make you fully qualified to do this job. Besides I agree on everything, it's hard if not impossible designers/players can teach a computer to play BETTER than them. Expecially a game so complex as WiF. And this a reason why a large part of computer games compensate the lack of AI with sad tricks with dice and combat results or, more honestly, openly give an handicap to the Human player(s). Not being this the case I would be happy to have at least a game where the AI knows better than me the rules and does not make obvious mistakes even if it's unable to play aggressively. But I see this too difficult for the countless options at disposal impossible to predict in detail at the design stage, also given our current tech level. My only concern and displeasure is that if the priority has been from the beginning just a game "knowing the rules" with no dreams about AI today we would probably already have WiF by years with no real harm for the AI project (that requires better machines and in general better resources not jet available at that time). But this has little to do with you or us, and much to do with Matrix policy, bearing in mind that Matrix has all the right to make the choices it likes, correct or wrong. Of course may be I'm wrong and you are able to complete this project soon and the need for a team of scientists working full time two years closely with your group of playtesters is an exaggeration. But, forgive me, I tend to believe that such a monster game will not see the light in the form we hope because commercially impossible ( where are the funds for the required team of scientist ?) if not technically impossible as well. Time will respond. [;)]

SamuraiProgrmmr -> RE: When? (6/5/2012 9:27:29 PM)
quote: ORIGINAL: Shannon V. OKeets And into those gaps in the human player's execution the AI Opponent can shove long sharp objects. If the AIO causes you to restore a game that was saved earlier, does that make it a good AIO? Or must it be able to defeat you despite your dozens of "oh, wait a minute, I need to go back and ..."? OOOH! The gauntlet has been thrown down. [:'(] With all of the discussion about AI (good and bad), I wanted to mention an honoroble contender....OGRE The computer version of this Steve Jackson Game was an excellent AI opponent, especially as the OGRE. This was because it could through Brute Force examine all possibilities. The game was simple enough. I look forward to the day when someone who has not been part of these discussions over the years floods the form with complaints of how the AI CHEATS! and think of the smile on Steve's face when that happens. Go for it, Steve!

ioticus -> RE: When? (6/6/2012 1:35:52 AM)
OGRE had great AI, I agree.

danlongman -> RE: When? (6/6/2012 2:31:30 AM)
I had those two GDW games. Played them on my atari800 which was not my first machine. Man I had Imagination in those days! Nobody in modern times would sit still for any of that stuff. I remember I got the Eighth Army behind the Afrika Korps in one scenario. I was so excited I almost wet meself. I cicked to end the turn and the machine ran and ran and ran all evening and overnight. There was an electrical storm the following afternoon and a power bump followed. In those days a powerbump while the disk drive was running on an atari800 fried the disk. That was the end of my North African Adventure. Oh we live in good times. cheers

Shannon V. OKeets -> RE: When? (6/6/2012 3:16:41 AM)
quote: ORIGINAL: ecz Steve first of all let me say once more I really appreciate your work for WiF and respect your extra-long curriculum that make you fully qualified to do this job. Besides I agree on everything, it's hard if not impossible designers/players can teach a computer to play BETTER than them. Expecially a game so complex as WiF. And this a reason why a large part of computer games compensate the lack of AI with sad tricks with dice and combat results or, more honestly, openly give an handicap to the Human player(s). Not being this the case I would be happy to have at least a game where the AI knows better than me the rules and does not make obvious mistakes even if it's unable to play aggressively. But I see this too difficult for the countless options at disposal impossible to predict in detail at the design stage, also given our current tech level. My only concern and displeasure is that if the priority has been from the beginning just a game "knowing the rules" with no dreams about AI today we would probably already have WiF by years with no real harm for the AI project (that requires better machines and in general better resources not jet available at that time). But this has little to do with you or us, and much to do with Matrix policy, bearing in mind that Matrix has all the right to make the choices it likes, correct or wrong. Of course may be I'm wrong and you are able to complete this project soon and the need for a team of scientists working full time two years closely with your group of playtesters is an exaggeration. But, forgive me, I tend to believe that such a monster game will not see the light in the form we hope because commercially impossible ( where are the funds for the required team of scientist ?) if not technically impossible as well. Time will respond. [;)] The time spent so far on the AIO pales in comparison to the time spent on the map & unit data & graphics, and to the time spent implementing the sequence of play, and to the time spent on the 150+ forms, and to the time spent on the 54 optional rules in the initial release. The AIO has maybe 10 or 20 thousand lines of code. There are over 400,000 lines of code in MWIF presently. There is an old Bill Mauldin joke where Willie tells Joe to stop complaining about how heavy all his equipment is: "Stop carrying around all that unnecessary stuff; throw out the joker in your deck of cards!"

brian brian -> RE: When? (6/6/2012 5:04:04 AM)
I think the AI will first make you cry as the Russians during the Barbarossa scenario, teaching you how to run the Germans well. Then it will be a little more even as you put on the black Panzer uniforms and you use the AI's tactics against it.....until a few thousand saved games have been run back through the top-secret AI learning module (I'm making this up as I go here) and then MWIF AI 2.0 will be tenacious even when it plays the Russians. Then how it handles the navies in the Guadalcanal scenario will be the next challenge. Once it masters multiple-impulse & multiple-turn planning in that one, humanity better keep a wary eye out. I do hope to someday watch the AI play itself though. I bet some hacker will be able to accomplish this by tweaking the netplay code.

pzgndr -> RE: When? (6/6/2012 12:59:01 PM)
quote: Well, I think almost every WIF player has 'holes' in their playing ability. They forget to do things that they know they should do, or planned on doing things but didn't when the time came, or failed to set up all the prerequisites for, and therefore couldn't do. Many other events arise during a game that cause players to fail to live up to their own opinion of their playing ability/skill. Even the best professional ball player screws up from time to time. And into those gaps in the human player's execution the AI Opponent can shove long sharp objects. Bingo, +1 [&o] quote: I do hope to someday watch the AI play itself though. It is imperative that Steve develop AI-AI capability to run through many many games to assess and adjust AI behaviors; it is a laborious iterative process. I must have run hundreds of partial and full games over the past several years fine-tuning my A3R mod, and that's just for European theater operations. Global operations in WiF are more complex but the basic principles that human players apply when playing are the exact same principles that can be programmed/scripted into the AI. Throw in some variability parameters and random events, and every game becomes different. I've been genuinely surprised on numerous occassions by a computer opponent that I personally scripted; I consider that "challenging" enough. [;)]

Shannon V. OKeets -> RE: When? (6/6/2012 6:34:30 PM)
quote: ORIGINAL: pzgndr quote: Well, I think almost every WIF player has 'holes' in their playing ability. They forget to do things that they know they should do, or planned on doing things but didn't when the time came, or failed to set up all the prerequisites for, and therefore couldn't do. Many other events arise during a game that cause players to fail to live up to their own opinion of their playing ability/skill. Even the best professional ball player screws up from time to time. And into those gaps in the human player's execution the AI Opponent can shove long sharp objects. Bingo, +1 [&o] quote: I do hope to someday watch the AI play itself though. It is imperative that Steve develop AI-AI capability to run through many many games to assess and adjust AI behaviors; it is a laborious iterative process. I must have run hundreds of partial and full games over the past several years fine-tuning my A3R mod, and that's just for European theater operations. Global operations in WiF are more complex but the basic principles that human players apply when playing are the exact same principles that can be programmed/scripted into the AI. Throw in some variability parameters and random events, and every game becomes different. I've been genuinely surprised on numerous occassions by a computer opponent that I personally scripted; I consider that "challenging" enough. [;)] Speaking of surprises, I often find the game mechanics permitting/prohibiting actions that I thought were impossible/legal. For instance, the beta testers have had trouble setting up situations where the die roll for the number of partisan units that can be placed on the board is three but there is only room for two. The program reduces the number to two in those situations. But the program finds hexes you hadn't considered for the 3rd partisan (stacking is 2 per hex). Karelia and German controlled hexes now on the Russian side of the frontline are two cases. Or a hex next to a German divisional unit, and hence not in a German ZOC. Bombers in England being able to bomb factories in Italy is another case, where the bombers fly around eastern Switzerland instead of western Switzerland. Assuming carrier air units are flying as fighters, not bombers, because there are no more air activities remaining comes up often. Likewise, if a carrier air unit doesn't have the requisite bombing capability (air-to-sea, or tactical) for a mission, the program doesn't ask if you want it to fly as a bomber. Even land moves can be surprising at times. The one I miss all the time is the reduced cost for moving along rail lines in the advance after combat phase when playing with the optional rule Rail Movement Bonus. If, as a human, you can't even 'see' all the possibilities for what your (and more importantly, your enemy's) units can do, do you really expect to be invincible?

Ullern -> RE: When? (6/17/2012 2:45:16 PM)
quote: ORIGINAL: Shannon V. OKeets If, as a human, you can't even 'see' all the possibilities for what your (and more importantly, your enemy's) units can do, do you really expect to be invincible? Of course I do! What kind of question was that? [sm=sterb029.gif]

kyle8 -> AI is not impossible (6/26/2012 8:56:56 PM)
ten years ago a fantasy game came out, Age of Wonders II, Not as big as WIF, but It was a quite large and complex game. Big three level maps and up to eight opponents, complete with diplomacy, resource management etc. I found that AI to be very good. It can beat you a lot of times even an experienced opponent. Eventually you do learn it's weaknesses and it cannot overcome it's basic programming so no it will never be as good as a human who can learn and adapt. But a competent computer opponent is possible.

kyle8 -> ai (6/26/2012 9:03:04 PM)
When playing the board game I have noticed that you can sometimes win by relying on a preponderance of air power. When one side achieves air supremacy the game usually tips in their favor dramatically. So it is not necessary that the AI be balanced in it's choices. It may go for an unconventional strategy and therefore surprise the human player.

Joseignacio -> RE: ai (6/26/2012 11:49:04 PM)
True. However, you can not always win this way. You can avoid losses because you can select not to fight or you can concentrate on the missile units and thus the enemy can not retalilate whereas you can make consecutive attacks of all your surviving units on one enemy one. Before the last patches, you could also have an exploit, you got out of the fortress and the y followed you, then you got back in, and they were all the times going to and fro the same area. In the meantime if you had built the magic tower, they were hit every turn. Only with these and other exploits I could beat it most of the times, although I am a pretty experienced player in these games. Of course, I always played in the max difficulty level. Good example!

Joseignacio -> RE: ai (6/27/2012 11:15:09 AM)
I guess in this case No News are not Good News. I counted on the game being released by early July, then Steven explained that was his projected release date, then I learn Matrix will release the product when Matrix is ready to. So, finally, Steve, tje product will be ready ( in it's first version) by 1st July? If not, is there another temptative date? If yes, any clue on how can it go the release by Matrix? Good work anyway!

gridley -> RE: ai (6/28/2012 4:35:29 AM)
quote: ORIGINAL: Joseignacio I guess in this case No News are not Good News. The Months are just flying by, aren't they? [&:]

Omnius -> AI vs AI??? (6/28/2012 4:06:10 PM)
I sure hope that WiF will have AI vs AI capability. I remember the good old days of the first computer war games and they did have the AI vs AI ability. I always liked that so while I was reading the manual I could set the AI's to play the game and I could watch the action and see what was going on. I've been disappointed as the industry has gone away from this early excellent feature and hope that it will make a comeback. For scenario design that feature was extremely useful.

Omnius -> Convoys Straight??? (6/28/2012 4:09:33 PM)
I bought the original computer version of WiF done by ADG and the one thing I didn't like was how the program over thought the convoy routine. Everytime that I set it the way I wanted it would change it any time I went into that feature to look to ensure it was still set or when it came time to do the convoy resources routine. I sure hope that my reading of the improvements on the resource convoy system are correct that once set it will stay set unless convoy attacks change things.

ioticus -> RE: Convoys Straight??? (6/28/2012 8:01:18 PM)
The Matrix newsletter I just got says WIF is coming soon!

gridley -> RE: Convoys Straight??? (6/28/2012 8:28:16 PM)
quote: ORIGINAL: ioticus The Matrix newsletter I just got says WIF is coming soon! "Coming Soon" for MWiF has not meant much in the past... Anyway, I checked out the Newsletter button on the top of the page and the latest one there is Jan 2012...is there a more recent one?

ioticus -> RE: Convoys Straight??? (6/28/2012 10:56:23 PM)
quote: ORIGINAL: gridley quote: ORIGINAL: ioticus The Matrix newsletter I just got says WIF is coming soon! "Coming Soon" for MWiF has not meant much in the past... Anyway, I checked out the Newsletter button on the top of the page and the latest one there is Jan 2012...is there a more recent one? Yep, newsletter dated June 2012 says coming soon!

michaelbaldur -> RE: Convoys Straight??? (6/29/2012 1:59:25 AM)
quote: ORIGINAL: ioticus quote: ORIGINAL: gridley quote: ORIGINAL: ioticus The Matrix newsletter I just got says WIF is coming soon! "Coming Soon" for MWiF has not meant much in the past... Anyway, I checked out the Newsletter button on the top of the page and the latest one there is Jan 2012...is there a more recent one? Yep, newsletter dated June 2012 says coming soon! coming soon... can be alot of things. but as I understand there are many new informations in the monthly report. so be sure to stay tuned.

petracelli -> RE: Convoys Straight??? (6/29/2012 8:44:17 PM)
Steve Am really hoping your hard work comes to fruition in the near future, credit card at the ready! Cheers Phil

Page: << < prev 91 92 [93] 94 95 next > >>

Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

1.203125