RE: is the braindead AI going to be solved??? (Full Version)

All Forums >> [New Releases from Matrix Games] >> Empires in Arms the Napoleonic Wars of 1805 - 1815



Message

lavisj -> RE: is the braindead AI going to be solved??? (1/4/2008 4:32:07 PM)

quote:

ORIGINAL: Mynok

It would be far more rational and less expensive simply to recreate the UMP rules of the original board game....where guess what? Yep...nations are controlled by humans.

If you don't like playing against humans, you bought the wrong game. Sorry to break it to you.



Mynok,

If that game is just a PBEM platform why should I spend 60 dollars to play PBEM when I can do that for free with Cyberboard or Vassal?
Is the only advantage that the rules are set in stone (which means we now have to play exclusively with the designers interpretation), and that the dice rolls and calculations are made for us? But in a PBEM you already have plenty of time, so doing the calculations is really not much of an issue.

So I would say that for this game to be viable it need to offer more than what free programs already do... and that is an AI. Otherwise the game seems close to worthless to me. This is also I think why the game is being marketed as having an AI and not as a simple PBEM platform. Otherwise it probably would not sell.

Interestingly it would be interesting to know how many copies have sold so far.



Jimmer -> RE: is the braindead AI going to be solved??? (1/4/2008 5:25:55 PM)


quote:

ORIGINAL: Mynok


quote:

Computers cannot think. They don't have personalities. They aren't human.

This is a good point. I have a good friend (who happens to post on this board as "gwheelock") that takes this "human interaction" part and puts it to good use. If a person EVER backstabs him in a game, his focus changes to the destruction of that player's chances of winning. He pulls out all the stops.

Typically, computer games never do this (although, it wouldn't be that hard to code in). But, many humans do. And, if another human is playing with gwheelock, s/he needs to know that it is a singularly bad idea to backstab him early in a game.

Then, there's people like me: I like to try to play the game historically. As Russia, I hate the Turks, and I use the Prussians. And, I don't care much for the Austrians, either, but I'll play nice with them. At least early. GB had better think long and hard about making Russia an enemy (THIS Russia has a long memory!) France? I hate him, but if someone closer goes against my wishes, sure, I'll join him. Fully expecting him to backstab me at some point, and preparing for that day.

As GB, I want to control the world. "Make the world British. SIR!" That means controlling all of the other nations, at least partially. With money, I buy Turkish favors. With money I assist Prussia, Austria, and Russia against those disgusting Frenchies. All other nations exist to serve the empire upon which the sun never sets.

As Spain, I just want to be left alone. And, of course, in control of Portugal, Morocco, Naples, etc. :) Eventually, France and GB will recognize the threat I present, but I just hope the rest of the world has grown up enough to take care of them for me. :)

As Prussia, I hate Napoleon. Not "I hate France". I hate NAPOLEON. With a bitter and undying passion. Nothing else matters. France is fine, as long as Napoleon isn't there. Oh, and she should get back to her 1796 borders, too.

There are other players that can be best described as "erratic". These are the ones who go to war at the drop of a hat. Computers can sometimes emulate this, but computers don't also PLAY erratically. It requires erratic play to be successful with erratic war declarations.

Some players are superstitious. To a computer, this would be "random chance" factoring in. But, there's this one guy: He tosses a handful of dice before a roll, and then picks out the ones that are sixes, figuring they are on a roll. The only problem is, it seems to work. His last name sounds similar to Heisenberg (of Physics' "Heisenberg Uncertainty Principle" fame). We call this fellow's dice rolling (replacing Heisenberg's name with his name) "The Heisenberg CERTAINTY Principle". This is because while his dice rolling is normal most of the time, when a critical event occurs, he can almost always roll a six at just the right time. We've nicknamed HIM "Certainty Principle" because of this.

Some players are brilliant planners. Some are brilliant strategic thinkers. Some are brilliant tacticians.

I would think that the perfect AI, would incorporate some of the above:


  • Brilliant month-to-month planner
  • Brilliant strategic thinker
  • Brilliant tactician
  • Lucky
  • Unlucky
  • Manipulative
  • Historical
  • etc.

Unfortunately, I fully realize just how hard such things would be to code. But, that's what humans bring to the table.



Jimmer -> RE: is the braindead AI going to be solved??? (1/4/2008 5:29:07 PM)

quote:

ORIGINAL: Murat

EiANW II will rock!

As long as they don't do what Hasbro (Atari ? ) did with Axis & Allies: Ruin the game by making it into a real-time simulation.

It seems to me that the guys at Matrix understand a lot about this game. As a result, I doubt they would ever make that mistake. I wish I could work for them.



joviel -> RE: is the braindead AI going to be solved??? (1/4/2008 7:20:26 PM)

I'm not happy at all with the manual, the bugs or the AI.  But I went back to some of the forum posts from when the game was in development and people were begging, literally begging for the developer to release it as was and improve it later.  So the poor developer gets put into a no-win situation.  Do I think the game was released half-baked?  Yes.  Did we ask for it?  Yes.  Will Marshal fix it?  He says so.

So I'll knash my teeth at a PBEM system which corrupts most games , at a manual written only for experienced players and at an AI which a drunk Spanish general could beat.  And I'll wait and hope the Marshal doesn't get so discouraged that he joins the French Foreign Legion.  [&o]

Scott



SamuraiProgrmmr -> RE: is the braindead AI going to be solved??? (1/4/2008 7:28:22 PM)

The begging, pleading, and whining about 'release it as it is' have certainly put Marshall between a rock and a hard place!

I too, will be patient and see how things improve.

Hang in there Marshall!




Adraeth -> RE: is the braindead AI going to be solved??? (1/4/2008 7:49:46 PM)

This game has astonishing potential, so keep the ball rolling and let the guns smoking Marshall [;)]



ess1 -> RE: is the braindead AI going to be solved??? (1/4/2008 8:26:35 PM)


quote:

ORIGINAL: dauphan129

quote:

There will NEVER be an AI for this game that is good for anything more than training until we have true artificial intelligence.


I disagree with this statement in that the game could be programmed to follow certain paths. Perhaps have three or four profiles for each nation in which the AI prioritizes it's options in different ways, I just think it would be exceptionally hard to do.

>> snip

Still, this game was designed to be played Multi-player, all Humans. I think someday it will be a great single player game too but that was not the initial design and I think it is awesome for what it is.


Here here.



Jimmer -> RE: is the braindead AI going to be solved??? (1/4/2008 10:56:01 PM)


quote:

ORIGINAL: joviel

I'm not happy at all with the manual, the bugs or the AI.  But I went back to some of the forum posts from when the game was in development and people were begging, literally begging for the developer to release it as was and improve it later.  So the poor developer gets put into a no-win situation.  Do I think the game was released half-baked?  Yes.  Did we ask for it?  Yes.  Will Marshal fix it?  He says so.

So I'll knash my teeth at a PBEM system which corrupts most games , at a manual written only for experienced players and at an AI which a drunk Spanish general could beat.  And I'll wait and hope the Marshal doesn't get so discouraged that he joins the French Foreign Legion.  [&o]

Scott

I wasn't here for the debates (didn't know about it until a coincidence got me to this web site -- on release day!). But, if I had been, I would have joined that same chorus. We can get the tweaks later. Let's play!

However, I do believe it would have been a good idea to prominently display this information on the "buy it now" page. I'm not sure how one writes it, such that half the customers don't fly, but I think something should have been there.

But, along with the rest of the supporters, I'll tip my hat to Marshall and the Matrix team.



Fflewddur -> RE: is the braindead AI going to be solved??? (1/5/2008 2:13:50 AM)

I sooo want to buy this game, but without a credible AI there is no point. If I wanted to PBEM I could use the game I've owned for 20 years, why would I want to buy it twice? ANd why hotseat when you can all sit around the map plotting your turn? I never liked sitting in a chair someone else made HOT anyway.

Yes EIA was designed to be played with 7 humans, but the UMP rules were put in the game because the designer realized it was never easy to find 7 people with no life all at the same time. Resorting to letting your roomate say "gobble gobble" before he would play his turn at Turkey, or watching Austria never use more then 40 factors against la grande armee do not make for a fun gaming experience. This game has needed a competant AI since its inception.

Fortunatlely, Matrix Games EIA WAS designed to have an AI. If you don't like playing against computer opponents, then you AI haters are the ones who purchased the wrong game.  Otherwise this game is just a play aid.  

I've been following the progress of this game in the forums for a while now, and I was pleasantly surprised when I saw it was released, but with such a questionable AI, I'll be waiting for a patch or 2 before I make my purchase.



helm123456789 -> RE: is the braindead AI going to be solved??? (1/5/2008 4:22:18 AM)

Like a lot of others I have been watching the forums for the game for a while.  I'm familiar with the board game and played a couple of times in my much younger days, but now with a job, three kids and a needy wife and many other responsibilities those days are over.

This brings me to the discussion about the need for an AI.  I have realistically 30-60 minutes at night to do any type of gaming at night and my game of choice is a nice war game that requires thinking with the ability to save at any point <Dad can you check my homework>.  I don't have the time for TCPIP or PBEM to play any games and I'm sure PBEM opponents would kill me for taking weeks to do my turns.  So the only way I'll be trying this game is if the AI is improved on where it is now. 

Well I have killed my PC time while reading this thread and typing the email.  Back to the real world <Kids/Wife>[:@]



Mynok -> RE: is the braindead AI going to be solved??? (1/5/2008 5:29:59 AM)


quote:

ORIGINAL: Fflewddur

I sooo want to buy this game, but without a credible AI there is no point. If I wanted to PBEM I could use the game I've owned for 20 years, why would I want to buy it twice? ANd why hotseat when you can all sit around the map plotting your turn? I never liked sitting in a chair someone else made HOT anyway.

Yes EIA was designed to be played with 7 humans, but the UMP rules were put in the game because the designer realized it was never easy to find 7 people with no life all at the same time. Resorting to letting your roomate say "gobble gobble" before he would play his turn at Turkey, or watching Austria never use more then 40 factors against la grande armee do not make for a fun gaming experience. This game has needed a competant AI since its inception.

Fortunatlely, Matrix Games EIA WAS designed to have an AI. If you don't like playing against computer opponents, then you AI haters are the ones who purchased the wrong game. Otherwise this game is just a play aid.

I've been following the progress of this game in the forums for a while now, and I was pleasantly surprised when I saw it was released, but with such a questionable AI, I'll be waiting for a patch or 2 before I make my purchase.


EiANW was designed to be a pretty faithful re-creation of EiA the boardgame. Since the latter was designed for humans, not an AI, there's no way that "Fortunatlely, Matrix Games EIA WAS designed to have an AI." could be true. They created an AI for EiA, and it is limited at best.

The game would be far better served IMO by improving the multiplayer aspect of the game by including the UMP rules and providing a TCP/IP option before attempting to create a "credible AI", whatever that means.

The time to play issue is very relevant and valid. Making this a game an efficient and useful expediter of multiplayer play would help all of us with limited game time able to enjoy this great game as it was designed to be played.

I won't complain if they satisfy the AI guys at some point. It's just that multiplayer is where this game shines, as anyone who has had the privilege of playing the board game knows. Helping folks with limited time experience that is the better focus right now. Just my perspective on things, and certainly not the only one.



morvwilson -> RE: is the braindead AI going to be solved??? (1/5/2008 10:00:49 AM)

Maybe I am off topic here, but first off, thanks Matrix! I have enjoyed this game greatly so far!

The AI deffinitely needs some tweeking. Here are some of the things I have seen the AI do.

I was playing SP and FR declared war. FR then sent the Art. corp alone with out supply into Spain.
  - Personally whenever I have played FR, Nappy always had at least three corp with him: The Imperial Guard, Artillery(when built) and usually I corp. Further, I never sent them someplace where they would not be supplied. (same with cav corps) Too expensive to replace simply because of attrition.
 
When playing RU, I have had SP declare war and not even a controled minor state from either power was close to eachother.
When playing GB, I have seen PR declare war on GB. (PR did not even have ships!)
While playing FR, I had all of my fleets in Amsterdam. GB would start out blockading them then for some reason drop the blockade enabling me to land troops in England.
Also, while playing FR I have seen the AI march the entire GB army across the chanel into France in the beginning of the grand campaign.

While I am not an experienced programmer, maybe a filter can be set up? (i.e. GB AI gets the idea to declare war on SP while still at war with FR but lacks sufficient shipping to blockade both fleets, filter then discards option. Or RU AI decides to declare war on SW in Jan05 and fleet is still frozen, filter then delays until March)



Grognot -> RE: is the braindead AI going to be solved??? (1/5/2008 1:06:00 PM)

To offer some pseudoconcrete suggestions:

The AI is presently ignoring targets of opportunity, and presents too many such targets.  For instance --

The AI leaves unescorted, empty transport fleets at sea within easy range of enemy fleets for no obvious reason.  A transport fleet that is heading somewhere to pick up troops should be defended, as should one that is carrying troops or providing invasion supply.  A transport fleet that isn't doing one of those things isn't even good bait (because it is captured without forcing the attackers to stop), so it shouldn't be out there.  In the worst case, it should be in a defended port.

The AI should not place a level-A leader other than Napoleon with any inferior leader, which is practically all unless we're talking Mack vs. John.  It's of somewhat dubious value to they place them at all. 

Great Britain should *always* block the channel crossing arrow.  Enemy fleets there must *die* -- it does not suffice to block the crossing with a weak fleet that's going to need to be replaced the next turn.  Blockading enemy fleets and killing any non-blockaded enemy fleets within range of the British Isles should be done -- hell, you have the money to keep the fleets at sea as needed.  Build more ships.  Sweden, Denmark, and Naples are minors of particular concern to GB.  Holland, if lightly defended and the Holland fleets are still around, should be attacked if there are suitable minors to provide the boots.

Russia, if DoWing Sweden, should not march into Abo and cause the war to lapse unless it's firmly aligned with whomever's getting the Swedish corps and fleets.  Which, for an AI, is not entirely all that certain.

Should an AI be about to lose a minor with an intact fleet to an enemy (say, French-controlled Denmark being invaded by the British), doing even an obviously suicidal attack with the minor's fleet is entirely legal, if not entirely realistic.  The same holds for a minor army that's away from the minor that's about to fall -- might as well do something with it before it's lost anyway.

Prussia and Austria should NOT DoW each other, and should back down from DoW'ing each other, at least until France has been beaten down with an unconditional surrender, and unless the one issuing the DoW is already far stronger than the other.  They bleed each other far too much.

AIs at war with each other for a considerable amount of casualties without a surrender should reach an informal peace.   They're sitting ducks for everybody else if they continue.  An AI that's far significantly below its previous peak head count probably shouldn't be DoWing other majors, as well.

It is presently safe to leave empty depots in front of even starving Turkish troops.  As Spain (w/ an entire Ottoman Empire in 1807 or so... oh, and having taken Wales, Ireland and Scotland by then as well, without a single battle against any British corps or fleet -- just the garrison of Gibraltar), I landed 3 Spanish corps in Athens, then took Salonika.  Using Salonika for incoming sea supply, I marched up to Thrace, built a depot, crossed Gallipoll, and besieged Constantinople (building another depot there).

I left the depots in Thrace and Salonika completely undefended.  I didn't bother swatting the feudal corps who were scattered around ignoring the depots.  Minimally, eliiminating my land depots would have forced me to use invasion supply, fed the feudals, and prevented me from reinforcing the siege without actually landing more corps (or marching Ottomans across Anatolia), and would have made it difficult for me to chase inland (away from the invasion supply).  Win-win-win, and in any event sending individual units to hit undefended depots that are part of an enemy's depot chain would be appropriate -- especially for units that would otherwise be foraging, and don't have a battle to go to.

And as I've noted before, the AI appears to assume that any blockading fleet is strong enough to not challenge.  To illustrate, the way I got an unconditional out of GB, as Spain, without a single battle (note:  they and Prussia DoW'd me in Jan 1805 -- a bit dangerous for GB, and a complete waste of 4 PP for Prussia)... almost the entire GB fleet was in Gibraltar (although Nelson was on a single heavy fleet in the Channel), which I kept blockaded with most of my fleet (at first) and then a single fleet with 2 light ships.  I besieged Gibraltar, whose garrison quickly starved -- but deliberately walked out without occupying the city in order to not force the GB fleet to evacuate towards my far weaker 'blockade'.  The blockaded fleet behavior is particularly glaring because fleet strengths are known exactly and there are fewer variables and possible outcomes... and a blockaded fleet has nothing else to do other than attempt to break out.

I launched an amphibious assault with a single corps into undefended Bristol (ignored by the British heavy fleet in the channel, and the two British corps in London).  My single corps garrisoned cities in Wales, Scotland and Ireland, then sat in Edinburgh until the British surrendered to me unconditionally, despite their not being pressed by anybody else, nor my lack of movement towards London.



tevans6220 -> RE: is the braindead AI going to be solved??? (1/5/2008 6:22:33 PM)

There's no argument that EiA the boardgame was designed for multiplayer. Correct me if I'm wrong but aren't almost all boardgames designed to be played that way? Supposedly one of the advantages of computer wargaming was to have a challenging computer opponent to play against. Multiplayer gameplay such as PvP and PBEM are only a bonus. I certainly don't know of any computer wargame that is PBEM or PvP only. I'm sure some exist but they are definately in the minority. Stating that since EiANW was designed to be a faithful re-creation of the boardgame therefore it couldn't possibly be true that it was designed to have an AI seems a little narrowminded to me. It definately was designed to have an AI or else it wouldn't have one. This whole discussion hasn't been about a missing AI but rather an AI that needs work to give players a challenge. Take away the AI completely in this game and you have nothing more than a computer aid to assist in playing the boardgame. Vassal and ADC already exist to do that. One thing to keep in mind is that there have been many computer recreations and adaptations of boardgames. That doesn't mean that they are still intended to be played as the boardgame was. With a good AI you can have just as much fun as you can with any human multiplayer boardgame. Whether or not this game was supposed to have AI is irrelevant. The AI is there. It just needs to be improved.



Obsolete -> RE: is the braindead AI going to be solved??? (1/7/2008 6:00:15 AM)

quote:

Have you ever played any of the Civilization series of games? They all have wicked AIs. They are also very complex games. They pretty much have exactly the same features you list in your post (The utilization of income, manpower, military production, strategy, intuition, and diplomacy).


Actually, the AIs are rather poor, and act foolishly as well.

As for the comment about Atari/Hasbro Axis & Allies game, I don't have anything else to add to that.  It was a junker, which was replaced by a much worse junker, the one what was full of IRON.

I don't understand why a publisher as big as Atari's calibre, doesn't have a person in management who actually tests their own titles before agreeing to anything.  It is said, bigger corporations have larger wastage, and here they really hit bottom.





Page: <<   < prev  1 2 [3]

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
3.578125