July Update

[Marshall Ellis]

Hey guys / gals:

Just wanted to drop a quick note and let you know where we are and what we're doing!

First of all, WE'RE testing! Wooo hooo! Second, it's early so there are a lot of issues but that means that we have a great test crew.

I would like to thank each of you guys in the test group and want you all to know that I'll do my best to keep up!

We're not testing the AI or PBEM yet since these are not quite stable yet but as they test and as I fix, I'm also finishing up the AI and PBEM.

It's always amazing to me how obvious some of the bugs that are found are and even a bit amusing how I miss them but I ALWAYS get a kick out of testing so to me this is FUN!

Anyway, I would like to ask for your input as to some of the best possible initial setups for each nation in the grand campaign for 1805. Whatever format you wish and to whatever detail you wish but I'm looking for some clever setups and reasons as to why they are clever to arm the AI.

Thank you all

[Wynter]

IMO the AI setup depends on the strategy it is going to use.
F.ex. let's look at Turkey (I like the Turks), and different (working) initial strategies for the Turk.
Turkey could be starting out by trying to get the Ottoman Empire, so obviously he will deploy forces in Palestine to invade Egypt and forces on a fleet ready to strike Tunisia.
Or Turkey could be starting out by invading Russia at the start of the game because Russia starts out weak.
Or Turkey could decide to take some Italian minors.
Or Turkey could ally early with France and invade Austria.

I think it is best that first the possible strategies for each nation are defined, and then that the nation sets up his forces according to a choosen strategy.

My 2 cents...

Jeroen.

edit: typo

[Supervisor]

I agree. The setup depends entirely on the country and strategy of the player, pre-set alliances, setup close to your ally for early stacking (Au + Pr usually), How aggressive of an early stand the country wants to take. What kind of attitude it wants to project toward other nations. Russia's navy is safe setup in a cold weather port to start (St. Pete's). When I play France, I always set my fleets an semi-even fleet counters (France has 4 so I use them all) and put them in the highest gun ports, this occupies a good amount of Britain's initial navy (blockading them). As Turkey I try and stay nuetral at the beginning, so I can go after Africa. Setup some corps very close to Egypt with a depot supply. As Russia, I setup Close to Finland so I can make a few corps push thru Finland into Sweden, this happens slowly but surely, and usually GB is too occupied with going after Denmark/Norway. As prussia I try to make an early alliance with Au and Ru and make a strong corp setup close to Hanover and other bordering minor countries, although I try not to antagonize France too much. I do the same with Au, but also make a strong push for parts of Italy, with less emphasis than Pr on antagonizing France. As France, in the beginning, I try to look aggressive but due to the amount of setup factors I don't want to seem too greedy as I don't want to bring on a coalition war until I am 100% ready to do so.

So in short, setup depends on pre-game alliances, diplomacy, and strategies. Are you planning on try to work the VP system, or are you going balls to the wall and trying to grab as many PP (thus VP) as you can, anywhere you can? What countries do you want to have a strong alliance with, a weak alliance with, no alliance with (at the start). The diplomacy in EiA is very different from the diplomacy in a game of "Diplomacy".

If I can think of more stuff later I will post again.

[Marshall Ellis]

Wynter:

I see what you're saying but I'm trying something almost backwards to what you're saying.

What if you were to setup the forces on the Egyptian border which then the computer would look and say "Wow, maybe I should invade Egypt since I have all of those forces on the border"

This is an experiment but if this works then it could allow an almost infinite number of strategies to be used not with extra coding but with different setups.

Does that make sense?

[Supervisor]

Originally posted by Marshall Ellis
Wynter:

I see what you're saying but I'm trying something almost backwards to what you're saying.

What if you were to setup the forces on the Egyptian border which then the computer would look and say "Wow, maybe I should invade Egypt since I have all of those forces on the border"

This is an experiment but if this works then it could allow an almost infinite number of strategies to be used not with extra coding but with different setups.

Does that make sense?

I think this could make sense providing there were multiple setups depending on pre-game situations, or early game decisions. However, I still think that setup should always follow strategy.

[Wynter]

Originally posted by Marshall Ellis
Does that make sense?

Yes it does. And it certainly is a good idea to let the AI make intelligent decisions based on military positioning. I will try to provide you with some different setups.

I wonder, though, what will the AI decide, in our Turkish example, after the conquest of Egypt? It has all this troops in Egypt, so it will need a some sort of 'master plan' to do something usefull with his troops there, no? That is why I was thinking about general strategies (the Empires in Harm alternative victory conditions are a good starting point for general strategies)

Jeroen.

[Pippin]

Marshall Ellis,

One thing I have seen time and time again, is an AI while even if made decently, will often be exploited one way or another. In some cases you could swear there is no algorithms at all except a random movement.
A great solution I’ve seen is with developers who make a scripted AI. This (in theory) not only takes pressure off the coding team, but also allows a much more versatile / patchable design even far after the game has been out of production.
In every case, I have seen clubs improve & re-code the scripts over time making them more and more robust. Certainly better than the first release from the developers. This is not a knock on you of course, just that it is understandable that when you have many people playing constantly, and having unlimited time to tweak the scripts, it results in a more stronger AI than one person during development can possibly achieve.
The only trick here, is to make the script versatile enough.

[Marshall Ellis]

Sorry for the new thread earlier (MY bad)

No knock on me was taken!

You're so right. One person develops band width issues. The same guy tuning the AI cannot be same guy who makes sure that menu buttons refresh properly for a new game

Keep in mind that our AI is scripted but only on a strategic level. A grand strategy is picked then the computer will adjust tactically to further her grand strategy. This is where I was attempting to experiment with different setups prompting the computer to do certain tactical strategies.

To answer your question about post Egypt, well the computer would look and see that Cyranecia is adjacent to a conquered nation ans possibly target it for invasion BUT forces could not enter on the first turn which SHOULD trigger a transport to be moved down to a port which would then trigger the computer possibly declare war since she should be able to land foces in the same turn (preventing lapse).

Remember much of this is theory since my lab experiments to date have created some .... ???? "Interesting" results LOL i.e. Turkey attempting to invade France, you get my drift ...

[ABP]

Hi,

It makes sence to me to use setup as a tactical trigger in the AI.
As mentioned it should be done with care. The multiple AI setup depending on Grand Strategy sound good.
However make sure you don't create a snowball effect. With the case of Post Eqypt. The turk has this nice army standing around in Egypt. Fine Cyrenaica next. The turk has this nice army standing around in Cyrenaica. Fine Tripolitania next, etc. etc. Suddenly most of the Turks are standing in Morroco and the lines are overextended. There should be a check in the AI to gauge when to expand and when to consolidate.
To be honest I think it is great if you can pull a good AI off, but I expect mostly to engage in 7 human player campaigns.

[Marshall Ellis]

ABP:

Me too! I'll be playing with humans more than anything as well. I do hope that the AI will be challenging even though I realize most people will be using it to fill a void.

It will be tough since the computer doesn't know the fine art of fancy diplomacy too well so I'm gonna have to improvise and do some serious testing.


Thanks

[Chiteng]

Originally posted by Marshall Ellis
ABP:

Me too! I'll be playing with humans more than anything as well. I do hope that the AI will be challenging even though I realize most people will be using it to fill a void.

It will be tough since the computer doesn't know the fine art of fancy diplomacy too well so I'm gonna have to improvise and do some serious testing.


Thanks

Just make a heuristic table =)
One for each major power. In fact the Air Force rules
'national aspirations' is a good start.

[Supervisor]

Well, some of us aren't so lucky as to have 7 people laying around, and as far as PBEM games go, if I had 7 peeps via PBEM, why would I need to pay 60 bucks for a computer game? The same goes for playing "in person". If I played just those two options and had all 7 players, I wouldn't spend the 60 bucks, I would just use cyberboard instead and use the actual EiA rules. I am looking forward to the comp having at least some decent knowledge of how to act/respond. If it takes another couple of years to do this, hey, whatever. I'm not going anywhere.

Didn't mean this post to seem rude or aggresive. Just stating my opinion.

[pfnognoff]

Setup must be in accordance with the initial diplomacy.

Like Ottoman options Wynter posted, they depend on diplomacy. You will not transport Ottoman army on ships to Tunisia if you don't run the idea by the British player. You go against Austria if France is an ally who pays, etc.

Maybe you should have all the different setup options stored and then pick one after the diplomacy phase according to your (AI's) alliances or agreaments.

2c

[StCyr]

Dear Marshall Ellis,
there may be gamers playing Turkey who would declare war on Russia simply because there is a fine chance to do so. But if you think about AI, please recall that a nation acts in long terms. Remember Clausewitz "War is the continuation of state policy by other means". I think you should give a Metternich & Tallyerand AI to the game, and not a Patton AI. Or to be more precise, let the diplomatic AI decide about when and if to go to war, and the military AI about how to fight the war. Take Napoleon as an examaple: A real military genius, but he never forgot about the grand design.
Please, you simply have to reflect the fact that Diplomacy is the key in relation between nations. And indeed EIA shows this perfectly as long as every player does care for their nations more than to bring down a special player no matter what it costs to himself.
In EIA, you place your troops in accordance with prior diplomatic thinking. And it takes a poor AI to brake any former agreement just because the computers ally in spe did make a setup that fits to beat a coming enemy of both and to backstab this ally now.
And please recall the politics in general of this time:
It was impossible (and is reflected in the rules) simply to ambush another nation without a declaration of war. It was the time of the Cabinet Diplomacy. The diplomats acted as men of honour, French was the only accepted languae (even when Nappy invaded Russia, the russian generals spoke French in the Headquarter), there was no "poor cheating".
Of course I dont ask you to programm this game in french , but can you think about an option to reflect "noble Cabinett Diplomacy" ? To brake an alliance (esp. combined with a declaration of war) should not happen very often. In fact, I like the optional rule that it is impossible to declare war and to break an alliance the same month very much, cause such an action was impossible by that time (if it is possible in our days as at all...).

[Le Tondu]

Marshall,

Maybe it is just semantics, but shouldn't the "Grand Campaign" run the full breadth of the game?

I ask this because I was under the impression that there was a pre-1805 starting point like 1796.

Please correct me if I am wrong or did you just mean to say "the 1805 Campaign" instead?

Thanks.

[Supervisor]

The 1976 starting Campaign is part of a supplementary game called the Revolutionary Campaign Game (produced by EiA lovers). The "Grand Campaign Game" in the original EiA game goes from 1805-1815.

[Reknoy]

Originally posted by ryta1203
Well, some of us aren't so lucky as to have 7 people laying around, and as far as PBEM games go, if I had 7 peeps via PBEM, why would I need to pay 60 bucks for a computer game? The same goes for playing "in person". If I played just those two options and had all 7 players, I wouldn't spend the 60 bucks, I would just use cyberboard instead and use the actual EiA rules. I am looking forward to the comp having at least some decent knowledge of how to act/respond. If it takes another couple of years to do this, hey, whatever. I'm not going anywhere.

Didn't mean this post to seem rude or aggresive. Just stating my opinion.

I agree 100%.

In fact, I would go so far as to say that I think the largest consumer base that this game will appeal to will be the lone EiA gamers.

I recall reading those 'lonely heart's' letters from solo players forced to use the (well constructed) solo rules by Mr. Glam.

And then you have the PBEM game where someone bails mid-stream and leaves the game in a lurch.

*Picture John (Prussian), having just routed the Russian army in a great land battle, finds out the Russian quit and the rest of the players don't want to continue*

"Geez!!!! Just when I was doing so well!!!!!!!!!!"



While I maintain contact with possible players, I expect I will be purchasing this for the fact that I want to play when I want to (and have time to!).

Having said that, I think there's no cause for alarm.

My ideas of what will make a good AI may not mesh with everyone, however.

I don't see the AI powers as requiring ANY historical "accuracy".

The game (at least the last time I checked) was based on victory points.

And DOWing an ally costs 5 PPs.

That will have long term consequences. Even so much so that you might lose the game if there is not a quick way to recoup.

Point is, if the AI is mindful of this (and perhaps there is a way to reflect the AI's reluctance to work with a player that repeatedly breaks alliances) then I think that's the main goal.

I mean to say, certain strategies for certain players make more sense (Spain might be in one war and then manipulate like mad; while France has no other choice but to fight every 12 months or so - actually not if they get into a nice box and can manipulate there - and that's the point).

Ramble ramble ramble.........

[John Umber]

Grand strategy

I like the concept that each country is given an overall strategy.
The options given i the Airforce Rules are good way of showing the which areas where important to which country (as well as the optional dominant power rule).

But if I am playing a game versus an AI, I wish it to be historical correct. I would be a little disgusted with a Austrian major power around the Mediterrean.

To make my point. Each country should have several strategies activated at once. The highest and most important should be a political one considering the values of the countries AND the status of the victory points. Victory points considering the event of an overall winner, not just the leading country, it is a very long game. This could spring some interresting events like switching sides. A not completely strange thing those days. For a Swede this was a important time in our history, we lost Finland because of this. Russia decided to go with the French against England-Sweden. This overall strategy should point the country in a direction of dominant status (all political leaders dreamed of this).

Second strategy should be a more short term variant. Considering what to gain with a alliance just now. What to gain with that ally or that. Consider Turkey, to attack North Africa means moving several troops to the Egypt area. What country will protect Egypt? Likely England or Spain? England might also try to send troops or at least cutting of naval supply. Even France should be interrested (historical any way). So to create the Ottoman Empire (a major goal for Turkey) means several short term conditions. The border towards Austria-Russia must be protected by troops or a political protection. An alliance with France should cause the Austrians to prepare a defence which binds several Turkish corps along that border. A condition Turkey would like to avoid. A short term strategical condition would then be ally with Austria, Russia or England to be ABLE to initiate the attack on Egypt. Also what a defeat could cause...

Third strategical level is considering the movement of the troops. If Turkey has a substancial amount of troops in the region and the defence is weaker, then attack. It must consider the options of a defeat in Egypt and possible retreats, supply and etc.

Must be quite a test programming this AI! Good luck to you.

Historical note:
Napoleon cheated in cards against his generals, not to steal money (he paid it back), but rather he hated to lose.

[John Umber]

Another item

The AI should also take consideration of earlier events. The AI should remeber what steps the countries leaders took before. A country that breaks an alliance is rather untrustworth? I remember people to be rather touchy about these kind of things. It can make it quite difficult later on. Also the computer should remeber if a "preliminary" unconditional changes at the table. If you surrender, the winner states the rules...

[John Umber]

Another thing

I am rather new to this computer EIA, but is there supposed to be a "Hotseat" version? With more than one human player? I would like this very much indeed, not always PBEM...