Bug?

[joviel]

I'm not sure if this is a bug or a RTFM situation, but I R the FM and couldn't find anything. [:D]

In a multiplayer game as Austria I declared on Bavaria turn 1. Managing player (France) plops the Bavarian army and a depot in the Munich area. I march a corp in and am asked if I want to stop movement and eat the depot. This corp does. I follow with two more corp. Come land combat there is no land combat. The hungry Bavarian army stares across the lines at the belching and sated Austrian army which is apparently in calorie shock. [>:]

Why wasn't combat offered? I thought it was manditory when two field armies meet?

Thanks,

Scott Uhrick

[Grognot]

Hmmm.  Is the Bavarian army inside the city?

[Marshall Ellis]

Yea, can I see a screen print of this?
 

[Jimmer]

I'm with Grognot: The Bavarian corps must be in the city.
 
Although, if so, this would be a huge step up for the AI. That big corps is a potential political point factory for the controller. As long as it is alive, unless the Austrians bring their whole army (pretty hard to do in Winter), it stands  to gain 1-3 political points for the controller (as opposed to about a 90% chance of losing one by staying in the field).

[joviel]

Yep, you're right, the Bavarians ran into the city.

Sorry for the unnecessary question,

Thanks,

Scott

[Jimmer]

ORIGINAL: joviel

Yep, you're right, the Bavarians ran into the city.

Sorry for the unnecessary question,

Thanks,

Scott

There are no bad questions except those which are not asked.

Further, this particular one led to the follow-on question in my previous post: Could it be that the AI wised up? Most of the time, the AI would fight the field combat, losing horribly a great percentage of the time. This would lose 1 PP for him and give one to the Austrians.

However, going into the city, especially in Winter, virtually guarantees a positive PP result, while not granting more than one to the Austrians.

[Marshall Ellis]

Jimmer:
 
Would love to take credit for a wisened-up but I think it was mere "stupid" luck Your explanation seems very sound as to what the AI should have done. Would it be prudent for most minors to "retire" to the city as opposed to fighting a field combat? Maybe have the harder AI settings do this?

[JavaJoe]

ORIGINAL: Marshall Ellis

Jimmer:

Would love to take credit for a wisened-up but I think it was mere "stupid" luck Your explanation seems very sound as to what the AI should have done. Would it be prudent for most minors to "retire" to the city as opposed to fighting a field combat? Maybe have the harder AI settings do this?

Can the system count the corps that are attacking it and then decide to retire or not depending on the force that is attacking it?

Would hate to see minors running into the city against 1 corps that may or may not be full.

Heck for hard AI let it cheat and know what's attacking it and compute the odds of winning the battle [;)]

[Marshall Ellis]

Joe:
 
You know it could be coded to do anything
I like this idea in the area of AI improvements and the hard AI cheating a bit. Maybe not "cheating" but making a more "informed guess" as to the force composition
 
 

[JavaJoe]

ORIGINAL: Marshall Ellis

Joe:

You know it could be coded to do anything
I like this idea in the area of AI improvements and the hard AI cheating a bit. Maybe not "cheating" but making a more "informed guess" as to the force composition

As long as you don't allow the AI to know the chit choice ahead of time![:D]

[Marshall Ellis]

Joe:
 
Yes, that would be "cheating" for sure!
 

[Jimmer]

ORIGINAL: Marshall Ellis

Jimmer:

Would love to take credit for a wisened-up but I think it was mere "stupid" luck Your explanation seems very sound as to what the AI should have done. Would it be prudent for most minors to "retire" to the city as opposed to fighting a field combat? Maybe have the harder AI settings do this?

Playing a game, most minors should go into the city (IF they fit). However, in real life, some minors were just too cantankerous to do that (North African states, for example, against anybody except the Turks).

So, what you might consider is giving a probability of sitting in the city vs. fighting.

This has the added advantage of having the players not "know" in advance that they're going to be attacking just a city. For example, in Egypt, the Turks in our current game have spent six whole months trying to get Cairo to give up the ghost. They finally succeeded in June. I think I (as GB, the controller) got 3 PP out of the deal.

However, Egypt, with their monster corps, could just as easily have decided to fight it out. If the get lucky with a battle on the border, the Turks might wind up having a lapsed war. (Although, as GB in my game, I did NOT want that to happen -- I prefer not being at war with the Turks ).

So, make it random, but have each minor (with corps) have their own percentage for city vs. field. Also, this might change against certain powers (perhaps add in the national modifier?).

[Jimmer]

ORIGINAL: Marshall Ellis

Joe:

You know it could be coded to do anything
I like this idea in the area of AI improvements and the hard AI cheating a bit. Maybe not "cheating" but making a more "informed guess" as to the force composition

I like "informed guess" (for the AI). That's what humans do all the time when making this decision.

One thing that really needs to be added, though: The ability to not have your corps' current status known. If one is in the city, that can be seen before one's opponent moves. So, opponent knows it will be just a city combat, and can react accordingly.

I recommend instead that corps always appear to be in the field (unless besieged, obviously), or that this be an "orders" function (retreat to city vs. stay in the field).

[Jimmer]

ORIGINAL: JavaJoe

As long as you don't allow the AI to know the chit choice ahead of time![:D]

Oh, I don't know. Civilization does this all the time on "Deity" level games.

Perhaps as a future improvement, a "deity" level AI would be appropriate for EiA: The AI knows all corps strengths at all times. Once the other shortcomings of the AI are corrected for, this might actually prove useful to the AI.

[zaquex]

Giving up PP to the opposition is bad, unless supported most minors have very little chance to hold out vs a major and should probably fight from there capitol. There are some exceptions, Sweden, Denmark and Egypt have a small chance defending against most majors, how small depends of course on how much resources the major comits and to some extent on what major is attacking.
 
Some things to remember thoug. A Corp fights for PP no matter where it is so just marching the corp in to the city doesnt deny the attacker the points. Citys with Flechettes (fortified) also gives PP regardless if there is corps in the city or not ( I think a corp safely can be used in a fortified city without giving away additional PP, not that there is many fortified capitals).
 
I think the AI should take care not to lose PP over unsupported minors unless they are very high on the status table - its silly for a AI controlled major to practically sink itself to unstability over minors they are not prepared to fight for. The war will come soner or later and then they will need those PP.
 
The AI should be given a rough idea about the strenght of enemy corps, maybe 1-4 (weak), 5-8 (understrength), 9-14 (normal), 15-19(overstrenght) and 20+(huge). The AI needs some input to base its decisions on, or it will be blind and incompetent. The AI can only act on the data its given and according to the logic its been given - its not like it has the same perceptive and lateral thinking abilities as a human and and neither does it normally learn by experience like a real player.
 
 

[Soapy Frog]

The corps should have the option of retiring into the city at the combat phase; This would make it much easier to program AI behaviour.
 
For example at the point of decision, the AI controlled corps could take an educated guess at the size and morale of the force opposing it and then either fight in the field or retreat inside the city based on it's perceived chances of success.
 
Unfortunately at this time the rules are not correctly implemented so this is not possible. BTW this makes it equally hard for players to make rational decisions regarding whether to keep the corps in the field or put it in the city.

[Jimmer]

ORIGINAL: Soapy Frog

The corps should have the option of retiring into the city at the combat phase; This would make it much easier to program AI behaviour.

For example at the point of decision, the AI controlled corps could take an educated guess at the size and morale of the force opposing it and then either fight in the field or retreat inside the city based on it's perceived chances of success.

Unfortunately at this time the rules are not correctly implemented so this is not possible. BTW this makes it equally hard for players to make rational decisions regarding whether to keep the corps in the field or put it in the city.

I agree completely, except for the "not implemented correctly". What you mean is "not the way the board game does it", which isn't quite the same. I think you are correct in your assessment, but it would require a rule change as well as a code change to modify it.