It's very easily accomplished.  I regularly get 20-30 colonies by turn 30 and out-expand the AI.  

This is where you're simply wrong. The AI issues are not the result of imbalances in the game. The AI issues are a separate issue in themselves. Fixing your very minor list of complaints (which, as I've stated, really doesn't take very long at all) does not magically fix the fact that the scripted AI cannot actually react to anything. As I said, the list of triggers for the AI is just turn, tech age, victory condition, war, and personality. That's it. Anything you cannot detect using those 5 conditions is not something that the strategic AI can recognize and adjust to. The Diplomatic AI, for example, will play exactly the same strategy between turns 25 and 150 regardless of anything that has happened or will happen in the game (aside from being DoW'd). It does not react and it cannot react.

This has absolutely nothing to do with the fact you can't help but use exploits. We're posting in a thread where the OP didn't even know how to use the production wheel; I rather doubt he was cheesing the hell out of the Death Furnaces, performing the constructor flip or maximizing the Hyperion Shrinker. And how, precisely, is him doing any of that sort of thing supposed to have caused the AI not to research for 300-odd turns? He hasn't impacted on the AI's play through anything he's done there - that's just outright stupid play from the script. The one script which it was using in perpetuity for 300 turns, because no matter how much changed in the mean time it had no means of recognising it and adjusting.

I'm sorry, but that's simply absurd. I'm the one lost in the small details when you're attempting to pin the AI's failure to match the player on minor balance issues? The base game's AI scripting should not contain any logic to check on the events occurring in the game beyond turn number and if it's at war or not? The AI should use only 3 different sets of spending settings for the entire game? Frankly, the only way that you could be making this statement seriously is if you have not understood the extent of the AI's problems.

It's roughly akin to an FPS AI that does not actually check if it can see the player, but instead just fires unaimed shots at set time intervals, and which follows a set path with no reference to the actual level map - even if that path takes it off a cliff, or leads to it running eternally into a wall. Now, I'm arguing that that's a major problem. You're arguing that it's absolutely dandy, and the only problem with the game is that the rocket launcher is OP. I really, really wouldn't want to play an FPS you worked on

The AI won't declare on you to punish you for being greedy regardless of the difficulty; it's declarations are based entirely on diplomatic modifiers and aren't effected by difficulty at all. Because, as I've noted, it cannot actually tell the difference between Easy and Godlike. It follows the same script and just has a different number of ships, based on it's production. If it seems more belligerent on high difficulty levels, that's just because it's more likely to get the 'you are weak' negative diplo modifier. It's not reacting to your expansion, because it cannot react to your expansion; it cannot measure it, cannot change strategy based on it, and cannot even observe it. Statements like this one make it clear that you're arguing the AI is fine without knowing how it's scripted.Oh, and you're argument is also logically inconsistent. If tech pace is broken, how come this guy has taken nearly 400 turns to reach the 'brokenly powerful weapons'? How come the AI still hasn't? Which leads us neatly to...:

You're needing to come out with a hell of a lot of excuses here for why the 'fine' AI isn't performing fine. There's an obvious reason why: The AI isn't fine, for dozens of reasons. But you have already decided that the AI is ok and the big problems are based around balance; this means you now need to find a way to blame the AI's failure on the player (a player who doesn't actually understand how to play the game fully, let alone exploit it). So now this is because a player who was operating at about 33% of his economic potential managed to out-colonize the AI.... but really, is that not also indicative of a serious AI problem, if it can't keep up with the expansion of a player who is only spending about 15% of his industrial potential on colony ships? 

I'm not rejecting what you say because I'm obsessed with the AI. I'm rejecting what you say because it's logically inconsistent. You're claiming a causal effect that is at best tenuous, and at worst wild supposition.

Well, the game before my last was just like this setup. I was in the middle. The Drengin, Krynn, and Yor declared war on me pretty early like they always do since I play benevolent. After they got their butts kicked and lost a few planets, the fighting stopped...and never resumed again. The Altarians surrendered to the Thalian for no reason, giving me all of their stuff, and the Terrans surrendered to the Yor, also gifting me with all of there stuff. I ended up with a ZOC cutting the map in half diagonally. From then, it was just turn ending while all my planets developed. I ended up with more wealth than I knew what to do with and bought out the other factions. It was hilarious because against factions immune to culture flipping, I would buy all of a faction's planets save for their home world, sold them all of my tech to recoup losses (but I already had 999999bc), and then invade their home world to finish them off.

Europa Universalis 4 is CRUSHING its competitors on Meta-critic.  Perhaps I will switch to it.  

Paradox Interactive has two of the top 10 user-rated games on Metacritic right now with Cities: Skylines and Pillars of Eternity (which I have).  And EU 4 was another major hit back in 2013.  Hrrmmmmmmm.  

  I'm always thought that the AI was pretty effective after Their Finest Hour, especially strategically. Maybe it was more the way I played, but HoI3 TFH and EUIV were always significantly more challenging to me than GC2 TA.

  CK2 seems to be leaps and bounds beyond HoI3. Do you feel the same about CK2?

50 turns for taking over half the galaxy, which is effectively a win condition.  I don't bother continuing at that point because it would just be mop up.

Is there a way to create a multiplayer game where I play against the AI on a medium map with standard number of races, common/common and you can watch or something? I know you're busy, but just by watching me play you can learn a lot more than looking at the chumps on the metaverse.  This is a boast, but I can back it up.

I mean, I just played a multiplayer game with Echillion on a medium map and colonized 3/4's of the Galaxy by turn 30.  

I don't even know if it's exploitative. I'm just using what you give us. Thalan tech, Hives, and Gaia Vortex with Malevolence Invasion line level III (free 30 population AND 5 free transports) is pretty powerful on the smaller maps. The Benevolence Expansion line level III and the Benevolence research line Level V are both absolutely game changers on the larger ones.  

On the smaller maps the game effectively boils down to who can get Malevolence Level V on the Invasion line.  If you have 20 colonies, that ideology grants 20 very powerful early game ships. 

I can show you how to do this, and hopefully teach you a thing or two about the game-breaking things that are currently in your game.  

He posted his steam identity a while back and admitted that he doesn't finish games due to the boredom of the end game mopping up.  I doubt you'll get anything from his profile, but he might send you a savegame if you explain how.  I seem to recall him asking how to do that and not getting a direct answer.

Yes, he's incredibly annoying and easy to ignore, but you might find his data useful if you ask him nicely.  Just keep in mind that his style of play is unique and shouldn't be used as a representation of your normal gamer.  

Here's a screenie of his metaverse profile...

To be honest, the metaverse scores are mostly rubbish anyway; there's far too many 'won tech victory in 15 turns with 17 trillion population' games on there with a bajillion points, and hard-fought, fair games get piss-poor scores. I wouldn't trust any scores on there.

Marigoldran does generally beeline for the most OP thing he can see (which makes him great for finding and removing imbalanced stuff), but he does make it pretty clear what happens when the game mechanics don't offer any sort of mechanism to slow the player down. We can basically pick up all the toys if we want right now, and the decisions we're faced with are more which order we want to take them in rather than between powerful choices - the actual enforced choices (like tech specialization) are generally rather weak.

If you watch a few of Macsen's Let's Play vids, he's using basically similar tactics (and getting similar results). But really, all that using exploits does is speed up the inevitable; even those of us who don't use them generally find the AI empires crumbling once they get beyond about turn 150 or so in vanilla regardless of the difficulty level. It's not even from obviously bad scripting now (like, say, the diplo script in 1.02 which told diplomatic empires to spend just 1% of their production on ships between turn 30 and turn 150); that's largely been iron out. The problems are more:

1. there's not enough scripts. Most AI players will use at most 5 for the entire game; 4 if they don't go to war. Simply by increasing this to give you more granular control over what it's doing (the AI should probably not be doing the exact same script 40 turns after starting it ever), you can sharply increase the AI's play. If anything, just being able to tell it to spend 10 turns doing 100% military manu, then switch to 10 turns of 100% research, then do 10 turns of 100% econ and then repeating the cycle would improve it.

2. The scripts themselves need more triggers, so we can check extra stuff and make the AI use different scripts for different situations. It's possible to script the AI to build up for war - but mostly by guessing what turn in the game it ought to start doing so and hoping that we're right. But this is heavily dependent on game settings - the AI should begin preparing for war more or less immediately on a tiny map, but shouldn't really worry about ships til turn 100 or so on Insane. The scripts can't tell the difference. It will gimp itself on tiny if it's optimize for insane and visa versa. It will fail to make good use of bonuses on Godlike if it's at all capable of playing on normal, and will be hopeless on normal if it's designed to play well on Godlike.

3. The actual effects that we can call from the strategic AI scripts are fit for purpose. The governors, less so. The manner in which it picks which planetary governor to use is arcane and tbh not very effective; the queues it uses are poor, and then it fails to specialize anything anyway. I'd actually suggest that this area is one where scripting should be dumped completely and replaced with an actual AI subroutine, which evaluates hte stats of the colony and reacts to the changing circumstances; the AI cannot recognize that it's approval is sinking, for example. It just builds as many approval buildings as it's been told regardless of if it needs them or if it needs more. This means that the AI gaining approval relics or wonders is useless, because we can't bank on it having them and so it needs to act like it doesn't regardless. Such a subroutine would also be able to increase social manufacturing during building phases and shift the wheel on a planet-by-planet basis, just like the player; this alone would basically improve the AI immeasurably.

I hadn't seen your post - you added it while I was writing mine, so it's not responding to you. I agree with your points pretty much entirely, especially on the defensiveness thing.

I actually am having some trouble understanding this one. Brad has always said that the Gal Civ AI is one of the only ones which plays the game rather than just simulating playing it. But then, in threads like this one, he makes it clear that he regards the AI and the scripts as separate things. So... the AI doesn't actually play the game. The AI just plays with the toys provided by the scripts. If he claims that it actually plays the game, then that relies on the scripts being classed as AI; yet if he claims this thread doesn't show any AI issues, well, that relies on the scripts being separate from the AI. Either is a valid standpoint - players will largely lump everything a computer player does into 'AI'; programmers will largely separate them into scripted modules and AI modules - but you can't really mix and match which definition you're using like this.

If scripted = AI then you don't get to say 'the AI's doing a great job here but just doesn't have the resources' when the scripts are falling down, like in this thread. If scripted =/= AI, you don't get to say 'the AI plays the game properly, unlike in other games', because the majority of the stuff that the AI actually does in this game is the same as the majority of stuff handled by the AI in, say, Total War (which literally just handles unit movement - it gets units and buildings handed to it for free  at a rate based on the difficulty level and never does anything with the economy etc). 

For what it's worth, the AI part of the AI is already good. It doesn't need much tinkering with. If you give it the resources it needs, it does a fine job. If we're going to use the programmer's definitions, then, we can say GC3 has great AI, but it doesn't actually play the game, and the scripts which do are very weak. If we use the player's definition, where scripts are conflated, then GC3 has below-average AI, as despite the good handling of the units it gets it's incapable of managing it's economy beyond the opening 25-30 turns.

There's two basic options from here:

First, they could basically dump the scripts and just switch to Total War style giving the AI units when it needs them, based on difficulty. This is largely the same result that we'd get from every-increasing handicaps anyway - we just cut out the useless script economy management and hand it the units listed in the build queue scripts on a turn rate. This is an easy way to do things, and is fairly cheap. I think most of us would also agree it's a bad solution; it might make the AI more challenging, but many players will be turned off by the cheating AI.

The second option is to seriously engage with us over the flaws in the scripts, open up more triggers and effects, stop quibbling over the AI/script difference or pointing to 'exploits' when players highlight problems, and actually start making the thing play the game in the way most game AI's don't. This is a more expensive and difficult option by far. It requires adding entire AI modules to the game, which are not glamorous and despite most of us complaining about the AI we're not likely to get and stay excited about AI improvements happening - they very much fall into the realm of 'things players expect to happen for free'. But it's also likely to pay dividends. Many players are already putting the game back on the shelf because the AI offers no serious challenge. These players are not buying DLC, because getting 5 extra Mega events and some ship parts is not making the game any more worth playing.

And really, if making designs exportable for 3D-printing was an important enough priority to be included in 1.2 (to cater to the massive 0.000003% of the playerbase who own a 3D printer and just have to own a model of constructor Mk 38), then I don't think the difficulty of implementing scripting improvements is really a justification for not doing them.

I'm trying to remain fairly civil in these posts, as I do love the game and I see enormous potential in it, and I'm more than a little worried that the easiest thing for SD to do would be to drop forum bans on those of us speaking out about it if we put a foot wrong. SD largely only ban those who are outright toxic, but we need to be able to offer serious critique on these problems. They're not value judgements of the developers, or personal criticisms of their abilities - only idiots think that a bad bit of code means that the writer is a bad person - but constructive criticism of flaws that are increasingly becoming fairly obvious, and will become more so as the game continues to improve. Once the exploits thatsjaminei raises are ironed out; once the economy is actually doing something; once tech prices start to be raised to prevent the player from picking up everything in 1 turn, the AI's flaws become clear. We're already seeing thread after thread on this, and the AI is actually holding back other areas of play. 

I for one, would gladly pay for an expanded AI DLC, alnog the lines of Naselus' expanded scripts suggestion, but I honestly will not buy the Mega Events DLC until it is less than a buck, or some mod that I like requires it. Also, there is no way in hell I would've ever willingly paid for the ability to export to 3D printer, that was simply a terrible waste of time.