[BTS] Development thread

[Newcomer24]

Just posting to say that I have played several games so far with 0.90 and it defenitely needs more testing and balancing.

First of all... I have seen a lot of stupid wars, the AI apparently tends to start several wars togeter.... which is caused probably by the new play to win system.
Now... if an AI has a big military advantage and/or is already winning a war it's ok to declare war on another small country. But I have seen a lot of AIs starting a second (or even third) war even if it's not able to win the first one.
I have seen AIs having hard times winning a war with a small nation starting all of a sudden a second war vs the strongest nation around (suicide strategy?).
The AI is forced to spread its armies causing the inability to do a proper offensive... most wars are often ended with status quo.
The AI is not good at managing multiple fronts so waging several wars togeter should be done only when the military advantage is huge.

Another thing to report is the stupid usage the AI sometimes does with Siege weapons (this is probably not 0.9 related though). It happened more than one time to see undefended Siege weapons invading enemy territory with the main stack being several turns behind. I don't remember this happening that often in the past.

Last thing for now, did you by any chance touch the city founding logic in one of the latest releases? I'm asking because it looks to have worsened... the AI often manages to find the worst spot in an area to found a city... but I guess it was just bad luck in my last games, still worth asking.

Keep up the good work, looking forward to newer builds.

 

[_alphaBeta_]

First of all... I have seen a lot of stupid wars, the AI apparently tends to start several wars togeter.... which is caused probably by the new play to win system.
Now... if an AI has a big military advantage and/or is already winning a war it's ok to declare war on another small country. But I have seen a lot of AIs starting a second (or even third) war even if it's not able to win the first one.
I have seen AIs having hard times winning a war with a small nation starting all of a sudden a second war vs the strongest nation around (suicide strategy?).
The AI is forced to spread its armies causing the inability to do a proper offensive... most wars are often ended with status quo.
The AI is not good at managing multiple fronts so waging several wars togeter should be done only when the military advantage is huge.

I've also seen the same. Ragnar was waging a successful war against Pacal II (his neighbor to the east) and he decides to declare war on me (I'm to the west) even though he was at least "pleased" with me if not "friendly." This was an impressive surprise given the excellent relations we had. It went nowhere, however, given that I was the second most powerful civ right behind him. I took a city off him and we settled for peace soon after. Ragnar was now fighting wars on both sides of him, wars that he started.

I don't see how this made sense to declare on the second most powerful civ whom you have good relations when there were better targets to be had. Less powerful civs were to the north.

Another thing to report is the stupid usage the AI sometimes does with Siege weapons (this is probably not 0.9 related though). It happened more than one time to see undefended Siege weapons invading enemy territory with the main stack being several turns behind. I don't remember this happening that often in the past.

Have seen this as well. In the opening stages of the war I spoke about, Ragnar decided to send in a stack of four catapults who sadly died at the hands of my marines. I also saw lots of catapults moving around unescorted one tile in from the border. I could have easily picked them off before they turned into cannons.

 

[entwood]

I posted this first in Better BUG AI and was referred back here:

Can the player still 'freeze' the AI? What I mean by that, and for example, I was playing a new game, destroyed Ghengis Khan's second newly built city and then I hung around entrenched in the borders of his remaining city, stayed in a state of war, and he just sits there, never trying to get out of trouble. You could actually forget you are at War because of the inactivity. In another case, another leader built approx. 30 archers in a city and never tried to get out of trouble with them. A leader could do a lot with a SOD of 25 archers leaving some behind to protect the city.

I think this freeze problem exists in many situations. Is this the best the AI can do when they are hit hard but, seemingly to me, could find some ways to get out of trouble?

 

[Afforess]

One suggestion I have for Better AI that I've already implemented in my mod:

Asset & Power values for promotions. Not static values like for Units, Buildings, etc.. but multipliers. E.g Combat I would have a power multiplier of 10%. When you give a unit combat I, 10% of the units original power value is added to your total power level.

An example of this. A rifleman with 100 power is given Combat I with a power multiplier of 10%. ((100 * 110) / 100) - 100 is added to your power level. This will help the AI, especially facing a human, who might have more heavily promoted units, gauge when to attack. It also scales with unit strength perfectly.

The one drawback is that when a unit dies, you will need to remove all the promotions from the unit, or the power and asset values from promotions will be left over as garbage values and skew your real values.

 

[jdog5000]

Are there any plans to add the most recent (v1.2) version of lead from behind? Does anyone have the earlier versions that would make that task a bit easier?

Yes, thank you for bringing that to my attention.

Could you give a quick, rough overview of the range of values for these new tags and how much of an affect altering them would have?

I updated the victory system thread with corrected numbers, here's how the distribution of tendencies should work:

- 0 is an extreme value, meaning the AI will never do a strategy in a normal game. These should be very rare.
- 30 is a good middle for the distribution and represents an average affinity for the victory type.
- 70 is also an extreme value for a normal game, they should be very rare.
- Each leaders 5 values should add up to 150, so that on average they start a game adopting 1.5 victory strategies at level 1.

As for what makes sense for a particular AI, Civ4 been around long enough for the community to know which leaders are good/decent/bad for different game styles. Having traits which play into a particular strategy certainly helps as do other AI characteristics like their odds for war and building units. Historical actions by the leaders may also be relevant.

hmm.... developer 0.90r:

		if( GC.getBBAI_DEFENSIVE_PACT_BEHAVIOR() > 0 )
		{
			cancelDefensivePacts();
		}

		if( GC.getBBAI_DEFENSIVE_PACT_BEHAVIOR() > 0 )
		{
			GET_TEAM(eTeam).cancelDefensivePacts();
		}

should not this be "if( !GC.getBBAI_DEFENSIVE_PACT_BEHAVIOR() > 0 )" [if not] ?

Thank you for catching that.

 

[jdog5000]

Just posting to say that I have played several games so far with 0.90 and it defenitely needs more testing and balancing.

First of all... I have seen a lot of stupid wars, the AI apparently tends to start several wars togeter.... which is caused probably by the new play to win system.

Yes, this was a big issue ... the AI would be over aggressive and get itself bogged down, make no progress in its wars and fall behind in tech. I had done some work to improve AI decisions on when to open a second front, and at your prompting just took another pass. Should be much better in the next version.

Can the player still 'freeze' the AI? What I mean by that, and for example, I was playing a new game, destroyed Ghengis Khan's second newly built city and then I hung around entrenched in the borders of his remaining city, stayed in a state of war, and he just sits there, never trying to get out of trouble. You could actually forget you are at War because of the inactivity. In another case, another leader built approx. 30 archers in a city and never tried to get out of trouble with them. A leader could do a lot with a SOD of 25 archers leaving some behind to protect the city.

I think this freeze problem exists in many situations. Is this the best the AI can do when they are hit hard but, seemingly to me, could find some ways to get out of trouble?

Yeah, it's still possible to choke the AI early. Once they have a couple cities, they can put up much more of a fight now. Will look into this.

One suggestion I have for Better AI that I've already implemented in my mod:

Asset & Power values for promotions. Not static values like for Units, Buildings, etc.. but multipliers. E.g Combat I would have a power multiplier of 10%. When you give a unit combat I, 10% of the units original power value is added to your total power level.

An example of this. A rifleman with 100 power is given Combat I with a power multiplier of 10%. ((100 * 110) / 100) - 100 is added to your power level. This will help the AI, especially facing a human, who might have more heavily promoted units, gauge when to attack. It also scales with unit strength perfectly.

The one drawback is that when a unit dies, you will need to remove all the promotions from the unit, or the power and asset values from promotions will be left over as garbage values and skew your real values.

Good idea! Would need to be sure pluses and minuses always add up right, but aside from those details this would definitely help the AI.

 

[Tholal]

Is there a different forum that discusses and tracks the latest dev version? If I go to the main Sourceforge page for Better AI, I only see up through version 0.84 available. But then I followed a link from the Better BUG AI thread and got to a developmental version, for which I then had to change the URL to end up on what I think is the latest version, based on a recent post in the Better BUG AI thread.

http://civ4betterai.svn.sourceforge.net/viewvc/civ4betterai/trunk/?sortby=date&pathrev=550

It's all a little confusing, especially when the first post in this thread points to an outdated dev version and the first post in the Playing to Win thread doesn't have any links at all!

 

[mt1]

jdog5000
I am not advanced in development, but I'm a big fan of your work and translator (rather adopts to russian lang) of your modifications on civfanatics.ru.
I and my friends look forward to v0.90, we would like to know when the release date approximately.
I apologize for mistakes, my native language Russian, best regards.

 

[entwood]

Is there a different forum that discusses and tracks the latest dev version? If I go to the main Sourceforge page for Better AI, I only see up through version 0.84 available. But then I followed a link from the Better BUG AI thread and got to a developmental version, for which I then had to change the URL to end up on what I think is the latest version, based on a recent post in the Better BUG AI thread.

http://civ4betterai.svn.sourceforge.net/viewvc/civ4betterai/trunk/?sortby=date&pathrev=550

It's all a little confusing, especially when the first post in this thread points to an outdated dev version and the first post in the Playing to Win thread doesn't have any links at all!

I would like to be be able to get the latest version also but I can't seem to find the correct SVN URL to use. What is it?

 

[Tholal]

I would like to be be able to get the latest version also but I can't seem to find the correct SVN URL to use. What is it?

I found a link to keep track of the latest dev versions - http://sourceforge.net/projects/civ4betterai/develop

 

[entwood]

Thanks. I tried to copy the latest build from SVN using
https://civ4betterai.svn.sourceforge.net/svnroot/civ4betterai/trunk/

and apparently got the 0.90m version but there is a developer version 0.90n version already available as a ZIP and Better BUG AI already has a 0.90r version and I still can't seem to obtain the latest version.

 

[Tholal]

Thanks. I tried to copy the latest build from SVN using
https://civ4betterai.svn.sourceforge.net/svnroot/civ4betterai/trunk/

Hmm... it works for me (though I'm pointing directly to the DLL subdirectory). What revision do you see in SVN?

 

[entwood]

I see another new version. By looking carefully at the SVN I think it can be obtained using URL

https://civ4betterai.svn.sourceforge.net/svnroot/civ4betterai/trunk/Better BTS AI/

I think I needed the next directory after trunk/

Edit: Yes, now this seems to be good, my mouse-over says better ai version 1.00a, so that is the latest. thanks for trying to help me out.

 

[jdog5000]

Sorry guys for the multiple dev versions in different places ... I've changed up how I have my folders set up so the mod version in svn should now be updated more regularly. It'll work like this:

svn mod file: basically what I'm working on, newest code but potentially rough around the edges

dev releases on CFC: occasional "beta" packages that are a little more polished, but not really tested

official releases: intended for games, mergers, etc, have been tested

Also, I decided to skip 0.90 and just number the next release 1.00. The change list is huge, and the overhauls to combat AI and the victory strategy system make it a major milestone.

 

[jdog5000]

jdog5000
I am not advanced in development, but I'm a big fan of your work and translator (rather adopts to russian lang) of your modifications on civfanatics.ru.
I and my friends look forward to v0.90, we would like to know when the release date approximately.
I apologize for mistakes, my native language Russian, best regards.

Nice to meet you! And thank you for helping make the mod available to more people, that's wonderful.

There are a couple more little things I want to do before the next release, to put the finishing touches on changes that have already been made. May 16th is my target date for the official 1.00 release, I'll post here if things go faster or slower than scheduled.

I'm going to put up a new beta 1.00a version very soon. This beta version seems pretty solid and has had some testing, so certainly if you've been playing with 0.90 I would suggest switching.

 

[jdog5000]

Development version 1.00a posted here at CFC.

 

[Afforess]

How soon until it merges with RevDCM's SVN? I ask because merging BBAI alone with my mod is an impossible task now.

 

[Maniac]

Also, I decided to skip 0.90 and just number the next release 1.00.

Does this also mean you will consider Better AI complete, and stop working on it further?

 

[Cybah]

How soon until it merges with RevDCM's SVN? I ask because merging BBAI alone with my mod is an impossible task now.

I know what you mean. U should compare the newest BBAI with the BBAI files you are using in your mod. Remember the files which differ and then compare these files with your mod files. So you won't have to compare ALL BBAI files with ALL of your mod files.

 

[Caboose]

Minor detail:

I have Angkor Wat and set my military city to prioritize production (hammer button). It removes a free priest (mercantilism) and adds an engineer. There is absolutely no reason why it should behave like this when I have the Angkor Wat. Is it possible to fix? I know I can force a priest, but maybe it's an easy fix anyway?