Real Strategy (AI)

Real Strategy (AI) 2.3.1

PSEUDOYIELD_CITY_DEFENSES - this controls the valuation of cities when preparing an attack, has nothing to do with defense
OP_DEFENSE - this is a tag for defense ops, on strategic level you can only control the number of simultaneous operations, not how they are executed; when a civ it attacked this mod increases this number by 1 or 2 depending on the situation, allowing the cit to protect more cities at the same time
BaseOpsLimits - this is the list to control number of ops
It is like I said - this mod deals with Strategy, not tactics.
You just guessing here. You don't really know how those database entries are connected to game core ddls.
What I know for sure is that with your mod city defenses become non-existant. My second game and again AIs are lame ducks with no walls and only have a few combat units. I've also noticed AIs exploring less. My agressive style gameplay experience deteriorated significantly with this mod.
 
You just guessing here. You don't really know how those database entries are connected to game core ddls.
What I know for sure is that with your mod city defenses become non-existant. My second game and again AIs are lame ducks with no walls and only have a few combat units. I've also noticed AIs exploring less. My agressive style gameplay experience deteriorated significantly with this mod.
1. I use the collective knowledge gathered in mods like AI+, Airpower Fix and few others. Many parameters I have tested quite extensively. I do not use parameters that I am not sure how they work (e.g. Sensitivities lists from R&F). I will eventually enclose a description of how things works imho.
2. I believe it is entirely up to you to select mods that enhance your playstyle.
 
I downloaded this and gave it a quick spin with FireTuner running to get a feel for what it was doing under the covers. I just loaded a game I started yesterday, played a single turn and thought I would load the same game again and play the same turn again to see what differences there were... I was not expecting any, but it surprised me a bit, so I thought I'd ask. Note: I am not a Mod developer and know very little about the internals of Civ, so appologize if I get it wrong.

First the numbers...

Turn 39 -> 40 first load:
Code:
 ...
 RealStrategy: 39       *** base priorities LEADER_BARBAROSSA     : CONQUEST 100.0  : SCIENCE 140.0  : CULTURE 80.0  : RELIGION 20.0
 ... 
 RealStrategy: 39    ...game turn adjustment (iMaxT,iCurT,perc) [1] 500 [2] 39 [3] 34.04
 RealStrategy: 39        specific & generic after turn adjust     : CONQUEST 10.6  : SCIENCE  5.1  : CULTURE  5.1  : RELIGION  5.1
 RealStrategy: 39                applying specific priorities     : CONQUEST 110.6  : SCIENCE 145.1  : CULTURE 85.1  : RELIGION 25.1
 RealStrategy: 39             applying a bit of randomization     : CONQUEST 135.6  : SCIENCE 162.1  : CULTURE 110.1  : RELIGION 39.1
 ...

Turn 39 -> 40 first load:
Code:
 ...
 RealStrategy: 39       *** base priorities LEADER_BARBAROSSA     : CONQUEST 140.0  : SCIENCE 160.0  : CULTURE 80.0  : RELIGION 20.0
 ... 
 RealStrategy: 39    ...game turn adjustment (iMaxT,iCurT,perc) [1] 500 [2] 39 [3] 34.04
 RealStrategy: 39        specific & generic after turn adjust     : CONQUEST 11.9  : SCIENCE  5.1  : CULTURE  5.1  : RELIGION  5.1
 RealStrategy: 39                applying specific priorities     : CONQUEST 151.9  : SCIENCE 165.1  : CULTURE 85.1  : RELIGION 25.1
 RealStrategy: 39             applying a bit of randomization     : CONQUEST 167.9  : SCIENCE 170.1  : CULTURE 98.1  : RELIGION 25.1
 ...


As you can see, I got different base priorities for Barbarossa on the second load than we did on the first. I had assumed that the base priorities would be static. Can you explain a bit more how the base priorites are caclulated and why they are variable between game loads / turns / strategy changes?

Also, the randomization seems a bit... random! I would have thought that you would create a single randomization factor per leader, per game, and apply this same factor throughout the game to adjust how they behave for 'this game'. However, the %'age differences for each of the priorities suggest that this is calculated at least at game load, but maybe per turn / strategy change. Can you explain a bit more about this randomization element?

Thanks
 
@docdef Randomization is applied when the mod loads, so each time is a bit different. As for now I dont save any data, but eventually it will be just like you described - generated once and used for entire game.
It’s a minor feature, I added it to get more variety in testing. If you want to play with stable base flavors just set it to 0 in RST_Params.sql.
 
@docdef Ah, there is also randomization during calculations (the line that says “applying a bit of randomizatiin”) each time a scoring is done. It adds 0..30 to score. This is ofc run each time, and is just pure random. And this is by design.
 
@docdef Randomization is applied when the mod loads, so each time is a bit different. As for now I dont save any data, but eventually it will be just like you described - generated once and used for entire game.
It’s a minor feature, I added it to get more variety in testing. If you want to play with stable base flavors just set it to 0 in RST_Params.sql.
Are there by chance more test settings that made it into your release that we need to know of?

@docdef Ah, there is also randomization during calculations (the line that says “applying a bit of randomizatiin”) each time a scoring is done. It adds 0..30 to score. This is ofc run each time, and is just pure random. And this is by design.
Is "by design" by Firaxis or by you?
 
Are there by chance more test settings that made it into your release that we need to know of?
Is "by design" by Firaxis or by you?
It is like in description: this is a playable version for testing purposes. All parameters are explained in RST_Params.sql and as for now all are sort of "testing".
"By design" be me.
 
Yes, I read that this is a test version ;). I find it informative to know what is considered core (or "by design") and what is considered experimental by you nonetheless, which you did now. Thanks.
 
For what it's worth, I think this mod would gain a lot more traction on Steam if you squeezed the word "AI" into the title.

Anyway, Infixo, do you have any insight into whether this is multiplayer compatible? As in, do you think the Lua functions would cause desyncs?

I tried using RST in multiplayer. It caused significant AI slowdown. Another side effect is that the player who is not hosting the game is regularly disconnected while the AI players are taking their very slow turns. I tested the same game with the same friend with and without the mod, and this mod is 100% the source of the trouble. I've attached the Lua.log output from that game.

Edit to add: no noticeable slowdown in single player.
 

Attachments

You can remove Airpower Fix from a save, but it is a bit tricky. You need to set the flag AffectsSaveFiles in .modinfo to 0, load a savefile and save it again. The mod will no longer be stored in the savefile, so then you can disable it and load a game without it. This trick is only possible because, in fact, Airpower Fix doesn't affect savefiles. If you do that with the mod that actually does affect - the game will either won't start or crash.
Is there a way to add this, the code, if this mod doesn;t have it or another mod? I was trying to do that to this mod and another but there is no AffectsSaveFiles in .modinfo. I am a novice to coding.
 
@Raven_2012
In .modinfo you need to add the line
Code:
       <AffectsSavedGames>0</AffectsSavedGames>
Some mods don't have this line because it is optional. And if it is not there, by default this flag is considered to be 1. So, each mod that doesn't have this line by default is registered in the save file and required to load the game later.
 
Infixo updated Real Strategy with a new update entry:

Version 0.6 Major update

Version 0.6
Major update with lots of tweaks and few new functionalities.
- Added: Support for MultiPlayer. See details in the description.
- Added: More Science - activates when a civ is behind in num techs.
- Added: More Culture - activates when a civ is behind in culture.
- Added: More GW Slots - activates when a civ has more GPs than slots for GWs.
- Added: Support for Government Plaza buildings.
- Added: A fixed BH "Manage Great People". Removed one node that caused the tree to get...

Read the rest of this update entry...
 
@Raven_2012
In .modinfo you need to add the line
Code:
       <AffectsSavedGames>0</AffectsSavedGames>
Some mods don't have this line because it is optional. And if it is not there, by default this flag is considered to be 1. So, each mod that doesn't have this line by default is registered in the save file and required to load the game later.
Thanks bud, I got it.
 
@Laurana Kanan
- More Science - compares num of techs to the average of known civs, if you are behind much, then it prioritizes science, e.g. enforcing building a Campus.
- More Culture - compares culture yield output, activates when below 75% of an average, then prioritizes culture, e.g. building a Theater square.
 
As anyone got to mid and late game with this mod? Is the AI build air and using it?

I'm not using the mod, but I changed the air units combat strength to the ones they will have in GS and the AI seems to be building quite a bit of them.

I'll definitely check this mod out :D
 
I'm not using the mod, but I changed the air units combat strength to the ones they will have in GS and the AI seems to be building quite a bit of them.
I'll definitely check this mod out :D

AI scores the units to build by their Combat Strength and tends to choose the strongest possible atm. So, yes, just by simply increasing planes’ CS you should see a bit more of them.
And, btw, this is what Firaxis was referring to during one of the streams, claiming that they “increased air power” or something like and people understood that they fixed AI for using planes. They increased planes’ strength which was clearly visible in the streams.
But I still want to see the actual usage which is consitent not accidental. E.g. fighters attacking bombers or ground units, and bombers taking down cities. Regularly.
 
Back
Top Bottom