AI vs Rules

[bo]

Probably a dumb question but what the heck I have asked them before[;)] Absolutely do not want to stir the AI contraversy again [you know AI or no AI] but this is gnawing at me it seems to me [assuming [&:]] that for Steve to incorporate the 4,000 rules into code[:@] why not do just some of the rules now and add some later, a few rules and playable is to me better than all the rules and two years down the road before all the rules are done and the game comes to fruitation. I would still prefer a good AI over anything but thats just me, the AI seems very important to many here and to a few, less important. To the purist here I guess I have suggested a trip to hell and I should rot in it. But in reality if Steve is going to do the AI and he said he was than do it and put a lot of the rules on the back burner[;)] I am not trying to be smart but sometimes common sense [ by Thomas Paine] suggests we cant do everything at once. What should be the priorty here?

Bo

[michaelbaldur]

without revealing to much

we have already removed some non critical options rules.

[MilRevKo]

AI is a pipe dream. There is no computer program in the world that can run a game of this scale. It really is that simple.

[bo]

ORIGINAL: michaelbaldur

without revealing to much

we have already removed some non critical options rules.

Thank you again as always Michael, but there is that mysterious felling I am getting again "without revealing to much" [:-] think of me as an investigative reporter Michael wanting to know more, I am probably on the lowest rung of the famous "need to know".
Ok moving on I would think in my little mind that the major holdup of this game is instituting all of the rules of this game and having to debug them constantly, right, wrong or NDA[&:] I admit I know nothing about programming but it is obvious even to a noob like me that something is holding this game up, maybe just maybe you should do away with some critical rules for the moment and add them later[;)]

Bo

[bo]

ORIGINAL: MilRevKo

AI is a pipe dream. There is no computer program in the world that can run a game of this scale. It really is that simple.

I agree 100% MilRevKo with you in fact no AI could come close to beating any of the seasoned veterans that have played the board game for years, so are you saying dont do an AI?

Bo

[Shannon V. OKeets]

ORIGINAL: bo

ORIGINAL: michaelbaldur

without revealing to much

we have already removed some non critical options rules.

Thank you again as always Michael, but there is that mysterious felling I am getting again "without revealing to much" [:-] think of me as an investigative reporter Michael wanting to know more, I am probably on the lowest rung of the famous "need to know".
Ok moving on I would think in my little mind that the major holdup of this game is instituting all of the rules of this game and having to debug them constantly, right, wrong or NDA[&:] I admit I know nothing about programming but it is obvious even to a noob like me that something is holding this game up, maybe just maybe you should do away with some critical rules for the moment and add them later[;)]

Bo

Doing away with a rule to remove complexity is like saying a large bowl of spaghetti would be organized if you removed one noodle.

[wworld7]

ORIGINAL: Shannon V. OKeets

ORIGINAL: bo

ORIGINAL: michaelbaldur

without revealing to much

we have already removed some non critical options rules.

Thank you again as always Michael, but there is that mysterious felling I am getting again "without revealing to much" [:-] think of me as an investigative reporter Michael wanting to know more, I am probably on the lowest rung of the famous "need to know".
Ok moving on I would think in my little mind that the major holdup of this game is instituting all of the rules of this game and having to debug them constantly, right, wrong or NDA[&:] I admit I know nothing about programming but it is obvious even to a noob like me that something is holding this game up, maybe just maybe you should do away with some critical rules for the moment and add them later[;)]

Bo

Doing away with a rule to remove complexity is like saying a large bowl of spaghetti would be organized if you removed one noodle.

An economic question using data from this thread (ie: a fantasy "what if")

Current postion:
Fantasy constant #1: MWIF is released in two (2) years including all "4000 rules" (YIPPEE).
FC #1a: It takes Steve 5000 hours of programming to complete the project.

Fantasy Proposal #1: MWIF-LIGHT is released in one (1) year but only included "2000" rules (Yip...OMG).
FP#1a: It takes Steve 2500 hours of programming to compete the project.

Fantasy Proposal #2: MWIF-LIGHT is upgraded to MWIF and released 1.5 years after MWIF-LIGHT (with all "4000" rules).
FP#2a: It takes Steve 3750 hours of programming to merge the missing "2000 rules" to bring MWIF to total completion (Yippee).

Explaination of timeframe for FP#2a or why did it take Steve longer to add in the missing "2000 rules"? The extra time is needed to add, blend, splice, connect NEW code to OLD code (it takes a lot of time and much more testing). It is quicker to program something originally then to add it in later. This is one of the reasons why "feature creep" can destroy projects IF the project manager allows it to go unchecked.

RESULT: FP#1 gets us the finished MWIF product in two (2) years (YIPPEE).

RESULT: FP#2 gets us "something" in a year and then the finished MWIF 1.5 years later (YIPPEE).

Here is my question: If FP#2 is used who pays Steve for the additional .5 year of programming work?
A. He does it for free (YIPPEE...sorry[:'(][:'(][:'(])
B. MWIF-LIGHT price is increased by $50.00 (ie: original MWIF cost + $50.00).
C. The completed MWIF is sold as an add-on for $50.00 (not a good idea IMO as I generally dislike this kind of policy).
D. The first sixty (60) buyers send Steve $500.00 each (please send this to me and I'll pass it on to Steve (- a small fee [:)])).

[bo]

ORIGINAL: Shannon V. OKeets

ORIGINAL: bo

ORIGINAL: michaelbaldur

without revealing to much

we have already removed some non critical options rules.

Thank you again as always Michael, but there is that mysterious felling I am getting again "without revealing to much" [:-] think of me as an investigative reporter Michael wanting to know more, I am probably on the lowest rung of the famous "need to know".
Ok moving on I would think in my little mind that the major holdup of this game is instituting all of the rules of this game and having to debug them constantly, right, wrong or NDA[&:] I admit I know nothing about programming but it is obvious even to a noob like me that something is holding this game up, maybe just maybe you should do away with some critical rules for the moment and add them later[;)]

Bo

Doing away with a rule to remove complexity is like saying a large bowl of spaghetti would be organized if you removed one noodle.

Without revealing too much Steve wouldn't you say it is the rules and all of their complexity that keeps you programming so many hours and probably loss of sleep, ok keep all the rules see if I care[:D] three noodles would organize anything anywhere[;)].

Bo

[bo]

Aw shucks my post made Flipper flip, could you please explain what you meant to say in
layman terms pretty please[&o]

Bo

[wworld7]

ORIGINAL: bo

Aw shucks my post made Flipper flip, could you please explain what you meant to say in
layman terms pretty please[&o]

Bo

Bo,

I didn't flip, I tried (spent 25 minutes) to make the scenario "simple". I don't see any terms that a layman can't understand especially for areas pertaining to programming. Maybe I should have tried using the bowl of pasta but my guess is that if I used sauces and cheese as variables you (me or anyone else) wouldn't be able to grasp the concepts as clearly as I laid out above.

Anyway, everyone have a Happy 4th of July weekend. GO USA...oops...GO BRAZIL...oops...Go GHANA...oops...

This World Cup has been a bit crazy (anyone seen the French team? Or are they still drinking to forget this year?)

[bo]

ORIGINAL: flipperwasirish

ORIGINAL: bo

Aw shucks my post made Flipper flip, could you please explain what you meant to say in
layman terms pretty please[&o]

Bo

Bo,

I didn't flip, I tried (spent 25 minutes) to make the scenario "simple". I don't see any terms that a layman can't understand especially for areas pertaining to programming. Maybe I should have tried using the bowl of pasta but my guess is that if I used sauces and cheese as variables you (me or anyone else) wouldn't be able to grasp the concepts as clearly as I laid out above.

Anyway, everyone have a Happy 4th of July weekend. GO USA...oops...GO BRAZIL...oops...Go GHANA...oops...

This World Cup has been a bit crazy (anyone seen the French team? Or are they still drinking to forget this year?)

Simple is as simple does Flipper[;)], wow 25 minutes that seems like a lifetime to me after you have waited 5 years at least eyeing these posts I think you are inhaling things in your Vuvuzela instead of blowing into it [:D] I thought my question was right on [of course I always think that [&:]]from the gist of the posts from Steve it seems the coding of these rules is eating up quite a bit of his time, Steve can correct me if this coding of rules is a minor thing. I wasnt trolling I asked a legitimate question and I get hyperbole [is that a word did not go to college so forgive me] 4000 rules 400 rules 40 rules whose counting, I as a novice think that is what is causing this delay. It has to be mind boggling to code these rules [a compliment Steve[&o]] I think Spain is going to win this farce, they should cut the field in half so it looks like a game. Sorry about the Vuvuzela remark [well not really] always remember Mr Cornhusker you are my favorite poster[:)]

Bo

[brian brian]

I think Bo is on to something here, and indeed the first few decisions on which rules options to exclude from the AI have already been made.

It probably would be easier to program the AI to play "WiF Classic" than it would be to play "WiF SuperDeluxe".

It takes a long time and a lot of game-playing to realize this, but more than a few steady players of the game move back from the cutting edge of adding every possible ADG kit to a pile of now more than 6,000 counters and just play Classic these days. Possibly Steve would disagree here, that the programming would be the same. But it seems intuitive to me that the AI will perform better at the lower counter densities and thus fewer decision points of a Classic game as opposed to one with light cruisers, construction engineers, oil tank farms, and sorting out things like this for example: gun divisions that are heavy/light/static/towed/self-propelled/sometimes-doubled/sometimes-not. Ditto on which types of divisions can interact with which types of ships of planes.

[bo]

ORIGINAL: brian brian

I think Bo is on to something here, and indeed the first few decisions on which rules options to exclude from the AI have already been made.

It probably would be easier to program the AI to play "WiF Classic" than it would be to play "WiF SuperDeluxe".

It takes a long time and a lot of game-playing to realize this, but more than a few steady players of the game move back from the cutting edge of adding every possible ADG kit to a pile of now more than 6,000 counters and just play Classic these days. Possibly Steve would disagree here, that the programming would be the same. But it seems intuitive to me that the AI will perform better at the lower counter densities and thus fewer decision points of a Classic game as opposed to one with light cruisers, construction engineers, oil tank farms, and sorting out things like this for example: gun divisions that are heavy/light/static/towed/self-propelled/sometimes-doubled/sometimes-not. Ditto on which types of divisions can interact with which types of ships of planes.

Thank you Brian [on to what] look I am just fishing for some answers nothing negative intended, see I dont understand classic from smassic[&:] but in trying to use common sense all those rules are creating program problems within themselves, I never thought of the rules before in this light good lord every rule most likely has such a variation on game play it would boggle any ones mind. Probably too late but really why not cut a lot of the rules out, some one said in a previous post that tha AI testing hasn't even begun, dont know whether thats correct or not, but if it has not we are in for a looooooong wait here.

Bo

[wworld7]

ORIGINAL: bo

ORIGINAL: flipperwasirish

ORIGINAL: bo

Aw shucks my post made Flipper flip, could you please explain what you meant to say in
layman terms pretty please[&o]

Bo

Bo,

I didn't flip, I tried (spent 25 minutes) to make the scenario "simple". I don't see any terms that a layman can't understand especially for areas pertaining to programming. Maybe I should have tried using the bowl of pasta but my guess is that if I used sauces and cheese as variables you (me or anyone else) wouldn't be able to grasp the concepts as clearly as I laid out above.

Anyway, everyone have a Happy 4th of July weekend. GO USA...oops...GO BRAZIL...oops...Go GHANA...oops...

This World Cup has been a bit crazy (anyone seen the French team? Or are they still drinking to forget this year?)

Simple is as simple does Flipper[;)], wow 25 minutes that seems like a lifetime to me after you have waited 5 years at least eyeing these posts I think you are inhaling things in your Vuvuzela instead of blowing into it [:D] I thought my question was right on [of course I always think that [&:]]from the gist of the posts from Steve it seems the coding of these rules is eating up quite a bit of his time, Steve can correct me if this coding of rules is a minor thing. I wasnt trolling I asked a legitimate question and I get hyperbole [is that a word did not go to college so forgive me] 4000 rules 400 rules 40 rules whose counting, I as a novice think that is what is causing this delay. It has to be mind boggling to code these rules [a compliment Steve[&o]] I think Spain is going to win this farce, they should cut the field in half so it looks like a game. Sorry about the Vuvuzela remark [well not really] always remember Mr Cornhusker you are my favorite poster[:)]

Bo

Bo,

I didn't say your idea was bad. And rather than just dismissing your idea, I took the long route and tried to give an example with some of the factors involved if it was implemented and the cost of such an action. Perhaps, some you may not have thought of or been aware of (like what does it do to Steve's compensation/per hour). As for 25 minutes, hey I am a terrible at spelling and tried to make my post correct.

I guess you (and some others) want a less complete MWIF pushed out the door, no matter the effect on the finished MWIF product (some others of us want). It isn't a right or wrong to want either, especialy since we don't get to make the choice on what actually is produced.

Also, as for my 25 minutes it has no connection to your 5 year wait, so I won't comment on that.

Happy Holiday,

[Shannon V. OKeets]

ORIGINAL: brian brian

I think Bo is on to something here, and indeed the first few decisions on which rules options to exclude from the AI have already been made.

It probably would be easier to program the AI to play "WiF Classic" than it would be to play "WiF SuperDeluxe".

It takes a long time and a lot of game-playing to realize this, but more than a few steady players of the game move back from the cutting edge of adding every possible ADG kit to a pile of now more than 6,000 counters and just play Classic these days. Possibly Steve would disagree here, that the programming would be the same. But it seems intuitive to me that the AI will perform better at the lower counter densities and thus fewer decision points of a Classic game as opposed to one with light cruisers, construction engineers, oil tank farms, and sorting out things like this for example: gun divisions that are heavy/light/static/towed/self-propelled/sometimes-doubled/sometimes-not. Ditto on which types of divisions can interact with which types of ships of planes.

Even if you stick to just WIF Classic there are a lot of rules. Consider the following (for debugging naval movement code):

1 - Move an individual naval unit during the Naval Movement phase (unlimited naval moves permitted) and test that it can correctly reach each sea area and port destination. Include the rules about passing Gibraltar/Skaggerak/Panama/Suez/London, going through Kiel, stopping in Brest even if London is not held.
2 - Move multiple naval units as a stack for the same as #1.
3 - Do all the above during a combined action and keep track of naval moves.
4 - Do all the above using submarines (which can go pass Gibraltar and only require 1 naval move if they all start from the same hex).
5 - Do all the above, then undo moves and see if the units return to the correct locations. The number of available naval moves increases when a move is undone.
6 - Do all the above including naval transports and have the TRS pick up units from port (both at start and when passing through a port) and from a sea area (both when starting in a sea area and when ending a move in a sea area).
7 - Repeat all of the above for the Returning to Base phase, forced naval rebases due to overruns, conquest, liberation, and Vichy France formation and collapse, and aborts from naval combat.
8 - Now repeat all of the above when enemy units can conduct naval interceptions. Once an interception is possible, undoing any moves is prohibited. Consider the following cases:
8a - no interception attempted. Moving units stop in the sea area (they can now load units from coastal hexes).
8b - no interception attempted. Moving units keep moving (perhaps after dropping off some units - dropped off units can load units from coastal hexes).
8c - interception attempted but failed. Moving units stop in the sea area (they can now load units from coastal hexes).
8d - interception attempted but failed. Moving units keep moving (perhaps after dropping off some units - dropped off units can load units from coastal hexes).
8e - interception succeeded and moving side stops in the sea area (they can now load units from coastal hexes).
8f - interception succeeded and moving side fights through (either voluntarily or forced to do so by the rules). This variation has several logic paths depending on the outcome of the naval combat.
9 - For all the above, consider that multiple players can be moving units simultaneously during a NetPlay game.

There are rules about weather, DOW surprise, neutrality, cooperation betweenn countries, control of minor countries, and foreign troop commitments that have to be considered in the above.
---
Given all this, adding in that SCS units can carry divisions, restrictions on the what the Queens and AMPHs can carry, and Rough Seas are minor additions.

P.S. I might have forgotten to mention some things. But I do know that all of the above have to be considered and coded (it is at least 10,000 lines of code). Note that I haven't even gotten into the details of naval combat in the above list.

[bo]

Flipper if you and I cant rag each other in good nature I have to give up on these posts and this game[:)] I took no offence to anything you said, actually it made sense but that does not solve the problem of too much coding for Steve to do, and trully I would like a complete game with all the rules intact, I know you want a complete game and there is nothing wrong with that but at what cost to players patience in the coming years.

Bo

[bo]

ORIGINAL: Shannon V. OKeets

ORIGINAL: brian brian

I think Bo is on to something here, and indeed the first few decisions on which rules options to exclude from the AI have already been made.

It probably would be easier to program the AI to play "WiF Classic" than it would be to play "WiF SuperDeluxe".

It takes a long time and a lot of game-playing to realize this, but more than a few steady players of the game move back from the cutting edge of adding every possible ADG kit to a pile of now more than 6,000 counters and just play Classic these days. Possibly Steve would disagree here, that the programming would be the same. But it seems intuitive to me that the AI will perform better at the lower counter densities and thus fewer decision points of a Classic game as opposed to one with light cruisers, construction engineers, oil tank farms, and sorting out things like this for example: gun divisions that are heavy/light/static/towed/self-propelled/sometimes-doubled/sometimes-not. Ditto on which types of divisions can interact with which types of ships of planes.

Even if you stick to just WIF Classic there are a lot of rules. Consider the following (for debugging naval movement code):

1 - Move an individual naval unit during the Naval Movement phase (unlimited naval moves permitted) and test that it can correctly reach each sea area and port destination. Include the rules about passing Gibraltar/Skaggerak/Panama/Suez/London, going through Kiel, stopping in Brest even if London is not held.
2 - Move multiple naval units as a stack for the same as #1.
3 - Do all the above during a combined action and keep track of naval moves.
4 - Do all the above using submarines (which can go pass Gibraltar and only require 1 naval move if they all start from the same hex).
5 - Do all the above, then undo moves and see if the units return to the correct locations. The number of available naval moves increases when a move is undone.
6 - Do all the above including naval transports and have the TRS pick up units from port (both at start and when passing through a port) and from a sea area (both when starting in a sea area and when ending a move in a sea area).
7 - Repeat all of the above for the Returning to Base phase, forced naval rebases due to overruns, conquest, liberation, and Vichy France formation and collapse, and aborts from naval combat.
8 - Now repeat all of the above when enemy units can conduct naval interceptions. Once an interception is possible, undoing any moves is prohibited. Consider the following cases:
8a - no interception attempted. Moving units stop in the sea area (they can now load units from coastal hexes).
8b - no interception attempted. Moving units keep moving (perhaps after dropping off some units - dropped off units can load units from coastal hexes).
8c - interception attempted but failed. Moving units stop in the sea area (they can now load units from coastal hexes).
8d - interception attempted but failed. Moving units keep moving (perhaps after dropping off some units - dropped off units can load units from coastal hexes).
8e - interception succeeded and moving side stops in the sea area (they can now load units from coastal hexes).
8f - interception succeeded and moving side fights through (either voluntarily or forced to do so by the rules). This variation has several logic paths depending on the outcome of the naval combat.
9 - For all the above, consider that multiple players can be moving units simultaneously during a NetPlay game.

There are rules about weather, DOW surprise, neutrality, cooperation betweenn countries, control of minor countries, and foreign troop commitments that have to be considered in the above.
---
Given all this, adding in that SCS units can carry divisions, restrictions on the what the Queens and AMPHs can carry, and Rough Seas are minor additions.

P.S. I might have forgotten to mention some things. But I do know that all of the above have to be considered and coded (it is at least 10,000 lines of code). Note that I haven't even gotten into the details of naval combat in the above list.

Yo Steve I am totally convinced that you have a monumental job on your hands, I did not and will not say anything negative about your work again on these posts, I only thought if less rules were involved you would be able to move on and get this game out, it was not negative in any way, people have to remember me and some other people are total noobs when it comes to this game, you and others can explain to the cows come home about coding, rules, etc. to us neophytes and some of it sinks in and some of it goes right over our heads [well maybe only me] when you look at the posters compared to a year ago there is a big drop not maybe in posts because the same posters post over and over and over but they are the diehards the rest have gone on to other games. They drop in now and then to check out your reports and dont come back till next month[maybe] I am also playing solitare with CWIF[no other choice] and even with the die roll changing combat situations it is boring to say the least. I have to be honest I am losing heart for this game.

Bo

[wworld7]

ORIGINAL: bo

Flipper if you and I cant rag each other in good nature I have to give up on these posts and this game[:)] I took no offence to anything you said, actually it made sense but that does not solve the problem of too much coding for Steve to do, and trully I would like a complete game with all the rules intact, I know you want a complete game and there is nothing wrong with that but at what cost to players patience in the coming years.

Bo

Bo,

I am not into ragging on each other, I just didn't want you to think I (not that my opinion is important) was dismissing your idea without providing a explaination as to why (and I only covered a couple of the many reasons it could lead to problems). It wasn't meant to solve the problem you think Steve has with coding "too much". Steve has a project plan (I'm sure of this) which includes X hours of coding/testing time (NOTE: X is a variable as it needs to handle issues and changes in a project). Eliminating say 50% of the rules would not automatically reduce this time by 50%. Since coding and testing are forged together (in a good project plan) an option could be the remainng 50% of the rules are given this extra time for more fixing and better testing before release. Maybe I should have used this yesterday as it is simpler.

Also, don't put too much weight into the number of people posting, many people are following this project just by checking this forum every month or two. It doesn't mean they have given up on it, it is just a way to keep tabs until the project is finished (that's how I followed CWIE project (Computer War in Europe) by Decission Games).

Anyway, I'm tired and need to sleep.

Have a nice weekend.

[pzgndr]

I agree 100% MilRevKo with you in fact no AI could come close to beating any of the seasoned veterans that have played the board game for years, so are you saying dont do an AI?

I can't speak for others, but for me having a challenging computer opponent for solo gaming at my convenience and for my entertainment is the desired goal. I get to define what convenience and entertainment mean for me, nobody else does. I do not expect the AI to be brilliant or capable of challenging seasoned veterans; that is not the goal. But it should be functional and capable enough to satisfy casual gamers. Plus with some difficulty settings and other handicaps, it is quite possible to have an AI that provides a challenging game.

It's different than playing a seasoned veteran of course, but can still be enjoyable. As I often enjoy playing boardgames solo, where I play both sides and let the dice decide how things go. It's just different. What continues to amaze me is how some folks want to dictate to others how they are supposed to play and enjoy games like this, as if there's only one way or the highway. Wrong.

Brilliant AI may be a pipe dream but it's still a goal worth shooting for. We already have games about as complicated as possible, with wonderful eye-candy and sound effects, and lots of bells and whistles to appease those who want everything. The great appeal of computer wargames was always the prospect of a challenging computer opponent, ready and willing whenever you are. Still true today.

[Zorachus99]

If you don't mind, I'd like to add a few noodles to the bowl of spaghetti. 

As a beta tester for some years now, I'm very encouraged at the steady progress Steve has made.  The codebase that was inherited from Windows 98 had to be merged with modern technology so that it would behave properly. I confess to screwing things up by bringing up architectural issues that needed remedy, so that the game would work on future platforms.  Frankly, Steve is some kind of genuis for unraveling assembly someone else wrote, and then writing documented functioning routines.

As far as AI goes, I have a lot of faith in some of Steve's ideas.  Take some time to read all of what Steve's written.  I suspect he could pull off a fearful AI.

Cheers & Happy 4th!