Re City defeners - that's a very good point.  I spent some time today re-designing units for a few of the factions, they're more axe oriented (and definitely more spear) oriented.  

I think hammers got the short-shift in this beta.

Right now, the AI doesn't design its units. It uses what's in there. From playing this weekend, I've realized that the units available to the AI are just truly awful. I'm not sure if the last FE updates went into LH (lots of player made units were made availabel to the AI). I've asked Kael about that. 

It's going to be really tough to optimize the AI's unit placement in tactical. One of the reasons why turns are so fast is that the AI doesn't have to put much CPU time into where it puts its units in battle but instead focuses on what units it wants to target. Will have to think about that.

Thanks for the feedback!

I think this is the right time to rethink the city defenders and city fights!

I think that possibly the faction power calculations are not weighting things very well, and if the AI is trying to have a high faction power rather than an army that wins fights, it is going to go in the same direction as the bias.

I would quite like to see a better troop power calculation, especially if this resulted in the AI fielding more effective units.

Are their weightings for attack power, hp, movement, armour (hopefully not multiplied by number of figures for armour)? At the moment, the power calculations are quite opaque to me. It would be interesting to see the current weightings for such calculations.

Two suggestions and examples in the hope of giving constructive feedback:

How about as someone else mentioned trying to get the AI to plan an attack before declaring war and thereby surprising and threathening a player the moment the ai declares and perhaps now and then catching a player unawares? For me just the uncertanity of having a friendly ai player mobilizing units near my borders not knowing but suspecting an attack is fun. Is this hard to do?

I am not sure if this is possible or not but as with the AI learning to use designed units let the ai learn the spells/abilities/traits a player uses and either use them to or use counters. For example player x generally uses fireball and he currently has a mage with the spell fireball then we need something to counter that or use it ourselves. I realize this probably is very complex but perhaps it could be done on some basic level as in last battle fireball/range/axes did this amount of damage and so fireprotection etc is needed. Impossible or possible in some sort of fashion?

I realize I am probably bad at giving examples and solutions but this game could really be one of the best strategy games ever especially now that LH actually made tactical battles more interesting.

I think it'll be worth the coding effort to give the AI the concept of a "strike force" which would be an initial declare-war force.  It's probably 60+ engineering hours but given we're not being released until May, I think there's time.

GalCiv had one. But I don't have anything like that for FE.  I am groveling with Kael to do this using Steamworks. 

Im also confused as to why the AI mobs, not sovs forces, dont seem to be attacking my cities, or the other sovs cities or whatnot.  They seem to just wander around aimlessly and...I dont know what they do.  I can understand if they are thinking "Hey I cant prolly take that city", but from what Ive seen they could do some dmg to them and to me.  Also, not every AI mob, should prolly use tactics.  Many undead types or even animal types may just charge in because...well...thats what they do.  The Ogre types, trolls, and such prolly dont think it matters if I have 100 troops or whatnot...they are just coming.  I see many AI cities and they are literally surrounded by mobs that do nothing...they dont even go in to destroy the sovs mines and stuff.   This is just what Ive noticed right off the bat.   Though I have been enjoying the game more since all this new stuff.

One last question.  Its on the Govenor trait class.  Are you planning on having someway for these guys to lvl while in cities then?  I know theres an adventurer guilds you can build, but is that how you think you want it to go?  I mean having someone who is better at building and maintain cities, should be in those cities...not running around in armies trying to gain lvls and xps.  Administrators maybe well be gifted in running civilian side of life, but im guessing that they may not be so good running the military side.  Thanks and keep up the great...yes GREAT work!

Kael, listen to Brad!  You shouldn't have to grovel.

 A good challenging AI gives the player a real sense of achievement and satisfaction, sorely lacking from a lot of games these days.  GalCiv2 AI was a work of art.

 Also Brad, check out what Dan is doing with Stardrive, there's a big discussion on the forum there http://www.stardrivegame.com/forum/viewtopic.php?f=1&t=2578 about the AI using areas of operations (AO's) for fleets (which is like a stack?).  Dan is trying to create offensive AO's and defensive AO's as well as a central pool of units the AI can use as it sees fit.  It will continue to create units as needed and added into each of these three pools if I understand it correctly.

 This might help as the defensive type unit designs can be tagged ie. spear, shieldmen and put into the defensive AO and the more offensive unit designs tagged and put into the offensive force.  Also units that are flexible or necessary like sov, champs, archers can be put into both and the pool.  The pool is basically a large stack the AI can do what it pleases with, scout, attack, defend, supplement the offensive or defensive AO.  

The offensive AO and defensive AO armies can be coded specifically to excel in those areas, as per the OP when defending cities etc

Not sure if this is something you've already done, but I think Dan is a pretty smart guy and the AI of Stardrive is quite impressive already.  I think you and Dan could have a good discussion about the concepts in AI design and although one is space and real time and the other is TBS and fantasy I think sharing your knowledge with each other will make both games have a better AI.  I own both by the way.

Tactical battles

problem1:  I'm fighting 3 spider-things.  Spider 1 successfully webs one of my units.  Spider 2 attempts to web the same, already webbed unit.  Spider 3 does the same.

Solution: check if target is already affected by the upcoming action.  (yes, I understand not wanting to slow down things by adding checks, but this is a pretty big deal -- seeing spider after spider try to web an already webbed unit looks bad)

problem2: spiders attempt to web ranged units that would likely hold back (so being webbed is least likely to benefit the AI), when there's plenty of viable melee type units unwebbed.

Solution: for webs or similar movement restricting attacks, prioritize melee over ranged weapon and spell casters.

-------------------------

I'm liking the changes in LH.  For example:

-the new swarm feature makes flanking/etc. more effective

-I initiated battle near a river and there was a river on the tactical map.  Tactical maps now sorta reflect the overland terrain.  That's a big deal! 
Choosing where to fight to achieve some advantage is basic strategy, and with a bit of refining (such as having tactical map terrain affect combat such as trees blocking ranged attacks, cover and concealment adjustments to defense, etc.) this would really add a lot to this strategy game.

-SPOILER... (don't look down or else you'll see something about a quest)

I had a quest pop up, something about bad stuff in dreams.  Hmmmm...  Made a mental note but that was it.  A few turns later I notice a few wandering skeletons in the area.  A few turns later I notice a crapload of wandering skeletons swarming my area.  I don't know if they'll keep appearing until I finish the quest or what, but it's a darn interesting/surprising event.  Good job to whoever thought it up!

Because of the way armor works, double attack means more than double damage against armored units. Crushing Blow is very good against high armor units.

Bottom Line: You seriously underestimate Crushing Blow. It is much better than what you make of it.

Large map with 8 AI. Challenging level.

The AI wasn't picking up loot. Undefended loot!

Increased the monster difficulty by one level in this game too. The monster groups aren't attacking the AI but then they also can't take out any of the Strong monster groups.

Does the AI prioritize research into increasing city research and production? All of the AI factions have less than half my growth. Seem to be way ahead in research.

I came here to post about the same thing.  Very annoying!  The spiders also love to beguile ranged units.  With the new 'close combat' layouts to most battles, the spiders would generally be better off just attacking.  The AI does seem to favor spells too much.  I'll have closed with an enemy hero, and be beating the... er... tar out of him, and they're still fiddling around hasting themselves.

Tactical battles

Problem3: I had 3 units -- 1 melee unit, 1 ranged unit, and 1 champ.  Spiders webbed the melee who had no enemies adjacent, so my melee unit was 'out of the battle' while web persisted and could have easily been bypassed.  Instead, enemy spider1 bypasses my champ (that's already been attacked by enemy spider2) and moves up to and attacks my webbed melee unit.  My webbed unit should have been ignored and enemy spider1 instead should have ganged up on my champ, or perhaps closed with my ranged unit.

Solution: Not a quick and easy fix as this would require several checks.  It appears enemy units choose their target very simply (weakest unit?), without regard for concepts such as 'is another enemy unit attacking the target'; or 'is the target out-of-the-battle-temporarily and if so then move to another target'; or 'is there an ally attacking this target' to preferentially combine attacks on a single target.

Problem4: tactical map had 2 impassable areas, forming a 1 tile wide gap between them.  I placed 2 strong melee units next to said gap so my 2 units could both hit the 1 tile gap, put my ranged units behind them, then waited.  Enemy units didn't move up to engage my 2 strong melee units, instead they moved the long way around and were picked off by ranged units without ever having a chance of reaching me.

Solution: Again, it appears the enemy chooses its target simply, without regard for considerations such as 'are there strong enemy ranged units that will kill me before I can get close so it's better to switch targets to one I can reach before becoming a flaming pincushion'.  Perhaps at battles beginning assess my units and if there's 'strong' ranged units then set a flag that nearer targets are preferential to distant targets.  Each turn see if the 'strong' ranged units are still alive and if so keep said flag, otherwise toggle flag off.