ai?

[freeboy]

I seriously don't have time for another great pbem game, although if I did this looks great. My ? regards the quality of the ai, does it get enough help to give competent to good players a challenge?

[Skanvak]

no

[freeboy]

thats what I heard at first too.. ok, perhaps in the future when we get better ai, cheers

[Skanvak]

I would wait until the game can be play TCP/IP or LAN. Really, it is designed for multiplayer. It is like Diplomacy, I don't think you will get anything interesting with the AI. Unless, Marshall implement pure battle scenario without diplomacy (which is feasible).

[Mardonius]

I'd second the comment about the game being about diplomacy but would add that I think that PBEM is just fine. The game will take a long time but is very enjoyable once you can get through the bugs.

[Marshall Ellis]

The AI can (has) and will get much better in the tactical/strategic military realm. I will attempt to make the diplomatic intrigue (Strength of the original game design itself) better as well BUT I don't think it will ever match the anxiety level of an unpredictable human across the table ... IMO.

[pzgndr]

Really, it is designed for multiplayer. It is like Diplomacy, I don't think you will get anything interesting with the AI.

It was originally designed by a 13 year old Harry Rowland as a game for fun, remember, and not something requiring a damn PhD in nuclear physics to figure out. It is not rocket science. At the highest level, each MP is making a few fundamental choices regarding who to ally with and who to go to war with and when, subject to change at certain event points during a game and dependent on the given MP's circumstances. This is complex to a point with 6 other MPs and various alliances to consider, but not impossible. If veteran players expect newbies to be able to play and enjoy this game, then don't try to make it more difficult than it is? And if newbies can be coached (programmed?) to provide "interesting" gameplay, then an AI can be too. This isn't likely to happen soon, but I'm optimistic that eventually it can. [8D]

From some of the stories I've seen here on this forum about real human players making poor diplomacy choices, spoiling games with questionable tactics, and/or quiting mid-game for whatever reasons (and I'm referring to board game players and not pbem glitches or other issues with EiANW), I reckon the bar is set low enough for even a modest AI to prove itself capable of being a halfway decent computer opponent for a PC wargame. Perhaps even better than some average players? That should be interesting enough.

[Skanvak]

The problem is to speak with the AI. Have you ever chat with the AI?
I encourage Marshall to implement the battle scenario.

I think we can design a kind of diplomatic engine for single player games (similar to solitary game) that will mimics the historical situation. But a lone AI in a 6 player game is just an improved UMP, and we should see it this way.

[pzgndr]

I encourage Marshall to implement the battle scenario.

Most EiANW customers of this computer wargame are expecting an AI computer opponent, not a "pure battle scenario without diplomacy" whatever that is. Freeboy certainly wasn't asking about that in his original post? Fortunately Marshall and Matrix Games are committed to more AI improvements and enhancements, as it should be since this is not simply pbem software. The AI is not there yet (still) for competent to good players but should be getting better.

[Marshall Ellis]

Thanks pzgndr!
This is true. We have focused little in ther past on the AI diplomacy and I intend to keep playing with this! Most AI improvement up to now have focused on the military aspect (that still needs improving) which is tons better than it was.

[Skanvak]

The battle scenario which are included in the original EiA to learn the game. They are the Danube campaign and the Russian Campaign. They are good and allow to learn the game.

For the AI I see that in a kind of incremental developpement. You develop an AI good enough for battle and after you deal with the whole game.

BUT the problem with the whole game is to make the diplomatic module interesting, interactive and understandable by the human player. I understand the relation between the major power and the minor power. It is neat and simple, but not refined. The UMP rule is simple too.

BUT what for the diplomatic model between 3 humans (or even one) and 4 computers? Well it becomes the war of the machines as humans don't argue with machines. May be you should look at other area. Developping an AI for diplomatic game is times more difficult that designing and programing EiANW.

I have seen new developpement of the AI by the creator of Balance of power. I think that Marshall (and us) can get idea from its story telling engine as it allow for interaction with the AI. Or put him on the project :p as it has designed an AI for a negociation game (ok one human and one AI).

[pzgndr]

They are the Danube campaign and the Russian Campaign. They are good and allow to learn the game.

I understand now. Yes, these would be good. I believe Marshall addressed this a long while ago and indicated difficulty in making these happen with parts of the map and and some MPs not in play, and how this would affect turn sequencing and such. It *should* be possible to toggle areas and MPs on/off in the editor and have the game simply skip over those not in play but that's another can of worms to deal with. Marshall would have to confirm. The priority should be to implement the classic map and grand campaign with original EiA OOB, and other high demand items, before tackling something like this?

For the AI I see that in a kind of incremental developpement. You develop an AI good enough for battle and after you deal with the whole game.

Here I disagree. The programming structure should address the whole game up front with grand strategy diplomacy, campaign strategy, operational maneuver and tactical combat levels. If the tutorial scenarios are ever implemented then they could help fine tune the AI's lower levels, but we shouldn't wait for them and hold up AI development in the meantime. PBEM issues and bug fixes have been a high priority for quite a while, and with those nearing resolution with sim dip/eco in v1.08 and possible file exchange streamlining for v1.09, it's time to pump some life into the computer opponent.

[NeverMan]

ORIGINAL: pzgndr

I encourage Marshall to implement the battle scenario.

Most EiANW customers of this computer wargame are expecting an AI computer opponent,

Is this true? What are the numbers exactly, can you give us a break down of people who bought the game for the purpose of playing:

1. single player with AI
2. PBEM with AI
3. pure PBEM no AI


Since you seem to have the numbers I would be VERY interested in seeing this breakdown. Thanks alot pzn.

[pzgndr]

No I do not have "exact" numbers.  But one website I found with a listing of worldwide EiA pbem players had about 300 names.  The number of active pbem players for this game on this website seems to be an order of magnitude less.  I would estimate the number of EiANW customers who bought this Matrix computer wargame version expecting it to have an AI to be an order of magnitude more.  So, 30-300-ish pbemers versus an estimated 3,000+.  I would ask you if YOU had any numbers worth citing, but it doesn't much matter.  Marshall and Matrix Games have repeatedly indicated that AI improvements and enhancements will proceed forward, as it should be, irregardless of the spurious heckling to the contrary.  So, pound sand.  And have a Merry Christmas. [;)]

[NeverMan]

ORIGINAL: pzgndr

No I do not have "exact" numbers.  But one website I found with a listing of worldwide EiA pbem players had about 300 names.  The number of active pbem players for this game on this website seems to be an order of magnitude less.  I would estimate the number of EiANW customers who bought this Matrix computer wargame version expecting it to have an AI to be an order of magnitude more.  So, 30-300-ish pbemers versus an estimated 3,000+.  I would ask you if YOU had any numbers worth citing, but it doesn't much matter.  Marshall and Matrix Games have repeatedly indicated that AI improvements and enhancements will proceed forward, as it should be, irregardless of the spurious heckling to the contrary.  So, pound sand.  And have a Merry Christmas. [;)]

Ok, I just wanted to make sure that your statement was pure guessing and speculation with no real evidence. Thanks.

The only numbers I have worth "citing" (as you say) are just from the people who post here, who are mostly into PBEM games.

[Skanvak]

Pzn,

Even if I could agree with your point on the AI, it does not address the problem of speaking with the AI... And I think it is a task too big for Marshall. So even if I agree on the point to do before battle scenario, AI for me is just a point on which developpement shoudl be stopped or given to a dedicated team of crack developper.

Pzn, THERE are no AI able to do diplomacy in ALL THE GAME INDUSTRY! You really think a miracle will happen?

[pzgndr]

AI for me is just a point on which developpement shoudl be stopped

Sure, whatever. Good luck with that.

[NeverMan]

ORIGINAL: Skanvak


Pzn, THERE are no AI able to do diplomacy in ALL THE GAME INDUSTRY! You really think a miracle will happen?

Yes, he thinks that one software developer with his own startup who makes wargames is going to be able to accomplish what hundreds of VERY TALENTED software developers and researchers have yet to do.

So, yes, he thinks a miracle will happen.

[Marshall Ellis]

I believe! LOL!

[pzgndr]

Yeah, I believe a better AI can and should be developed.  Setting an unrealistically high expectation for a master-level genius AI to challenge experienced veteran players is not the issue here except for Skanvak's strawman argument being presented, and that's no reason to stop AI development completely and not make any improvements or enhancements that other players expect.  This is a silly argument, and presumes to dictate to other players how they should play and enjoy the game as they see fit.  For players like freeboy or other small groups of 2-3 players looking for a fun fast game against relatively decent computer opponents that at least don't suck too badly, I guess they'd be out of luck if Skanvak and Neverman were running the show?? [:-] 
 
Fortunately Matrix Games realizes there is a broader customer base of computer wargamers looking for occassional fun gameplay than just the handful of vocal pbem veterans here.  I've met and chatted with David Heath several times over the years and he recognizes the market.  If anyone has some data for numbers of newbies flocking to freebie CyberBoard and Vassal pbem EiA games to jump into 7-player games, let us know?  I maintain (unofficially) that many more customers bought the EiANW computer wargame version for the convenience of a computer opponent rather than simply for pbem which is already freely available elsewhere.  Freeboy is just one example of a Matrix Games customer looking to play a decent game against a computer opponent for fun and entertainment.  I'm guessing (unofficially) there are lots more like him lurking and waiting patiently (or not so patiently as the case may be).  [;)]