Ship copying is uncreative, not helpful, and extremely easy to manipulate. 

First of all, the AI can never copy my ship designs because the techs that I use have not been unlocked by them. Even if they had unlock them, they might not have the race traits necessary to make it work. I am talking about Specialization here, particularly mass/capacity in my case. By endgame, none of the AIs have anywhere near 650 capacity on their huge ships, and to even get there, they would need a planet dedicated to the Hyperion Shrinker. Then, they would also have to specialize in nothing but mass reduction. Of course, they could circumvent this by being able to build these designs anyway, but then that would be taking cheating to a whole new level, and make research pointless for the AI.

Secondly, powerful ships in the hands of bad AI are just going to be cannon fodder anyway. Slightly more annoying  and expensive cannon fodder but still fodder, especially without maxed logistics. 

Finally, depending on how much copying the AI does, you could probably just flood your design list with a bunch of expensive, bad designs (bad, not weak, mind you) that you can then exploit with one or two good designs.

Instead, make the AI smarter and design ships better, especially in response to their enemies. 

The fewer credits at start would go a long way to stiffing expansion by the player. I use those credits to buy a sensor ship, then I will rush both factories and research buildings if i am building them on my homeworld. 

Lack of income about 1/4 of the way into the game is what dramatically slowed a player down from over expanding in Gal Civ II. I would colonize 5 or 6 planets and literally run out of money and be unable to do anything except explore for like 50 turns! No research no building nothing, just getting caught up out of the hole of Empire expansion. 

While I really hated this mechanic it really did fix the problem we face now.  Each planet we colonize could possibly be  a flat credits drain/cost on the empire till it becomes able to float its own in support. Say each planet you colonize drained 100-200 credits per turn for 50 turns unless it was only set to economy. That would halt player expansion like a wall. 

Yeah I didn't mind the economic wall actually. Does anyone remember posts about an expanding to fast. The problem is when the ai is getting it right different people complain. Advanced algorithms kind of solve this. Then people can turn on a smarter ai. Other solutions would be have a different ai  for smaller maps, or uncommon settings than have for bigger maps, or abundant settings. You could use the civilization page for the ai to identify play style. 

There are two ways to handle this. First through a bunch of nested decisions the computer could figure out a counter to your stradegies. I originally suggested this because I imagined this is how you make a game ai only to find out designers are more lazy. They just write out a stradegy, so this prompted me to come up with a simpler method. You will still write out strategies only different ones which will be activated depending on how the player holds up on the civilization page. When I say lazy I mean it would require dozens nested if then else statements.

I’ve given Larsenex’s map settings a try (Insane and everything rare except abundant stars, Godlike of course) and had a Research Victory on Turn 187, refer attached save game and in short none of the AI are near a Victory condition.  This follows the save games posted in other threads for a Turn 53 Alliance Victory on Huge and Turn 20 Alliance Victory on Tiny. 

https://www.dropbox.com/s/1d0zgb2o4fsyold/Insane%20Rare%20Turn%20187%20Research%20Victory.GC3Sav?dl=0

I did plan to keep playing with Research Victory off but once you start pumping out Huge ships with Beyond Mortality technologies it becomes clear there is little point playing on, see the screenshot below at Turn 201.

https://www.dropbox.com/s/7vtmz1jzofbsbk0/Screenshot%202016-01-09%2021.55.34.png?dl=0

I’m still interested in any other settings anybody wants to propose where they believe Godlike would be difficult. 

For the record there were no reloads (beyond having a life in between gaming sessions).  I’ve used the first map generated and it was an absolute shocker, out on a thin spiral slice at the edge of the galaxy with a huge void to get to the main part of the spiral … and in particular no nearby Precursor Worlds.

Given how ridiculously easy Alliance Victories have become on smaller maps it was nice to play a game that made Alliances more difficult.  Since you don’t discover many AI for a while, and have very limited resources or anomalies, the AI have the chance to build up and it’s very hard to avoid appearing weak which has a massive negative effect.  As at Turn 187 I have Alliances with most races and but remain at war with a couple and recently conquered the Borg.  Anytime a bridging starbase was built by the AI nearby I destroyed it and was routinely upgrading sensor boats around my space to ensure there were no surprises.  I never felt threatened militarily despite all the AI bonuses and lack of resources/anomalies to help catch-up to the AI. 

All of the Precursor Worlds were at large distances in difficult to defend locations and those I could get to all required Atmospheric Cleansing.  On Turn 100 I managed to get the 4 closest Precursor Worlds (which is a huge advantage given how few colonies there are on these settings) and the strategy to do so was simply parking a colony ship right next to those planets ready for when the required technology was available which was given priority.  Due to the difficult to defend locations I put planetary defences on, but due to the rare settings struggled to find Elerium for the Elerium shield until recently so they were somewhat vulnerable … but the AI never invaded.

Spamming economic starbases led to massive bonuses and is also really boring and my least favourite part of the game.  I just expanded as per normal, spammed starbases, and then at about Turn 100 focused on the research victory technologies.  Even if you are behind the Godlike AI on research output, none of them seem to target a research victory.

With respect to having Abundant Minor Races, as much as I enjoy having them in the game that setting just means there are even more factions where cash or technology can be farmed from the AI.  As usual I had diplomatic technologies early and could use diplomacy to trade for more cash than I would ever need (to the point where I really shouldn’t have bothered with having any credit focused worlds at all and just made them research focused).   I could focus on Research Victory technologies early because I didn’t need to research anything else, I just traded for them.  Free Trade Agreements alone give you about 3,000 credits every time they spawn for all races you have less than terrible relationships with.  Mid-game, often being flush with credits, I asked some of the AI’s to declare war on each other, which both distracted them and weakened them (with a particular focus on races that had declared war). 

As Frogboy has said and as we all know very well, all 4X games have exploits … BUT sometimes it reaches a point where those exploits become excessive and break the game. With that theme in mind there are a couple of changes I would recommend here to Stardock:

1. Give the AI’s better late game strategies e.g. to rush technologies needed for research victories etc.
2. With starbases implement the changes Naselus has made where starbase range is massively larger and they can have multiple functions (e.g. economic and mining) which would both reduce starbase spam and limit the bonuses that can applied to individual worlds. 
3. The AI designs need much better movement speed so that war declarations actually have some menace. 
4. The AI needs to give Precursor Worlds enough focus, either in colonisation or in war.
5. The AI should use the Logistics reduction building and have a fleet filled completely with reduced logistics ships (i.e. a fleet of doom) and send that fleet in first during war, it would certainly give some players an “oh my god” moment or two.
6. Teach the AI to use the planetary wheel since it seems to be remaining in the game for a while.
7. Totally overhaul Diplomacy so it’s not so easy to excessively exploit.

I want to emphasise that I enjoy Gal Civ 3 a lot.  I’m hoping somebody at Stardock is reading this, loading the various save games provided and agrees there are massive opportunities to improve the game in the ways we describe, because it means a lot more fun in the future.  That said, I’m sure I’ll always play more games with expansions and good DLC as an Elite Founder.  Frogboy will say “hey Ice you have 100’s of hours of playtime so you got good value” and my response is yes I totally agree.  But my counterargument is that the game is good but not great.  The games I always come back to in time, which I consider to be great, are the ones which don’t become too easy and where the strategy I had on Turn 1 has to be totally changed on the fly.  It’s only when more people actually use these strategies in Gal Civ 3, express their surprise about how easy it makes the game rather than being defensive, that there will be a realistic chance of significant change.  At the end of the day obviously Stardock run a business.  Stay quiet, refuse to acknowledge legitimate issues, and the game will remain good but not great.  Be vocal (remember the response on the planetary wheel? remember how Distant Worlds improved after community feedback?) and there is a chance this game will become great.  And isn’t that what most of us want?  Now if you already think the game is great now … well all I can say is those rose tinted glasses are powerful!

The AI could be coded to exploit all of this a lot better.  However, if that was done without also balancing/improving the mechanics then a lot more new players are going to find Normal really tough.  They are going to feel the AI is cheating when it isn’t.  So as is often the case the best option is to combine the proposed AI improvements with balances/improvements.  That said, there really are some quick wins as well e.g. increase starbase range/multiple functions, remove a 0 from the value of Free Trade Agreements etc.

I haven't posted in this thread yet but with all the talk about AI and game difficulty recently I thought I'd give myself a potentially pretty tough challenge...

So here's a link to my new Let's Play Terran Custom Race -2 ALL Racial traits yep everything Cargo Space, Diplomacy, All Production types the lot! Godlike difficulty.

Daily updates on:  https://www.youtube.com/user/MacsenGaming/

Well, nobody seems to beat the Godlike with tech set to very slow. Is this because the great players only play insane and don't want to take 7 months or does it actually favour the AI?

It definitely crossed my mind in video #2 I was thinking about it but this wasn't my 1st attempt I'd done a couple test runs and the AI had caught me out I didn't realize how early it can DoW, I was more used to turn 70 odd not 40 ish.  An early DoW is avoidable and a fail as far as I was concerned.  I didn't fancy starting again and it was too nice a start not to play out really, glad I didn't its been a ton of fun to play so far.  

It's just because no-one plays with slow research, because the setting doesn't work very well. The slowdown is pretty much nothing if you specialize planets.

Well, what kind of response do you want? #1 is a bug that humans are exploiting and I hope they fix, most of the rest are AI inefficiencies. They're working on it. An effective pseudointelligence will never be quick in the making.

It is my understanding that several folk at Stardock regularly troll the forums.  They have stated this several times.  There have been replies from various members of the crew on various threads.  They do not make a habit of acknowledging every thread or posting a bugfix schedule.  Considering we would rather have them working on things than getting caught up in forum discussion time sinks, many of us consider this a good decision.  So, given the responses we do see, both in postings and actual progress in the game, I have confidence that the devs do see this info and do respond by fixing or adjusting or redesigning things.

To turn the question around:  On what basis do you have doubts they are working on it? 

I ask because, if you are not demanding, you are doing something rather insistent, and somewhat less than complementary in implications.  The "good to great" line seems to imply that this was never Stardock's goal in the first place.  I understand where you would feel that rudeness has gotten reaction in the past, and share your frustration about that, but it is still no excuse for further rudeness over a game we already like.

I'm reasonably confident that stardock devs has seen this thread. Only way to not see this thread at all is to not even go to the forums at all in first place.

If you count the Betas as "patches" then Beta 6 (from memory) and patch 1.4 saw AI improvements, that's 2/12 "patches" hardly virtually every patch...  If any of the others had improvements they were incredibly minor, not noticeable at all imo.  GC3 could certainly benefit from Brad Wardell/Frogboy giving it some attention (he's working on Ashes not GC3?!?) for 1.6/the expansion, imo.

Macsen thanks for putting the Terran Resurgence on the Steam Workshop, I'll continue to watch the series until the end.  Out of interest I played a game with them this evening (Huge, abundant, 9 opponents), the only change I made was to an ability (to Intuitive).  The end result was a Turn 51 Alliance Victory while being dead last on the power charts.  Apart from what I've posted before, this time I focused on diplomacy research immediately and embassies (which I normally don't really bother with).  Appearing weak didn't help but the combined effect dwarfs the negative diplomatic modifiers of the custom race.  It's also worth noting I have Huge Hulls already thanks to an Artifacts Event, LOL!

https://www.dropbox.com/s/egh9bvvsp9m6qnx/TR%20Turn%2051%20Alliance%20Victory.GC3Sav?dl=0