while you're at it: i'm pretty sure one reason why many people stomp even high level AI is how easy it is to get units the AI can't possibly deal with.

the  2 prime examples would be

- tamed bears with maul (mostly early game, but if used properly, there is no late game since you already won)

--> adjust maul to be less overpowered

- trained units with ridiculous initiative and fire damage buffs from essences (aura of grace & heart of fire)

 --> remove the 2 buffs, or tone them down to some sane level. as they are now, there's really no point investing time in tactical AI. with super buffed player units, there's no way the AI could possibly handle them, no matter how smart they move. and don't underestimate their effect. read some of the guides/AARs/watch some LP's and you'll realize that at least half (if not more) of the players who beat ridiculous/insane abuse those buffs to trivialize/bypass the whole tactical combat system.

The extreme faction differentiation makes conceptualizing how a great, straight-up AI solution would work difficult. I play so much to the strength of the specific faction I am playing, I would imagine it wouldn't be possible to really get competitive without designing for each faction (it will be hilarious if we ever see complaints like, "Kulan has allied with half the wilds, no fair"). I was amazed how well this worked in Gal Civ 2's expansions, but the challenge to do the same in this game seems far, far greater.

Now that we have access to some of the process in the game files, it will be really interesting to see what happens. This problem has always struck me as nearly impossible as a problem can be. I have always wondered how you approached it. If you ever describe your process in detail, or if you have and I have missed it, I would be fascinated (my desire is to do AI research). But I don't know if you consider it trade secrets or not, or if you consider it too technical for a wide audience. 

It seems almost impossible, or at least extremely impractical, to manipulate the weights fruitfully without knowing how they are implemented. I am guessing it might be possible for someone with a lot of time, but I doubt they will come up with any major improvement without expending an unbelievable effort compared to what it would take if you could at least know how they are used (much less tinker/add/rework them to a personal approach to the problem). I am almost certain by the time a problem is solved via playing with the weights, it will no longer exist in the game anyway. 

I could be wrong though, but I have tasted how hard even simple problems in producing intelligent behavior can be to solve, and this isn't simple.

In one thread I tried to give someone advice on how to make the AI quit declaring war. This isn't trying to improve the AI, it is just trying to eliminate a behavior. There were weights called something like warPriority in that file, one for each stage of the game, which had some small positive value, and I thought setting them to 0 would have the effect of the AI not declaring war. It seemed reasonable, but it apparently didn't work that way. That is the extra problem, you would have to tinker with something, see if your hypothesis on how it even works is accurate, and then fine-tune it if you ever got to the point you understand what it does (if it happened to do something relevant). That extra step of removal from an already difficult problem seems like a nightmare.

I would be fascinated if someone actually made significant improvements this way, but I am dissuaded from even trying it because of the above difficulty. It seems a little like making blindfolded swings at pitches to me.

Any update on the AI love? I would love to know what's going on!

The AI figured out how to cast strategic spells in real life.  Now our only hope for salvation lies in the hands of the one man brave enough to lead us: John Connor.  (Fortunately, Dereck never designed a "time travel" spell for the game, so I think we're gonna be fine.)

My take on the AI is this: As long as the tactical AI is not able to kill player units, all strategic improvements will not do much. Just look at DAS24680 ongoing letsplay on youtube. He was able to eradicate his biggest rival faction in just a few turns without loosing any unit. The AI needs to focus on killing the weakest player units. If it can not decimate the player armies, bigger numbers on the strategic level won't matter. The player will just get stronger. 

EDIT: Just look at the battle in DAS latest episode starting around 22 min in (http://www.youtube.com/watch?v=XhkeBZLPhsU&html5=1). The AI player is superior in terms of raw power. However, instead of going  for the healer and the mages in the back it wastes its strength by attacking the hardest to kill units in the frontline. If the AI had gone for the softer targets, the battle would have seen a different outcome. 

The strategic AI can benefit from some aspects of the tactical. For instance, if it were to build more effective troops then the tactical AI would struggle less. Even better, if the strategic AI could analyze whether or not its troops would be effective.

Could you have players opt in to upload troop effectiveness stats for some en mass machine learning?

I haven't had time to play LH in a while now so haven't posted much on the forums recently (that time thing again!). But I wanted to just pop in and point out that this sort of problem is exactly why you still have people like legaceez creating his (poorly written) rant thread. Like most of the people on the board I am pleased that LH is more polished than any of the previous iterations but the fact remains it is still remarkably buggy and full of exploits.

Before you say I'm being unreasonable, let me ask you a question. If you played Civ 4 or 5, one or two expansions after first release what do you think the chances would be that when you reload a game it would be in a substantially different state from when it was saved? OK, it was a rhetorical question because I'm going to answer it for you, the answer is negligible. How about the chances of there being such an easy and overpowered exploit as queuing pioneers to completely get around the city population limit and ensure you always get maximum growth? Again, negligible. Yet both of these are known problems that have existed for ages in LH/FE.

There are plenty more, mostly less critical, but in total they are still rather frustrating. I would really like Stardock to focus on polishing the game more. Maybe they will next?

I think Civ IV BTS is clearly a better designed strategy game than Civ V but I would still rate Civ V in its current form (ie after the first expansion) as a well above average 4X game.

In any case my point was more about the polish of the games. As you say hopefully there will be further patches fixing more bugs and exploits.

I would like to see the AI polished more and it's kind of sad that it has to be an after action item to clean up simple stuff like the AI focusing fire for swarm attacks or prioritizing obvious targets.  I'm not that great of a player and immediately grasped how ludicrously overpowered the swarm mechanic is - the AI should be locked in to maximizing it like crazy (otherwise it gives the player a ridiculous advantage).  You would think that even the most basic testing/playing would reveal the completely obvious things (and if *I* figure it out it must be pretty obvious).

Target priority is a tough one because players are going to have lots of different opinions on what should be taken out first.  Some targets may seem obvious under certain circumstances but what's obvious to a human could involve a boatload of "thinking" for an AI.

Ultimately the tactical AI is decent.  I would like it to be stronger but you have to realize that there are a bazillion factors that a human processes for every battle - factors both pertinent to the current battle and the big picture.  The AI does pretty well for the most part considering the complexity.

What does this have to do with a discussion about AI and bugs?  But I'll bite...

I love Civ 5 and consider it best in series but there are times when mid/late game drags and/or I don't finish a game that's a sure win due the tedium of finishing long drawn out games that are sure wins.  I'm hoping the xpac spices up late game a bit more.

However.

I don't think LH is any different and most 4X suffer the same thing.  Mid/late game is often less exciting and/or the turns get longer and longer as slug thru wrapping up games.  I find LH tedious in mid/late game because:

- tech tree seems tiny, once you get rolling there's little left, it removes some spice when have nothing to look forward to for tech.  Picking the next 10% repeatable upgrade is pretty much button pushing and is unremarkable regardless.

- I love the tactical combat but the autoresolve is so iffy that I often feel that I need to manually do what should be trivial battles at times, multiply buy having several active armies needing to resolve 1 or more battles in a turn and you get some long turns.  Manually slugging thru trivial tactical is just button pushing (and holding down shift to speed it up).  It requires next to no though for me to get thru such battles and I don't often use magic or any hero abilities for them, but too often if I leave them to autoresolve my armies take incomprehensible amounts of damage.

- road system is ultra wonky, not easy to get a "road builder" (I don't really wanna give a hero that ability), maps are huge, takes forever to get armies to places they can be doing something.  Slowing to a crawl on roads in enemy territory is necessary but brutal.  Setting long paths for units is generally not a great idea since you have no idea how/where they might path and/or what random stuff they may encounter in the fog, controlling every single unit/army for every single move ends up being a LOT of button pushing.

- early city management is interesting.  Once your core cities are "done" and you're conquering, it's a little less interesting to clean up the AI cities you capture, especially the runts they tend to eventually spawn at former outposts (I'm guessing they upgrade the land via spell then create a city on the usually crappy resulting plots).  You can raze the runts but if there's any space to build a city and a spot to build it on, then as in most games of this genre, if an AI can build a crappy city, it will build that crappy city, no matter what.  So, mid/late game city management is less thrilling.

- one reason I sometimes don't finish civ 5 games is because I just don't want to slug out the latter part of games - I personally usually find the earlier stages of the game more interesting, the exploring, the REX, trying to grab some early wonders, fighting early wars to establish a solid position, early worker actions that have more impact, overcoming the threat of barbarians, etc.  I feel pretty much the same way in elemental.  The world is less exciting once you clear most of the monsters, have every reasonable city plot settled, and are to the point where your armies are ludicrously powerful (either by xp, gear/unit size, or both).

What I usually say when people discuss this particular topic is, so what?  It's how the genre rolls.  Like it and play, or don't.  It's not something that can be easily cured because the very nature of these games is to usually create a huge empire, or have a long term goal, and things like that take time.

But it's kind of silly to bag on Civ 5 for having mid/late game tedium when it's something you can almost always find in the genre, and when LH is surely not lacking for it either.

Wait till 1.2

Things are about to get harder.

Not to get too far afield, but I am not bagging on Civ 5 for having mid/late game tedium (mopping up is a plague to the 4x genre).  My issue is that Civ V is tedium from the word go.  IMO, they boiled out all the decision making and refined the game to the point that there's very little left.  Polish is a double-edged sword; Civ V is living proof that you can polish a game until there's nothing left.

The way the game is structured, you're supposed to finish clearing out the monsters just about when you are ready to wage war with the AI.  Unfortunately, the AI kinda sucks at preparing for and executing a war.  It's terrible at recognizing that things have changed and adjusting to new circumstances.  Which means that if your stack can beat the AI's stack, you're going to conquer their entire kingdom.  After you've cleared out the monsters, there's really nothing left to do.

I am not so concerned about whether they fix the Autosave bug or pathing around rivers or make the hotkeys consistent (all those being worthwhile concerns, mind you).  To me the far more important question is whether Brad & Stardock can get the AI sorted out so that it can offer an interesting mid/late game challenge; so that conquering the whole continent requires diplomacy, strategic and tactical skill, and some creativity.  Of course, I've been asking that question since FE's beta and I am really not sure why I still expect it to happen.  GalCiv II, I guess.