Does the AI ever build starbases at stars?

I've never seen the AI build starbases on uncolonizable gravity wells, star or otherwise.  Similarly, it doesn't organize its own offensives against these starbases very well, and usually ends up getting its fleet toasted. 

Does the AI ever build starbases at stars?

No!

Ok thanks people. It makes my sp multi star games all a bit pointless then when I can use such a simple, and somewhat tedious,  strategy .  I suppose I should go back to vanilla sins if I want loooong games against decent opposition, but I do love all the goodies in Entrenchment. I guess expecting an "AI" to be able to react to all possible player strategies is asking to much, but the AI seems particuarly uncreative in its approach to using starbases itself or in attacking the players starbases.

             It seems the devs might of gotten ahead of themselves..they are working on the second expansion but there are issues that need to be addressed now.  One being all the formentioned posts about lack of ai building sb's in star wells.

       I've come to the conclusion that Vasari need a anti-structure ship because the vasari ai will not build a starbase to take down yours. At least when you play against the Advent you have to worry about those star antisructure ships. This is primarily a single player game these issues should be resolved. Ai still doesnt make superweapons, even on unfair.

If you're looking for decent opposition, then I advise you to come online. I know, I know. You probably don't want to play rush maps that are decided witihin the first 45 minutes, you'd rather play larger maps like the one you mentioned. You know what? There are plenty of people, myself included, that if you suggested it to them, they'll say yes, and do it gladly!

Yes, the people online are very good, but we're also very friendly (for the most part), and have no problem showing players who are new to the online community a trick or two. The learning curve online is only as steep as you make it. If you decide to figure it all out on your own, ur gonna get frustrated quickly. If you ask for help, there are plenty of people who will help.

Hope to see ya online. If you see Raging Amish don't hesistate to ask me anything that comes to mind.

Thanks Raging Amish. Yes online is probably where I will end up  eventually  - once I have exhausted the possibilities of the AI. At the moment I am still way too much of a noob at this game. Just reading the forums re the great LRF/carrier debate shows me how much out of my depth I would be against real players

I was just very surprised to find that the poor Vasari AI effectively has no counter to starbases - at least the Advent have an anti-structure ship.

If the starbase doesn't have too many upgrades, the trusty standby of assailants and nanos will work.  Once it has enough upgrades (or you suspect the "big red button" may be in play) you'll have to go with bombers.  Vasari bombers are arguably the best against starbases anyways thanks to phase missiles.  Bombers really are the best starbase counter anyways.

My point being is at the present moment Vasari ai does not know HOW to counter starbases.

None of the AI's know how to counter starbases.  They'll all run their armies into range of its weapons and leave their Ogrovs or starfish to get minced by a few fighter squads.

What's the point of the devs moving on to other issues before these are addressed?

The AI's a joke right now, that's for sure.  The fact is that once an AI encounters a starbase strong enough to kill its fleet, the game is over, because it will suicide its fleet trying to kill the starbase, and then will repeatedly suicide with weaker fleets over and over again.

I don't know if I'm a better player than I was and I love the sb's, but the ai doesnt give me the same challange as before Entrechment. The devs seem to have a problem with the starbases either being pushovers for online players.....or indestructable when the ai goes up against them.

RTS ais are never that strong. an ai will never be able to out think and outsmart a human player. The only real way for it to win would be give it a huge resurce bonus. Really Its too bad less people play online. The ai is a joke and will always be a joke no matter how much work is put in it. A computer will not be able to outsmart a human player, and it is too predicitble. the only real way to get a chalange is to play online. There is many helpful people online as long as you listen to advice and imrpove you will fit in well with the online community.

The AI needs to learn the following behaviors:

1) keeping its distance: it shouldn't enter weapons range of a starbase unless it has a decisive advantage, otherwise staying out of range and using bombers and/or anti-structure frigates.

2) bypassing:  the AI needs to know when it's worth its time to bypass a gravity well and attack the next one.  The fact that it never does this greatly reduces the difficulty of defending.

3) ignoring:  the AI needs to know when a planet is just too fortified to take down, and instead to find something else to attack.  The AI gets tunnel vision, and often you can leave half your empire completely defenseless against it because it refuses to try a different direction of attack.

The AI will never be a true match for a human, but it could at very least be a threat.  Right now its love of the siege frigate and predictable and simplistic offensive and defensive behaviors make it shockingly weak.  It's also a poor judge of fleet strength, usually overvaluing its horde of light frigates and seeing them torn apart by a smaller force of lrf.

I just got out of a game where my ally quit partway through and was replaced by an AI, and I can tell you it would be very nice if he hadn't suicided his fleet THREE TIMES against the same starbase.

It is pretty funny to watch an AI win the game for you though, even if it has a massive advantage in number and strength of ships it will repeatedly beat its head against the same unimaginably fortified world until it has nothing left. It is almost sad, all the advancements in technology and thats the best we can do... 

Ironclad had two very severe budget limitations for developing that AI.  The first was monetary; Sins is their first game, and at a certain point they had to say that the AI was "good enough".  The second budget, and arguably the more restrictive, is the performance budget.  The system requires for Sins are low (something done on purpose), and in order for that to work the AI cannot be a CPU hog.  It's got to use minimal resources so as not to cause performance issues.

The problem with the starbases are the starbases themselves.

Remove the starbases from Entrenchment and you'll have one hell of a game.

Making the starbases "togglelable" (is that even a word?) would be nice and not too difficult for the devs.

We can already toggle on/off the pirates, why not starbases.

I love Entrenchment but the AI cant beat the starbases and they also make games last a lot loooooonger; and they lasted long enough in vanilla.

So my recommandation is to make starbases togglelable (if its wasn't a word before, it is now!) before starting a game.

I haven't gotten Entrenchment yet, waiting for a reply to my question post on that before I go in- however, you cannot look at a campaign on the basis of start to complete, but rather, of testing your manueverings on internal-external lines- how defense operates in a dynamism, manuever capabilites, how to manage your convoy formation (I really wish there was formats where you could choose how fleets moved together- like a honey combpattern with certain unit types being here or there on the pattern- so you can take advantage of tactical depositions. It would be silly simple in my opinion to do that- the AI would have to collectively go off the leadership and computing capabilities of the fleet to maintain formation).

If it is too simple, reduce your scope and set it up so it is not. Place a purely defensive war with weaker units, or make your frontline worlds UNFORTIFIED. It was the belief of the Mohist that a section of the wall had to be left undefended in order to control the direction of the attack in order to make preporations for a successful siege successful- if you know the behavior of the AI so well, exploit it as the Mohist did to creat increasingly flawless and skilled victories. The AI uses simple, Jomini style strategy- make it as complicated as you want, master the art of the ambush, tactical withdrawl, and economy of force at hardto defend points. Always look for the nature of the underlying strategic enviroment- you are part and parcel to it, learn to operate in it. CHess varients exist for those who found orthodox rules to easy- so change your approach. Read up on strategy- real strategy, nor forum FAQs, and test them out.

I never played online, so I may be looking into that here soon. Mennonite Devestor or something will be my name, I'll decide later.

In multiplayer starbases aren't a big deal.  Most people end up with huge bomber fleets late game anyways, and you need one hell of a starbase to actually survive that, even with fighter backup.  It's the AI that has trouble with starbases.  As I said, if the AI knew to keep its distance, bypass, and ignore starbases, it wouldn't be a problem.

What are you talking about?, looks like you're trying very hard to sound like you know better than most how best to impliment strategies. Anyway, we are not trying devise new strategies here, we are saying that the AI can't best starbases. Its too dumb for that. Right now the AI sacrifices high level Cap ships for no good reason, thats the problem!

I do know most- I studied strategy for 16 years.

The issue with thecap ship sounds simple, give it a  2-3 default pattern for a convoy, and make sure that they don't enter into a gravity well without skirmishers within that convoy and a second one tagging with a radius of the main convoy- best that second floater group operating within that main convoy be carriers. This is done with geometrical formating- of how things will move in a group. I recommended the honey comb because you can line your larger ships up where the angles meet, and the skirmishers can be occupying the angle or inside the honey comb- the honey comb can have a orbit around it (the radius) in which the smaller, second group would operate. The ability of the ships to maintain synchronized turning would be dependent on the existence of a Cielo, if it is not there, they should be far more garbled and messed up. I think the encoding for the behavior would be rather eas as you not really reprogramming how every ship individually moves- but rather two to three level to the relativist harmony of a group movement- hence the 'convoy'. You can have a larger second radius that stops AI convoy with bombers or missle capability upon touching the battle base. This would solve most of the issue here brought up, and negate the pressing need for the AI to have bettle bases in the first place- a strong fleet, dynamic fleet built to react aggressively by default is far better than battle bases everywhere.

And yes it is egotistical, but tough, it's the damn truth and I'm not dumbing my self down for social harmony in massaging another's world outlook on ehat is the acceptable exchage of information is- I make allowances for difference of opinion and the development of ideas, that is what matters.

Admittedly, most people here have zero background in military tactics.  Our experience is with games alone, so a "2-3 default pattern" doesn't make sense to us.  You'll have to explain what you mean.

That said, I do know there is a stark difference between real-life military tactics and military tactics in a game.  A chessmaster doesn't have the training or experience to become a military officer, just as a military officer might not know the first thing about chess.  Both are strategy, but the "rules" of their games are very different and how they go about winning them diverge accordingly.

From my perspective, when it comes to engaging a starbase in Sins there's no need to overcomplicate the matter.  If it's TEC or Advent, keep your distance and bombard it from afar.  Use whatever formation you want, so long as you keep out of range of its weapons.  If it's Vasari, either kite away from it while your bombers do attack runs, or rush it with all your forces and bring it down quickly.

That said, I do have one question... how would you actually benefit from a "honeycomb" formation?

I do understand what you are implying by creating a honeycomb-like formation to provide support for the cap ship, but IC is not gonna rework the entire pathing system of the game and the entire fleet management just for this issue. Do you know how many hours of work that would require? Maybe if they did a Sins 2 in a couple of years but right now if you're gonna make suggestions, I recommend making feasible ones. Like Darvin3 said, they are on a strict budget and they are currently working on making the diplomatic expansion so they dont have the manpower or cash to rework the entire fleet management system with multiple configurations just to please your egotestical personality.

The good news is that you know you are egotestical, so there is still hope for you.

The bad news is that egotestical people believe they always have the right answer, or the truth.

If I read this correctly, what he means by 2-3 paterns is for example;

Configuration 1: Cap ship in the middle, heavy cruisers in the front, flak/support cruisers left/right and lrm frigates at the rear. Good vs. starbases and heavy fleet combat.

Configuration 2: Cap ship at the middle with flak frigs at the front, support cruisers/heavy cruisers lerft/right and lrm frigates at the back. Good vs. fighter/bomber fleets.

Configuration 3: fleet separated in 2 groups, one for distraction/kitting, one for heavy assault.

And so on and so forth....

Basically what he is proposing is to have the option to reconfigure fleet formations to face different threats while providing cover for the cap ships.

BTW Darvin3, what you said about chessmasters not being good at military warfare, and vice versa is the best example someone could have made. Just because someone "studied" military tactics for 16 years, does not make them good at chess.

Games have rules and boundries that you cant cross. Real military warfare cant be applyed to gaming. In real life there is always the possibility that something "improbable" could happen (a soldier is distracted by the tought of his wife, a shell getting stuck in a barrel, a targeting computer freezing...).

Btw would it kill the devs(love ya btw) to fiqure out a way for the ai to use superweapons?