Artifical Intelligence for World in Flames

[Shannon V. OKeets]

ORIGINAL: CBoehm

If you would like to provide me with some advice on strategic plans for any of the major powers, I would greatly appreciate it.

Im willing to give a go at the CW which happens to be my favorite country exactly because of its multitude of strategic options. - should I post it here or email it to you?

A few detail questions?!
- you state that there will be no "master allied ai" ...does that mean that conditional strategies cannot be designed for the CW, ei. do this if US do that?
- "which VPhexes to garrison" how about other important cities/resource hexes?! (a key hexes are not VPhexes).
- can conditional ai behavior be implemented depending on which optional rules are used?

One of the elements of the AIO is a decision maker who is tasked to coordinate with allies. Here is his prospectus.
===============
3.2.4 Foreign Liaison
The Foreign Liaison (FL) is charged with the responsibility of coordinating with allies on the strategic allocation of resources and operational and tactical deployment of units in the field. One way of thinking about the separation of powers between the Commander in Chief and the Foreign Liaison is the CC deals with countries that are enemies and the FL deals with those that are allies.

Lend leasing and trade agreements are the purview of the FL. Making sure that necessary actions are accomplished but not performed redundantly are also part of the FL’s job. Mostly, the Foreign Liaisons of two friendly counties act as the linkage between their respective Admiralties, Air Marshals, and Field Marshals. This frequently comes up for joining naval forces, protecting convoys, transporting units and resources, and working side by side along a joint front line. Joint air operations are another area that requires the FL’s gentle touch.

Arguably the most important task the FL performs is making recommendations to the Grand Strategist about US entry choices concerning: China, the Commonwealth and France, Atlantic convoys, and the USSR. These recommendations require the FL to assess how well the war is going for all current and potentially future combatants. Since those assessments cover the entire globe, all units of all branches of service, all national production capacities, and the territorial holdings of every major power, this is both an enormous and very detailed task. It is essential information for the Grand Strategist when reviewing and choosing the grand strategic plan.

The Foreign Liaison has no direct subordinates and reports to the Grand Strategist.
=====================

Feel free to include crucial hexes beyond those of victory hexes in the strategic plan.


The AIO has to deal with the optional rules all the time, so conditionals concerning same are to be expected.

[wfzimmerman]

The AI's mastery of gambits is an important issue.

I would define gambit as a particular operation or series of operations which, if not guarded against, can decisively tip the course of the game.

A lot of the fun of playing games like WiF is to try out gambits and have them tried out on you.

The AI should be smart about looking for the opportunity to carry out gambits against the player. It should punish you severely if you leave yourself open. However, it may be worthwhile to provide a slider control that makes the AI a little less adept at gambits for players who like a gentler trip up the learning curve.

[Shannon V. OKeets]

ORIGINAL: wfzimmerman
The AI's mastery of gambits is an important issue.

I would define gambit as a particular operation or series of operations which, if not guarded against, can decisively tip the course of the game.

A lot of the fun of playing games like WiF is to try out gambits and have them tried out on you.

The AI should be smart about looking for the opportunity to carry out gambits against the player. It should punish you severely if you leave yourself open. However, it may be worthwhile to provide a slider control that makes the AI a little less adept at gambits for players who like a gentler trip up the learning curve.

I agree with you about 'gambits' for MWIF. However, my background is in chess where the word gambit has a very well defined meaning. In chess, a gambit is a sacrifice of material for some temporal advantage, with no guarantee that the material will ever be recovered. Essentially the player gives away one or more of his men/pieces for a positional advantage. In chess a gambit is always a risk with an uncertain outcome.

As you use the term, I believe you are refering to tactical opportuniites. If you read Patrice's posts on AI for Germany in this forum, you will find that he recommends garrisoning all capitals all the time; Patrice is very concerned about defending against invasions and paradrops. The Commonwealth leaving Malta undefended, Italy having its whole navy in a major port without a land unit to defend it from invasion, paradrops behind the enemy lines in Poland that cut off all supply to the frontlines, are examples of tactical opportunities. In more common terms "Oops!" [actually more forceful explicatives are commonly used by players experiencing these events].

I would classify a Commonwealth invasion of Denmark in 1941 as a gambit. The USSR declaring war on Turkey might be another one.

In any event, your point about keeping a keen eye out for opportunities for unexpected, dramatic, tactical attacks on vulnerabilities is excellent. I am compiling a list of them for the AIO to run through every impulse.

I also liked your idea of having the AIO play at various levels of skill. I detest giving the AIO any capability/advantage that the human player doesn't have (e.g., change the combat tables, or cost of building units, or anything like that). This means that I am somewhat in a quandry as to how to mitigate its skill smoothly. Since getting the AIO to play well is an awesome task, I have given the problem of degrading its skill level virtually no thought at all. I am open to suggestions.

[Froonp]

If you read Patrice's posts on AI for Germany in this forum, you will find that he recommends garrisoning all capitals all the time; Patrice is very concerned about defending against invasions and paradrops.

Well, not all the time, only when threatened, or soon being threatened. [:)]
Threatened meaning there's a threat within striking range.

[Shannon V. OKeets]

ORIGINAL: Froonp

If you read Patrice's posts on AI for Germany in this forum, you will find that he recommends garrisoning all capitals all the time; Patrice is very concerned about defending against invasions and paradrops.

Well, not all the time, only when threatened, or soon being threatened. [:)]
Threatened meaning there's a threat within striking range.

I apologize for the misquote. Garrisoning Washington DC isn't usually warranted.

[Greyshaft]

ORIGINAL: Shannon V. OKeets
Since getting the AIO to play well is an awesome task, I have given the problem of degrading its skill level virtually no thought at all. I am open to suggestions.

To degrade the AI, I suggest make it more human [:D]

* Leave every nth enemy unit out of your calculations <akin to a careless player not noticing the enemy dispositions>
* miscalculate the gearing limits <haven't we all done that far too often!>
* indulge in riskier attacks <ooops!! overlooked that extra TAC factor I could have thrown into the battle!>

The possibilities are endless...

[Neilster]

Since getting the AIO to play well is an awesome task, I have given the problem of degrading its skill level virtually no thought at all. I am open to suggestions.

If the AI generates several possible courses of action and then selects the "best" one semi-randomly, we could increase the probability of picking a lower-rated one.

Other than that, we could introduce "noise" into the AI's calculations, inversely proportional to the skill level.

This reminds me of something I considered ages ago. As well as a skill level for the AI, what about an aggression level? This would add a significant amount of interest and replayability to the game. For example, an aggressive Wallies might attempt an invasion of GER in 1939 to assist Poland. An aggressive Axis might attempt a "France first" strategy, try risky Gib assaults or a Sealion. A less aggressive Axis might attempt a Med strategy rather than Barbarossa. The aggression could be handled in a similar way to my first suggestion. Possible courses of action are rated for risk and at a high aggression level the riskier actions are more likely to be selected.

I think the interplay between skill and risk would be interesting. Historically the Wallies were pretty risk-averse whereas the Baddies took a lot of risks early in the war.

Cheers, Neilster

[Shannon V. OKeets]

ORIGINAL: Greyshaft

ORIGINAL: Shannon V. OKeets
Since getting the AIO to play well is an awesome task, I have given the problem of degrading its skill level virtually no thought at all. I am open to suggestions.

To degrade the AI, I suggest make it more human [:D]

* Leave every nth enemy unit out of your calculations <akin to a careless player not noticing the enemy dispositions>
* miscalculate the gearing limits <haven't we all done that far too often!>
* indulge in riskier attacks <ooops!! overlooked that extra TAC factor I could have thrown into the battle!>

The possibilities are endless...

Maybe. This doesn't strike my fancy - why not, I don't know.

[Shannon V. OKeets]

ORIGINAL: Neilster

Since getting the AIO to play well is an awesome task, I have given the problem of degrading its skill level virtually no thought at all. I am open to suggestions.

If the AI generates several possible courses of action and then selects the "best" one semi-randomly, we could increase the probability of picking a lower-rated one.

Other than that, we could introduce "noise" into the AI's calculations, inversely proportional to the skill level.

This reminds me of something I considered ages ago. As well as a skill level for the AI, what about an aggression level? This would add a significant amount of interest and replayability to the game. For example, an aggressive Wallies might attempt an invasion of GER in 1939 to assist Poland. An aggressive Axis might attempt a "France first" strategy, try risky Gib assaults or a Sealion. A less aggressive Axis might attempt a Med strategy rather than Barbarossa. The aggression could be handled in a similar way to my first suggestion. Possible courses of action are rated for risk and at a high aggression level the riskier actions are more likely to be selected.

I think the interplay between skill and risk would be interesting. Historically the Wallies were pretty risk-averse whereas the Baddies took a lot of risks early in the war.

Cheers, Neilster

This suggestion has more promise but the emphasis on choosing a poor strategic plan seems excessive.

Perhaps operational missteps, such as not building enough of an important unit type - pick almost any unit type you want.

Or tactical miscues, such as leaving convoys poorly defended or making stacking mistakes leaving open the possibility of very effective ground strikes.

After a whole 6 or 7 seconds of reflection, I think maybe intentional mistakes in deciding where to deploy units could cover just about everything. If the human is clever enough to notice the holes that are created by the AIO's poor positioning, he can make progress quickly and painlessly.

[Neilster]

This suggestion has more promise but the emphasis on choosing a poor strategic plan seems excessive.

I wasn't thinking strategic; probably more operational or tactical.

What do you think of adding noise to the calculations?
What about my suggestion of an aggression level?

Cheers, Neilster

[Shannon V. OKeets]

ORIGINAL: Neilster

This suggestion has more promise but the emphasis on choosing a poor strategic plan seems excessive.

I wasn't thinking strategic; probably more operational or tactical.

What do you think of adding noise to the calculations?
What about my suggestion of an aggression level?

Cheers, Neilster

Noise doesn't give me the control I want in the degradation of performance. It may impact the results or it may not. Against a novice, I want the program to definitely give the player openings to exploit.

Changing the aggression level is very dangerous. If it is too high, for example, Germany can destroy its entire army against mediocre USSR defenses in a couple of turns. The human player doesn't really have to do anything, just wait for the AIO to vanquish itself. As I see it that would not provide much of a learning experience for the novice. I want the beginner to have a lot of opportunities for success, not be punished rigorously for every mistake, and become a better player after each game he plays. Yeah, I know, it's ambitious. But bad choice of builds + bad positioning + failure to exploit every tactical mistake can do that; or at least I think it can.

[c92nichj]

I know that you will not like this suggestion but I think that the AIO should have a higher likelyhood of rolling good die rolls. I wouldn't want ito always get +1 to the land combat rolls but maybe it should have 5% instead of 10% chance of rolling a 1. I would not like it to have increased production or cheaper units, but wouldn't mind it to be a litlle bit luckier than I am.

[wfzimmerman]

ORIGINAL: c92nichj

I know that you will not like this suggestion but I think that the AIO should have a higher likelyhood of rolling good die rolls. I wouldn't want ito always get +1 to the land combat rolls but maybe it should have 5% instead of 10% chance of rolling a 1. I would not like it to have increased production or cheaper units, but wouldn't mind it to be a litlle bit luckier than I am.

why?

[Shannon V. OKeets]

ORIGINAL: c92nichj

I know that you will not like this suggestion but I think that the AIO should have a higher likelyhood of rolling good die rolls. I wouldn't want ito always get +1 to the land combat rolls but maybe it should have 5% instead of 10% chance of rolling a 1. I would not like it to have increased production or cheaper units, but wouldn't mind it to be a litlle bit luckier than I am.

Please visualize me making the sign of the cross and muttering Latin incantations to ward off evil. I will pray for you, my son.

[Froonp]

ORIGINAL: Shannon V. OKeets

ORIGINAL: c92nichj
I know that you will not like this suggestion but I think that the AIO should have a higher likelyhood of rolling good die rolls. I wouldn't want ito always get +1 to the land combat rolls but maybe it should have 5% instead of 10% chance of rolling a 1. I would not like it to have increased production or cheaper units, but wouldn't mind it to be a litlle bit luckier than I am.

Please visualize me making the sign of the cross and muttering Latin incantations to ward off evil. I will pray for you, my son.

Same.
I would hate it.
anyway, I intend to play TCP/IP versus my best friend.
Cheers !

[pak19652002]

Nicklas' admirable, but ultimately futile trial balloon underscores player sentiment that the AI will be...how to put this delicately..."peformance challenged" despite best efforts to the contrary. I was surprised by the feedback posted on this topic on the Online WiF board.

Peter

[SamuraiProgrmmr]

I have been following the posts about the AI for quite some time. I also have been following the posts in the Yahoo WIF group about the same subject.

I have a few comments.

First, do NOT listen to the naysayers!!!!!

THEY said man could not fly.

THEY said man could not reach the moon.

THEY said a computer would never beat a grandmaster at chess.

THEY WERE WRONG.

And now they are saying that a computer will NEVER be able to play a good game of WIF.

We will never know until it is tried.

I hope that the first complaint you get about the game is "The AI CHEATS" or The AI is too hard".

I hope that you are bombarded with questions about what free help you have given the AI.

I hope you succeed and I am pulling for you.

It is a daunting task. Of that there is no doubt. But, as a professional programmer with 20 years experience, I can tell you this. I would not have dreamed that Windows and the internet would accomplish what they have. The only think it took was willpower and money. Perhaps this is the time for WiF.

Go for it! Write a kick a$$ AI and make them run with their tail tucked between their legs.

I believe! [&o]


Dean

[Froonp]

Yeahhhhh Samurai !!!! Well said !!!
Go go go Steve !!!!
[&o]

[terje439]

ORIGINAL: SamuraiProgrammer
I hope that the first complaint you get about the game is "The AI CHEATS" or The AI is too hard".

I agree with you on the wish for a "The AI is too hard"-comment, but I certainly would not want a "cheating" AI. If I loose a game because the AI is good/me not good enough that is great, that means I will go back and try again and again until I get it right.

A cheating AI on the other hand is something I truely hate, to loose a game because the AI can build cheaper units in a shorter time as an example is something that would make me play the game once.

[SamuraiProgrmmr]

No, I don't want an AI that cheats either.

What I meant was that they do such a good job at creating the AI that everyone is convinced it is cheating.

Hope this helps.