Yes.  

I agree with naselus. It is not only time to begin putting up, it is time for Stardock to say what they intend. I think there is an understanding at Stardock that the AI (broadly used - I consider everything the AI 'player' does to fall under the category of 'AI') is not adequate, or at least an understanding that there are a lot of players who believe the AI is not adequate. Stardock would not be launching threads asking for improvements if they intended to ignore what people say - at least that's not my experience with the company. And at this point I think there is some consensus among the players - explained in deep and comprehensive detail - as to where the issues lie.

Second, while there are a lot of good suggestions (IMHO) here, Stardock has already pulled most of the labor off of a 'finished' game and is moving on other projects. This limits what they can do and there is no point in us debating whether that was a good decision or not - it is done, and was done for good and sufficient reasons, and we can't revisit it.

Third, opening up the AI by providing new variables (triggers) looks to me to be a win-win suggestion, satisfying the customer base while using minimal programmer resources. But I'm not a professional game-programmer or Stardock employee and there may be reasons they don't want to do that.

Fourth, the 3-D ship modeling was probably something a Stardockian developed because they were interested in it, and it was put out to the public as a 'this is really cool' way of separating the game from others of the genre. And it is extremely cool, and it probably did not soak up major time or resources, and we would not have gotten a better game without it, so criticism here is (I think) misplaced. This falls under the category of, 'the meal includes desserts one of which is ice cream, and perhaps you should not eat it if you don't like ice cream, but the restaurant should not remove it from the menu because you don't like it'.

Fifth. Brad, I understand this is your game, and I understand there are times when you can or can't talk to the public - and about what - and how. I think we've exhausted this topic and I sense some frustration that there hasn't been a substantive response from you. I don't take your previous comments on the subject as defensive - mostly attempts to explain the nature of the problem - but I can see how others would. I think - hope - that not saying anything means you guys are taking notes, mapping options and preparing to do whatever it is you want to do. And I can see that you don't want to make promises that you later find out you can't keep, and so would rather show than tell. Still...

I'm asking you - please - to take the time to give us a long-form answer. Tell us if you agree there are issues, lay out what you are looking at, and sketch out a proposal or two. Because a lot of us love this game too and are frustrated, and nothing will get better out of silence. I know that you don't want customers you want raving fans and you have them. But part of having fans is that you wind up with a lot of people who feel invested, who feel like they are your partners even though we aren't. Sometimes that's great, sometimes it sucks - but if you want the fan involvement this is the downside. Sometimes we need you to explain.

So I'm asking everyone to calm down, post on this only if you have something new and original to say, and give Brad and Stardock a little breathing room (say, 15-20 minutes - no, I kid - I kid - make it 30). I'm confident we will get communication from them. I'll surely be disappointed if we don't.

Personally I'll probably wait until Nov 10th for significant AI improvement. After that, I'm sure I'll be busy learning to play and mod FO4, then probably on to Stellaris Springtime next year.

I appreciate the information. I don't generally watch the developer videos. I do think everyone would be better served with a written response. Given the length, and depth, and strength of the dis-satisfaction, I think it is in Stardock's best interest to do this.

And I appreciate the modding - but I have gathered from Naselus's posts (spelling?) that the issue with modding the AI is the lack of triggers in the code, so as I understand it what a modder can actually do now is limited. Maybe Stardock will fix it and maybe they won't, but I'm going to go do other things for a while.

Hm, what settings are you all playing at? I just play my third game and hasn't made a science of finding out how the AI works so I surely still make a lot of mistakes (or, rather, don't use my means effectively; until I read this thread I never used the production wheel at all).

My second game was like that: insane map, tight clusters, nearly everthing set to rare, slowest tech and game pacing, played as the Yor with all other seven major races as opponents. I found myself in a corner of the map and was only able to find three colonizable planets within range of my ships. By the time I extended my range enough (through research and starbases) all other planets near my territory where taken by AI players stronger than me. I soon found myself in a war against the Krynn who were the most powerful faction of those I encontered thus far (four of the seven in game apart from me). They had more and better ships than me (and adapted to my missiles), way more techs, but didn't do very much to fight me so I eventually managed to take one planet from them. Slowly I became better, especially after an event that spwaned two new class 25 planets near my homeworld, and managed to leave the bottom of the ranking and overtook the Thalans and the Iridium Corporation. Then, around turn 500, the game was suddelnly over because the Drengin whom I never met won a Technological Victory.

So I don't see a non-researching AI, I have no clue how such a game where it takes tens of turns to research a single tech can be won by turn 50 or 100 or even 200. I also have no clue how one can get a fleet with 10,000 weapon strength while the best AI has some 500 ...

But I also have the feeling the AI is not aggressive enough against weaker opponents. In MoO 2 (sorry, I must refer to the still best space 4X game I ever played) you could not remotely hope that your opponents would leave you alone when you were so weak.

But whether that is so because of bad AI logic or missing bonuses or whatever, I cannot determine. Only because there are only a few triggers in the XML files that doesn't mean there are not a lot more in the code that simply are not exposed. How can you be so sure what problems the AI has when we have no information about the internal logic? Or have you?

Anyway, as I stated in another thread, I would like SD to give us a real scripting interface with a lot of callable functions instead of only some XML files ...

Techs shouldn't take tens of turns... Specialize with the planetary wheel some of your worlds into research planets and build up their population as well as make multiple research buildings, preferably in hub-like formations. You can get tons of research with dedicated planets.

econ worlds are almost (if not) more important than  research/mfg ones

You trippin?

I don't disagree with the importance of having econ worlds.  Trade can only get you so far, especially if you are going down the conquering route.  You don't need many research or industrial specialized worlds to bang out advancements or fleets.  Funding those worlds and your military requires a good economy, which you need econ worlds for.  I totally am all about being self-reliant on my own internal income, which is why I agree that econ worlds are much more of a priority than research/industrial worlds.  

1. The bonus qualifier is the weight required to pick that build queue. It works in tandem with 2 lines in GalCivAIdefs.xml - <PlanetTraitStatSamplingFlatBonus> and <GovernorWorstAcceptedScore>. The exact calculation is not known(I suspect it adds up all the %es on the planet and picks the best one over 20%), but basically it 'samples' the bonuses on the planet, weighs them, and then selects the build queue which closest matches it. There's a few other triggers listed in the schema - including one to check if it's the homeworld - but most seem to have no effect.

2. As above.

3. They're built immediately. This is a bad thing, which I actually wish WAS in the xml so we could change it, as it means that AI wastes time opening up extra tiles when it hasn't used up the ones it already has, at a point when it has the worst industry that world will ever have.

4. Yes, the AI determines placement itself, and was made aware of adjacency in 1.1. As the patches progress, I suspect most problems will be solved in the C rather than through additional script options.

5. I think there is, actually - again, in the C. In globalAIdefs.xml, there's a definition for 'minimum level to be immune to replacement'. This was also added in 1.1.

An easy fix would simply be to make the slider settings usable in the build queue, tbh. That'd do in the short term while SD wrote up an actual subroutine to deal with it, which could be mercilessly efficient from day 1.

ai yor will frequently start a planet with assembly before building even a single factory as another "wtf" inspiring choice

-: I'm not sure, but it would probably be in GovernorDefs.xml if so

2: there is already a guide

News Flash!

There is no such thing as transcendence!  There is no AI anywhere that has the tiniest bit of independence or unique will or creativity and I dare say there NEVER will be, NEVER!!  That being said, any AI is just a collection of <if then> statements and this AI is amazing!  The fact that it can manage as well as it does on such a complex game as this is a Fantastic! complement to the AI programmer.  Please keep up the good work!