Europe at War - great game, if it had AI

[Rocko911]

Well there is hope , the USA landed in France. I have only had 1 unit there for the past 4 turns so I am a little disappointed, but I was invading England. That went poorly and was destroyed after occupying for about 4 turns.

[Vale]

I started this thread about the passive AI. After posting, I had planned to quit the CEOW forever, but instead decided to scrap my current game and start over, this time on a slightly harder difficulty rating.

Things are going a lot better with the second game. The Allied navy actually attacked my transports (albeit with a weak force) as I prepared to invade England. On the eastern front, the Russians are making great use of air power, although they still can't punch through my line, because the AI is incapable of mustering an attack force.

The AI seems much better at tactics than strategy. For example, at one point I decided to encircle part of the Red Army. The AI was unable to react until I had completely encircled about a dozen units. I killed half a dozen of the surrounded units, and then much to my surprise the AI responded very effectively, and rescued the remaining six units.

I can't fault the AI for not responding as my units began their encircling movement. Conceptually, I don't understand how to design an AI smart enough to recognize and respond to such a maneuver. On the other hand, I seem some faults in the game which I believe can be remedied. For example, it's a no-brainer than the British should defend their little Isle, and conceptually it's easy to under stand how to do it. It's also a no-brainer that important strategic resources, such as the oil fields in the Caucasus, should be heavily defended (my little Italian strike force took them with ease). If the developers don't want to surrounded the oil fields with historically inaccurate defenses, then the AI should be capable of saying "we lost our oil?!! Get it back at ALL COST!".

To recap, here's a list of areas where I believe the AI could be easily improved:

-- better defenses for Great Britain. The British need to make use of their sea power, and they need to build ground units to repel invaders.
-- The AI needs to react more effectively when strategic resources are attacked.

I'm sure other gamers can add to the list of things to improve, but I've yet to finish a game, so I'm just reporting what I've seen so far.

[Von Manteuffel]

Yes, England needs to defend herself much better. My current game is at medium difficulty and I was able to crush England with air based in France. By 1942 I had taken England using only a few troops and it didn't even hurt me in Russia that much. My fighters even took out her over-powered carriers!

Anyway, it was too easy... The small land force I went in with should have been repulsed. I would think taking the british home isle would take a major effort. As it stands, it's a walkover and can be done concurrent with Barbarossa.

[Warfare1]

Two weak areas that need defending:

1) Station over-strength, entrenched garrison units in all British ports (even make them with "O" movement).

2) Soviet oil fields need several entrenched defenders so it's not a cakewalk for Axis to seize them.

[geozero]

So far Slitherine's only defense to this AI question has been to play the game with the settings set at max difficulty. All this does AFAIK is give the opponent more resources and production, so in a way that makes the game harder... but it DOESN'T change the way the opponent (AI) thinks about your moves, etc.

This is where clearly the ball has been dropped. AS Joe D. pointed out there needs to be If, Then, Else command type scripts programmed...not so the opponent plays exactly like history, but rather allow it to react to you better, thus making the game challenging intellectually (instead of difficult due to an opponent's immense production capability)...

JudegeDredd and Warfare1 also pointed out the absurd suggestion from Slitherine that chess AI is easy.

What Computer Chess game has a difficulty setting that allows new Queens and Knights to be built every turn??? Have you even played Computer Chess? [:D]

You have a beautiful map, clean interface and fairly fun game...it just becomes tediously absurd after a few hours. AFAIK there is no real AI whatsoever. The game can't think or make decisions based on events.

Is this game programmed in JAVA? That might explain a lot of things.

[Charles2222]

ORIGINAL: Vale

I started this thread about the passive AI. After posting, I had planned to quit the CEOW forever, but instead decided to scrap my current game and start over, this time on a slightly harder difficulty rating.

Things are going a lot better with the second game. The Allied navy actually attacked my transports (albeit with a weak force) as I prepared to invade England. On the eastern front, the Russians are making great use of air power, although they still can't punch through my line, because the AI is incapable of mustering an attack force.

The AI seems much better at tactics than strategy. For example, at one point I decided to encircle part of the Red Army. The AI was unable to react until I had completely encircled about a dozen units. I killed half a dozen of the surrounded units, and then much to my surprise the AI responded very effectively, and rescued the remaining six units.

I can't fault the AI for not responding as my units began their encircling movement. Conceptually, I don't understand how to design an AI smart enough to recognize and respond to such a maneuver. On the other hand, I seem some faults in the game which I believe can be remedied. For example, it's a no-brainer than the British should defend their little Isle, and conceptually it's easy to under stand how to do it. It's also a no-brainer that important strategic resources, such as the oil fields in the Caucasus, should be heavily defended (my little Italian strike force took them with ease). If the developers don't want to surrounded the oil fields with historically inaccurate defenses, then the AI should be capable of saying "we lost our oil?!! Get it back at ALL COST!".

To recap, here's a list of areas where I believe the AI could be easily improved:

-- better defenses for Great Britain. The British need to make use of their sea power, and they need to build ground units to repel invaders.
-- The AI needs to react more effectively when strategic resources are attacked.

I'm sure other gamers can add to the list of things to improve, but I've yet to finish a game, so I'm just reporting what I've seen so far.

I'm all for better AI, but I hope you are trying to merely make a point and instead don't actually play the way you described.

I can definitely see complaining about the AI not defending obvious targets, but do consider this. Would you have dared do such a thing with the Italians against a human? What I'm basically saying, it's as though you're playing to exploit the AI, as though you will try to do the most ludicrous things because the AI usually isn't programmed to stop ludicrous strategies.

When I played WIR for example, I didn't send a panzer corp from the northern front straight to Moscow without stopping at all. This is an example of a ludicrous strategy, whereby one tries something completely illogical in the hopes the AI doesn't account for it. In short, to play this way is to try to beat the programmer, not win the war. In playing the war your game is vastly improved by limiting yourself to things which are mostly logical. Even pulling 66% of the WIR southern forces, after turn one, and moving them to the north would in some ways be more logical than what you talk about doing with the Italians, and it's not too difficult to see that doing such a thing in WIR probably would either result in absolute disaster for Germany, or overwhleming victory. If you are playing against the programmer, instead of playing the war, then when that likely fails such a person would resort to the exact opposite, and so on and then talk about broken AI. If you did such a lame thing against me I certainly wouldn't play you again. I know when I played WIR as the USSR I certainly did pretty much abandon those oilfields, because there was no sense in leaving units down there if nothing could get to them in 3 months.

While I know it's irritating to find something completely defenseless, that should be defended, the short of it is that in order to get the AI to play more staunchly elsewhere, some places HAVE TO be left defenseless. Look over some USSR history and see just how quickly they sent forces to those oilfields because the Germans (let alone the Italians) were such a threat to take them early on. Even Moscow wasn't in such a rush to be defended from the outset, since it does take time for the enemy to get there. Why don't you just pretend those places are defended and play a bit more customary? I like better AI, as I said, but if I find the USSR can't hold Moscow because there's a large not needed force in the Caucuses because of your observation, that is truly when the AI is broken.

[Dave Ferguson]

ORIGINAL: Charles_22

.......I can definitely see complaining about the AI not defending obvious targets, but do consider this. Would you have dared do such a thing with the Italians against a human? What I'm basically saying, it's as though you're playing to exploit the AI, as though you will try to do the most ludicrous things because the AI usually isn't programmed to stop ludicrous strategies. ......

Well that sums it up for me, a lot of these strategies and techniques are designed to exploit the AI. try playing against a good allied player and the Axis might find they can't afford all those Uber weapons as they are spending the bulk of their pp in maintaining a field army. Play a human and russia can have nearly 30 corps ready for summer 41, NOT including the units on the border, plus those russian tanks are dangerous.

[Vale]

I don't accept the argument that playing the game in a non-historic fashion is exploitation. CEOW is a strategic wargame, not a simulation. Regarding the Italian attack on the oilfields, prior to sending the force on its way I had no idea how easy it would be to seize the area, as I had fog of war enabled.

Edit: sending the Italians after the oil was actually a mistake, as I now realize. Had I been playing a human opponent, I would have made the same mistake. The reason the invasion was a mistake was due to the long supply lines and the difficulty of reinforcing troops in the region. Against a human player, the Italian invasion would have ended as a humiliating defeat. Against the computer, this stupid blunder turned into a crushing victory. The Italian invasion was not about exploiting the AI. The Italian invasion was about a mistake that turned into a victory due to passive AI.

[Rocko911]

I think a resolution to our concerns are easy to do, much easier than the company wants to admit.
  1) Make subs defense go up after they are hit with an initial attack via a destroyer group by giving the player a choice to submerge and evade on contact (higher defense) or to stay on station (which means probably sacrificing your sub for a delaying action or another attack)
  2) Have England always put at least a destroyer group and a battleship stationed on the English channel side, which is very realistic during the actual war. This will make a player think twice about even doing an initial invasion.
  3) England should have more partisan units show up early if their is an invasion by Germany. This is realistic as England had one of the best Civil Defense programs in the world at that time.
 
  I will state that the games AI has surprised me and is better than my initial observations , but the above recommendations are easy to implement and would change the game play for the better.
 

[targul]

The idea of setting the defensive style of the Sub after round one of combat I like.

Forcing the British player to place a BB and sub in a place where no convoys or action will ever happen I disagree with. If the British player feels secure that England cannont be invaded it should be his choice where to place his assests.

[Warfare1]

ORIGINAL: geozero

So far Slitherine's only defense to this AI question has been to play the game with the settings set at max difficulty. All this does AFAIK is give the opponent more resources and production, so in a way that makes the game harder... but it DOESN'T change the way the opponent (AI) thinks about your moves, etc.

This is where clearly the ball has been dropped. AS Joe D. pointed out there needs to be If, Then, Else command type scripts programmed...not so the opponent plays exactly like history, but rather allow it to react to you better, thus making the game challenging intellectually (instead of difficult due to an opponent's immense production capability)...

JudegeDredd and Warfare1 also pointed out the absurd suggestion from Slitherine that chess AI is easy.

What Computer Chess game has a difficulty setting that allows new Queens and Knights to be built every turn??? Have you even played Computer Chess? [:D]

You have a beautiful map, clean interface and fairly fun game...it just becomes tediously absurd after a few hours. AFAIK there is no real AI whatsoever. The game can't think or make decisions based on events.

Is this game programmed in JAVA? That might explain a lot of things.

Is this game programmed in JAVA? That might explain a lot of things.


Why does programming in JAVA mean that certain things can or can't be implemented in the game?

[Warfare1]

ORIGINAL: LRRP

I think a resolution to our concerns are easy to do, much easier than the company wants to admit.
1) Make subs defense go up after they are hit with an initial attack via a destroyer group by giving the player a choice to submerge and evade on contact (higher defense) or to stay on station (which means probably sacrificing your sub for a delaying action or another attack)
2) Have England always put at least a destroyer group and a battleship stationed on the English channel side, which is very realistic during the actual war. This will make a player think twice about even doing an initial invasion.
3) England should have more partisan units show up early if their is an invasion by Germany. This is realistic as England had one of the best Civil Defense programs in the world at that time.

I will state that the games AI has surprised me and is better than my initial observations , but the above recommendations are easy to implement and would change the game play for the better.

I agree. I also think you are referring to actions by the AI (especially for England).

These things should be easy to implement. I understand that in the first patch the AI is being addressed and it will also come with a map and scenario editor.

[geozero]

Why must we always play WW2 historically? Why can't Germany not invade Poland, and instead attack France Sept 1 1939? Why should the French sit on t heir borders and not attack a weak Germany occupied with battle in Poland? Why can't Russia decide not to honor the treaties with Germany and invade from thea east? What if Japan never attacked Pearl Harbor? The U.S. would likely not have come into the war until much later...

It's rather dull after a while (playing AXIS for example) to attack Poland, then Denmark/Norway, then Low Countries and France, then BoB, attack Egypt, send in Afrika Korps, attack Yugo, etc etc.

Which is why diplomacy, random events (as an option), and other factors could make a strategic level WW2 game fun again.

[Warfare1]

ORIGINAL: geozero

Why must we always play WW2 historically? Why can't Germany not invade Poland, and instead attack France Sept 1 1939? Why should the French sit on t heir borders and not attack a weak Germany occupied with battle in Poland? Why can't Russia decide not to honor the treaties with Germany and invade from thea east? What if Japan never attacked Pearl Harbor? The U.S. would likely not have come into the war until much later...

It's rather dull after a while (playing AXIS for example) to attack Poland, then Denmark/Norway, then Low Countries and France, then BoB, attack Egypt, send in Afrika Korps, attack Yugo, etc etc.

Which is why diplomacy, random events (as an option), and other factors could make a strategic level WW2 game fun again.

I agree with most things.

Hitler and Stalin were opportunists.

One of the main criticisms of France and Britain was that they did nothing when Germany attacked Poland.

In addition, had Stalin attacked Germany when it was engaged in France in 1940, I think events would have been quite interesting.

For these reasons, I think there should be opportunity costs for actions taken by the Axis.

For example. If Germany attacks Poland, then Allied war effort should increase for all countries concerned (France, Britain, USA, etc). The more countries Germany attacks, the higher the war footing AND the resources/troops the Allied countries should start to receive (via events and NOT through production).

If Germany doesn't guard its border with France then MAYBE the French might invade it.

If Germany is stalemated in France or if it takes longer than it did historically, then the USSR MIGHT (random %) strike first, or at least build up its troop levels.

If Germany is attacking neutral nations then other neutral nations should adjust their position accordingly - increase spending on military, start warming to the Allies, etc....

Randomness (on a % basis) should be built into the game where it makes sense historically to do so. This will help keep the player on his toes.

Many of these things have already been done in games such as SC.

[geozero]

Warfare1 you are on the right track...all your questions start with "IF". This is the basis for a thinking AI.

What IF Turkey had come into the war? What IF Spain had not remained neutral and attacked Portugal, France or even Gibraltar? What IF Japan had not attacked the US? What IF Germany's industry had shifted earlier to rockets, jets and advanced aircraft?

What IF Churchill was assasinated? Or Hitler? Or Stalin just dies? How would these simple events "either random or triggered by some diplomatic/spy player network) affect the game on a "strategic" level?

IF, IF, IF... this is what makes playing WW2 interesting and fresh. We all know how it ended. Why would we handicap ourselves playing it again? You're better off watching a history channel re-run.

IF we can't make a difference in how events unfold, then there's no point in playing... HOI, for all it's micormanagement and other issues had it right to allow players to do nearly anything they wanted.

[Warfare1]

ORIGINAL: geozero

Warfare1 you are on the right track...all your questions start with "IF". This is the basis for a thinking AI.

What IF Turkey had come into the war? What IF Spain had not remained neutral and attacked Portugal, France or even Gibraltar? What IF Japan had not attacked the US? What IF Germany's industry had shifted earlier to rockets, jets and advanced aircraft?

What IF Churchill was assasinated? Or Hitler? Or Stalin just dies? How would these simple events "either random or triggered by some diplomatic/spy player network) affect the game on a "strategic" level?

IF, IF, IF... this is what makes playing WW2 interesting and fresh. We all know how it ended. Why would we handicap ourselves playing it again? You're better off watching a history channel re-run.

IF we can't make a difference in how events unfold, then there's no point in playing... HOI, for all it's micormanagement and other issues had it right to allow players to do nearly anything they wanted.

I get warm and fuzzy feelings when I hear "IF" questions raised for WWII events. heh

What IF Rommel had been given the supplies and troops he wanted?

What IF Germany had captured the Middle East oil?

What IF Hitler avoided attacking the USSR in 1941 and instead, built up troop levels, subs, and fought in the Middle East first?

I definitely like the idea of what Spain and Turkey would do. Actions based on a random % of how events unfold in-game.

Hopefully, Slitherine can address many of these issues. They are releasing a scenario editor with the first patch, so we'll see.

I certainly don't want a non-historical game. On the other hand, I think some random "what ifs" would make it interesting each time you play it.

What does Norway currently have that makes it worthwhile for the Axis to invade it? Does the Axis AI invade Norway?

[geozero]

Historical vs. non-hostorical has been a long time question.

Back when Avalon Hill still existed they typically advertised their games with "IF" questions. So even then, they were thinking of putting YOU (the player) into the hotseat and letting YOU decide how events would unfold...whether it was Third Reich or France 1940 games.

In a computer game it seems to me that it would be easy to put in various options the user can toggle on/off to give the whole range of historical vs non-historical situations.

This is why diplomacy, production, unit upgrades/development, etc are such key points. I remember another big board game called "Global War" by SPI. They had production, and commiting to building the wrong units was a real pain...but fun.

Does anyone remember High Command? They also made the V for Victory games...but High Command really had a lot of new features in a PC game, like the aforementioned diplomacy, etc.

You can created a whole system where resources are spent to develop new weapons or simply influence another country.

More IF's...

What IF Italy had not attacked Greece? What IF Germany's invasion of Russia had not been delayed by the Balkan wars? What IF Romania had not joined the Axis, and thus severely cut Germany's oil. What IF Sweden had joined the Axis?

It can go on and on... [:D]

[Warfare1]

I think the key idea in the "IF" discussion is to provide lots of toggle options, so the player can vary the random event situation to his liking.

This also increases the game's playability.

Many of the "IF" situations are also quite historical. What IF some things had gone in a different direction....

In SC (the original) there are lots of toggle options to vary the gaming experience.

Might be the way to go.

[geozero]

ORIGINAL: Warfare1

I think the key idea in the "IF" discussion is to provide lots of toggle options, so the player can vary the random event situation to his liking.

This also increases the game's playability.

Many of the "IF" situations are also quite historical. What IF some things had gone in a different direction....

In SC (the original) there are lots of toggle options to vary the gaming experience.

Might be the way to go.

Again you are spot on...

There could be toggles for various major events, or even a "slider" that would change the randomness factor. Almost like rolling d100. Of course the randomness can not be silly like Superpower where China attacks Uruguay or Haiti decides to attack Uganda... every country depicted must have certain loyalties...i.e. USA and Britain are moreloyal to each other, but Spain has more loyalties towards Germany, etc.

Some events might not cause too much permanent changes, just brief ones...like a sabotage event that disrupts industry or rail movement, or sets back production one or two more turns.

[Warfare1]

ORIGINAL: geozero

ORIGINAL: Warfare1

I think the key idea in the "IF" discussion is to provide lots of toggle options, so the player can vary the random event situation to his liking.

This also increases the game's playability.

Many of the "IF" situations are also quite historical. What IF some things had gone in a different direction....

In SC (the original) there are lots of toggle options to vary the gaming experience.

Might be the way to go.

Again you are spot on...

There could be toggles for various major events, or even a "slider" that would change the randomness factor. Almost like rolling d100. Of course the randomness can not be silly like Superpower where China attacks Uruguay or Haiti decides to attack Uganda... every country depicted must have certain loyalties...i.e. USA and Britain are moreloyal to each other, but Spain has more loyalties towards Germany, etc.

Some events might not cause too much permanent changes, just brief ones...like a sabotage event that disrupts industry or rail movement, or sets back production one or two more turns.

I like those ideas.

They don't increase the game's complexity; they do, however, increase the game's playability. The whole "simple to play, but difficult to master" idea.

Toggles and sliders for events and "what ifs", combined with different difficulty levels, would provide for a good level of playability.