Home Trending Latest threads New posts Developer diaries

AI choosing events

Jump to latest Follow Reply
  • We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
unmerged(2036)

unmerged(2036)

Lazy poster - avid reader
Mar 20, 2001
220
0
Visit site
It been awhile since I read the FAQ etc. (progress seems somewhat slow these days :rolleyes: ), so i migth have forgotten some of the finer details of the game - but here goes:

From what I recollect there is gonna be an event system similar to that of EU2 (hooray for that), however the events are gonna be less specific in their nature.

Still rather than adopt the 85%/15% -thingy from EU2, when "chosing" between choices, I would imagine that AI choice could be based on some AI analysis of it's current situation.

The extra memory/cpu needed would only be proportional to the number of events, which (I imagine) wouldn't be significant to decision making that the AI, needs to do anyway, during the game.

What do you think ?
-or is the CK event-system gonna be so different from EU's that the above is out of context?
 
From what I've seen mentioned by Sergei and Boreal the question might be out of context. I believe the event system will be different from EU (and HoI for that sake), but not how much different...

What I would like to see is a more flexible solution where you can put several different sets of conditions into the triggers, and different effects based on what trigger set fires the event...
Code: 
[color=silver]if(condition set 1) {
   <do one thing>
   }
elseif(condition set 2) { 
   <do something else>
   }
else { //if nothing of the above
   <do some other stuff>
   }[/color]
 
Originally posted by mrmister
[B...............
Still rather than adopt the 85%/15% -thingy from EU2, when "chosing" between choices, I would imagine that AI choice could be based on some AI analysis of it's current situation.

...................[/B]
Click to expand...

That sounds like a good idea. In some cases it might be possible, but others it might need too much information (like what the AI's future plans might be).

On the other hand, these events might be set so as to direct the course of game history in a certain direction. Additionally, monarchs did not always make the best decisions for theircountries/dynasties. A lot of decisions were based on personal preference/whim/ego/etc

So having the AI chose event options in a 85% - 15% type of randomness seems reasonable to me..:)
 
Re: Re: AI choosing events

Originally posted by Sonny
......
On the other hand, these events might be set so as to direct the course of game history in a certain direction. Additionally, monarchs did not always make the best decisions for theircountries/dynasties. A lot of decisions were based on personal preference/whim/ego/etc

So having the AI chose event options in a 85% - 15% type of randomness seems reasonable to me..:)
Click to expand...

Unless the CK AI is a huge step upward from EU/HOI (which I hope, but less will do :) ), the AI is gonna need all the help it can get, short of cheating. So there is really no reason to let it take bad event choices with complete randomness. :rolleyes:
To defend my position I am just gonna present a short a EU2-inspired example :)

Say there is an event where you get the option to construct some improving structure at the cost of hard currency.
Choosing not to build the structure will just result in some minor penalty (VP, stab, what ever....)
So 85% of the time the AI will choose to build the structure,
while what it (IMHO) should do is look at the following:
To the base chance of chosing option A (85%) add/subtract if:

1)"Do I already own said structure?" (-50% chance to build)
2)"How is my economical situation?" (chance to build: (avaiable_ducats/cost_of_building)-1 * 10% )
3)"Am I at war?" (-10% chance to build)
etc.

Altering the propabilities of AI chosing events in such (a simple)
maner should be doable (although I concur that for some events it is harder to determine what should affect the desicions).

Of course if Harvard is right this had little relevance :)
(Which means I will have to move this to the HUGE EU3 proposal thread :cool: )
 
Re: Re: Re: AI choosing events

Originally posted by mrmister
Unless the CK AI is a huge step upward from EU/HOI (which I hope, but less will do :) ), the AI is gonna need all the help it can get, short of cheating. So there is really no reason to let it take bad event choices with complete randomness. :rolleyes:
To defend my position I am just gonna present a short a EU2-inspired example :)

Say there is an event where you get the option to construct some improving structure at the cost of hard currency.
Choosing not to build the structure will just result in some minor penalty (VP, stab, what ever....)
So 85% of the time the AI will choose to build the structure,
while what it (IMHO) should do is look at the following:
To the base chance of chosing option A (85%) add/subtract if:

1)"Do I already own said structure?" (-50% chance to build)
2)"How is my economical situation?" (chance to build: (avaiable_ducats/cost_of_building)-1 * 10% )
3)"Am I at war?" (-10% chance to build)
etc.

Altering the propabilities of AI chosing events in such (a simple)
maner should be doable (although I concur that for some events it is harder to determine what should affect the desicions).

Of course if Harvard is right this had little relevance :)
(Which means I will have to move this to the HUGE EU3 proposal thread :cool: )
Click to expand...

Yes, you are right. In EU2, it was sometimes a problem if the AI had to pay a lot of money, which it didn't have, for a project, which it didn't need, just because it was the most historicla answer. Anyways, I don't know how the economy system works in CK, but my medieval knowledge tells me, that hard currency was not so widespread at that time. Building something great would rather mean enserfing more people or taking serfs from one vassals fief to this new assignemt, which would make either the peasents or the nobles angry at you. Building the Chinese great wall or the Egyptian pyramids costed many lives of many unlucky serfs. Although these are examples from ancient times, I think it kinda also works in the medieval ages.
 
Re: Re: Re: AI choosing events

Originally posted by mrmister
.............................
Choosing not to build the structure will just result in some minor penalty (VP, stab, what ever....)
So 85% of the time the AI will choose to build the structure,
........................
Click to expand...

And what if that stab hit causes it to go to -3 and revolts pop out all over the place?

See, there are just too many situations where making a logical decision is difficult. Not only that but (at least in EU) you know what the result will be years down the road (why conquer Hungary when you are going to inherit it anyway - syndrome). The AI does not - and should not - know these things. So what you classify as a good decision may be based on knowledge you should not posses at the time the decision is made.

Maybe the % should not be 85-15 but the idea is the same. And as I stated already, monarchs did not always chose the best path even when some of the results were knowable.:)
 
Would it not be a better idea if the AI never had random events; if EU II is any help, it never handles them well and this further handicaps the AI.
 
Re: Re: Re: Re: Re: AI choosing events

Originally posted by mrmister
So it is better to "build the useless piece of c***", and go bankrupt a short time later ? :D
Click to expand...

and all the provinces which revolt defect to another country...:D
 
You must log in or register to reply here.
Top Bottom
Back