Quite often in a game I conquer an AI developped planet that has specific tiles, like bonus to population, bonus to economy, etc, being misused.
It may be that the AI was desperate and had to use these tiles for something else, or it may be that the AI is ignoring the tile's bonus. For example, a market center built on an influence tile. A research center built on a economy tile. Little stuff like that annoys me. I like efficiency, the AI does not seem efficient
I'm playing on normal, so maybe it's only something that kicks at an higher difficulty level?
I have yet to see the AI build anything. Not on any of the planets I inherit...
Really? I'd say the AI is a lot better than with beta 5 at that, but while it is lacking, it is building.
I also observed that the ai wasnt building anything. And if it built stuff, it doesnt group the same type of buildings to harvest adjecent tile bonuses. I am playing on the most difficult level with the godlike ais and they are still so stupid. Will that be further improved?
The AI is still building shipyards first on just about every colony.
I think if they can get the AI to at least build up a planet like a average player the AI would be much harder to beat on.
I'm not sure the AI has been informed of all the changes. Shipyards use to be easy to build, now they are quite costly. I have found though that you can bi pass this by creating a constructor and then building a ship yard...much cheaper. This seems kinda silly that you can do that if the change was indeed meant to make ship yards harder to get / easier to defend early game.
I think that the AI does a decent job on larger planet, before it gets to terraformed the tiles. As a player, I often build up with the space I get, then as I can terraform tiles, I will destroy building to move them to a more appropriate place. The AI simply won't do that, it will uses the tiles as they come, and worst, it does not seem to recognize tile bonus.
I still like the change. No more camping shipyards. Also the travel time kind of makes up for the constructor cost.
Why not, they were rushing them for free in a soak game I watched today
Playing with all the AI on Gifted. Every AI planet I've conquered has been very poorly built up. No focusing on manufacturing at all. They build ships at each planet's shipyard before having built up the planet.
Hovering over AI planets to get their stats. Most have very dismal manufacturing and science production.
This is pretty much the crux of the problem with the AI. All of the complaints about AI planets flipping and overall weak play would be solved by better tile selection by the AI. Can they be made to do it? I don't know.
Hell, just filling up the tiles would help a lot, even if the selections were poor. Surely they could be made to fill up the tiles.
Hope Stardock fixes it.
Perhaps someone could create a game mod that greatly simplifies building up a planet. Makes everything so easy that the AI can't fail at it.
Remove everything complicated from the tech tree that concerns planet management. Change the adjacency bonuses to 1% or remove them. Make building upgrades super cheap.
It seems to me lack of AI development of its planets is a critical flaw in this otherwise superb game
I've also seen this. Conquered two Yor planets so far. On the first, there was one Yor Manfacturing Building. One the other, there was nothing (with a shipyard in orbit).
I'm only playing on Normal AI at the moment, but if it's always building a shipyard first... that's just silly. Especially when shipyards can be jointly used.
Not the only problem with the AI, In most of my games they seem to be unable to put more than 3 ships in a fleet.
Failing at simple basics imho
I played a game last night with Genius AI, but very limited planets to limit their current growth issues. Around turn 150 the Altarians attacked me without cause save for trading with opposing ideology and I was an opposing ideology. The Altarians actually had about 15 fleets of ships ranging from 4-11 in size. I thought I was going to be in for a fight. I only had 2 fleets with 9 ships each. I crushed all their fleets and after watching one battle it was obvious why. The fleet make up was very poor 85%+ of the fleets were all carriers with little to no attack or defense. My fleet make up was 6 small escorts (12beam/12ish all def) and 3 medium (16beam24gun/20ish all def). I never researched past the lv1 weapons or defense. I used the same fleet setup for the next 100 turns and beat any AI fleet from 3 other AIs. I am used to GalCiv2 where the AI sees my attack and changes it up. I have to change my ship designs every 10-20 turns.
Another thing I noticed was when I took over their planets I had to destroy 80%+ of their tile improvements. Almost nothing was linked and the AI likes to build the resource builds yet have them linked to nothing. I saw a Thulium data archive next to farms and even the farms were not next to each other. About the only thing that was linked was Durantium refineries, yet almost no factories linked to them.
The AI needs a good polish to ship design, fleet setup, and tile management.
I'm seeing this too. I'm not familiar with this area of the code but I will be looking into it before release. I do t think I'll be able (or allowed) to check in a major overhaul of this before release but it should be pretty easy to address some of the obvious stuff (like the shipyard ability).
theres a lot of good stuff for the AI with planets in the future that should let it crush you humans. Ais are way faster at optimal placement than humans. It just needs to know what optimal is which comes from playing.
i will be asking for saved games from people winning on hard difficulties so I can check out your planet strategies.
They also queue up durantium refineries right after the shipyard, despite the 30 turn build time, the minimal benefits, and the use of a limited strategic resource.
The problem is that there are no direct optimal dedsign of planet since it depend on when in time and the empire as a whole.
A human player can design a colony one way in the start and a complete different at the end of a campaign. The AI would need to be able to destriy and restructure their colonies just as an efficient human would.
At the start of the game you want to concentrate on lots of basic buildings. Once you can get some approval and food buildings an populations start hitting the cap of the colonies you will destroy some buildings to make room for more population. As technology rice and you get more buildings that will add to the overall effectiveness you will start to replace and/or upgrade buildings.
It is not even clear that upgrading buildings are the best path as soon as you can do it, many times it actually is not since there are more pressing matters to attend to.
My worlds are often living entities that change focus throughout a game... some more than others.
This is very true but there are also some fundementals that are still important. For example, two or three (depending on planet quality) factories next to each other and if possible next to the capital or around a nice manufacturing bonus tile should always be the first thing done. Afterward, a quick look at the planet can set which way you would build most of your buildings afterwards (Science, Manufacturing, Wealth). Just setting this in place would do wonders for the AI. After that, with all the extra buildings comes the back and forth consideration that is sometimes difficult to replicate in an AI.
This is actually an area that AI's excel at compare to humans. If the AI knows what to do, it can check all planets every turn and make small adjustments to keep things optimally...like building a farm to immediately coincide with a population cap. This kind of micromanagement, while possible for a human, is hard to do at scale. So once the AI is good at this, it will generally do it better than us fleshbags:)
Yes.. I do agree that the AI is much better than us to keep tabs on thresholds and things like that. What is is rather bad at is judging when it should change focus and turn a research world into a completely different focus and increase industry or even changing the world into a wealth/tourist world.
The problem with the AI is to stick with a certain path or when to change it.
The problem right now is just getting the AI to work. What you are talking about is something that will be tweaked until the end of time.
Yes, the AI could be made excellent at the tedious micro. Then we would have to rise to its level. Would you want to do that with 50 planets? Even only 25? I can do it but its not adding to my enjoyment of the game. In multiplayer, it would be a nightmare.
I like the research choices. The ship design. Starbase building is a little tedious but not too bad.
The United Planets voting is not having any effect on my games. Star Ruler 2 is entirely better in this area.
If building up a planet was made very simple. Everything made so easy it was impossible to be sub-optimal. I don't think my game enjoyment would be any less. Planets generating bigger numbers doesn't excite me.
Being threatened by an AI's fleet of ships coming at me is exciting. This brings enjoying gameplay. Tweaking sliders and such. Again and again. Plunking down buildings in obvious patterns. These don't do much for me.
Why not there are still many area were an experienced human still have the advantage over the ai. So it could help even the balance between ai and human in game. And if a player doesnt want to deal with that at all, he could probably lower the difficulty setting, hopefully they will allow us to fine tune those setting further later.
There are many great features available to you once you register, including:
Sign in or Create Account
