A trip through the fantasy worlds I enjoy

 

  We did a lot of work on the AI this week (and by we, I mean Brad) and made some significant changes.  I even had to ask Brad to tone down the aggressiveness based on a game where I got dragged down and beat like the Goonie kids, if they'd accidently wandered into the Hellraiser movies*.

  But the trouble with testing the AI in a game that contains as many moving parts as FE is that it can be very situational.  It also has a lot to do with balance.  Some abilites are to good, some are to weak and a human is always better a figuring out how to exploit them than an AI player.  Some items may randomly swing the game ot much (if you get a berserkers axe in the first 20 turns, etc).

  I can only play a limited amount of games per week, but if I can get feedback from you folks we can start to learn form that experience and figure out what is keeping the AI from being competitive.  I would love to get the following from anyone who is interested.

 

1. General feedback on if you felt the monsters were a threat (were they too strong, too weak, or about right).

2. General feedback on if you felt the AI players were a threat (were they too strong, too weak, or about right).

3. What difficulty did you play on?

4. What faction/sovereign did you play (if you made a custom sovereigns, what traits did you pick)?

5. If you think the AI was too easy, what did you do to lead to your success (did you outfit your soveriegn in leather armor, find a good sword and then single handidly kill everyone, did you start producing an endless stream of spearmen and destroy with your armies, did you learn fireball and use it to destroy the world)?

6. Attach a save game at the point where you believe you have the game beat so we can check it out.  A save at the tipping point, where you believe you have the game won and have to play it out is more useful than one where the opponents are all crushed because it allows us to see exactly what is going on when the AI loses.  If you get to a point where yo don't feel a threat from the AI anymore thats a good point to get a save.

 

Thanks,

Derek

 

* If you weren't born in an era where killing a duck with a three pixel long sword wasn't awesome then please replace the above mixed analogy with "beat like Justin Bieber, if he'd accidently wandered into one of the Saw movies."


Comments (Page 9)
12 PagesFirst 7 8 9 10 11  Last
on May 07, 2012

I've been playing this a bunch over the weekend- it is getting better and better.

 

1) Monsters feel pretty close to just right to me.  They definitely beat me in a few games.  The only complaint I have is that it seems like sometimes they target the player unfairly over the AI.  No idea why that is, and the possibility is that it's just confirmation bias, but I've seen monsters go after me when there are weaker AI units nearby.

 

2) As for the enemy AI- it feels like Brad's AI, it seems to build well, but doesn't fight that well.  I've seen some herp derp behavior, too many pioneers, trying to walk weak units past towns I just conquered, non-optimal use of magic.  I do think the AI is getting there strategically, it's the combat AI that needs the most work right now.  That said, combat in a fantasy game is twice as hard as a space game, so Brad has his work cut out for him.

 

- One definite: the AI needs to recalculate its movement paths whenever a city is captured: I think that would stop some of the AI's bad choices.

 

3) I play on Challenging.  I want the AI to fight me on even terms without cheating either way.  Challenging world difficulty also, unsure if that's the desired setting or not.

 

4) I usually play as Irane or Ceresa.

 

5) Ceresa's spells are gamebreakers.  If you level up Ceresa she can one-shot armies with Dirge of Ceresa and monsters with Soulburning.  All she needs is a minimal meatshield.

 

The whole surround and zerg rush, I notice it from the AI as well, and wish other strategies were more viable for the AI.  AI would be much more interesting if it could be effectively unpredictable.

 

My verdict right now: this game can be great, it's good enough now that I'd be satisfied if this was the release product, but the greatness is going to depend on two things: game-to-game variety (not enough right now, it's about by my esitmates 1/2 of FFH at best), and the AI improving.

 

I'm still laughing that my Spearmaidens are popping up in other folks games.

 

on May 07, 2012

I've commented on specifics as per the request and my comments stand, however, i've noticed 2 or 3 things that are worth mentioning.

- The AI's do not band together to take down someone leading the field by a country mile. For example, I have 600, the next 2 have 300 each so what do they do, they make war on each other.

- At about level 7/8 my Sov can kill any monster in any stack. Him and his hero can take out the giant elemental guardians, pyre of man etc from the wasteland zones.  Either creatures need to steadily get stronger or get stronger at certain key points in the game. I'd also make their lairs muchy tougher as otherwise they get cleared out too easily later on.

- AI does not research armour quickly enough. Ever.

 

on May 07, 2012

I played a lot this weekend.  I played most the games on challenging with dense monsters.  I did try a few on hard and with normal monsters but didn't like it as much.  This version is much, much better.  I think one challenge is that game configuration makes a huge difference in game difficulty.  I've found:

 - Small maps are easier, AI can't really handle a killer stack (usually heroes in my games)  early game and get run over.

 - playing with around 6 opponents on a large map with challenging settings gets very interesting.  I could dominate through 100 to 150 turns by building a killer stack with my heroes.  Attacking an AI early is also good for expanding.  but since I wasn't building many troops, it became a big tug of war trading cities and outposts with the AI that had freedom to grow.

- I think the hero skills need another pass.  There are still craptastic upgrades like finesse and great ones like potenital I.  I don't mind having some rare cool ones, but there should be no crappy ones that will be avoided.

- I found food too hard to come by.  I guess if your intent was that players would have to get lucky or make serious investments to get level 3 or higher cities, then you succeeded.  I felt I had to invest in agriculture in every game.  I didn't see any spells to help my food supply, I think that would help.

- Strength of units vrs. heroes seemed right.  A good hero could dominate, but I found it challenging to have more that 1 or two really good heroes.

- I would like outposts to have defensive upgrades.  It would be perfect for a tech upgrade

- I think cities are too hard to defend, it might be good to allow earlier defenses.  Maybe allowing upgrades for militia and city walls that were not so difficult to get would help.  Increased city defenses would help stop early rushes and so many nasty surprises from weak wandering monsters.  It is just really hard to make any units if you don't have many cities.  I had quite a few games where I was totally constrained for building new cities.  I could only get one or two before I hit high level creatures or unbuilldable land.

-  The spell that creates outposts feels overpowered.  I expanded so fast that I was able to keep casting it rather than build any pioneers for outposts.

- City management is better, but I feel really constrained.  I think with slow tech development and the various branches, it is really hard to get any city specialization.  All the cities are pretty much the same.  I think you need a whole lot more base upgrades, but make them more expensive.  This way a player has to choose where each city invests.  Only place I saw this in the current build was around army upgrades.  I usually only upgraded once city.

- I think the dense vrs normal monster setting could use some work.  Having more monsters is good for upgrading heroes if they aren't too difficult.  You may need another setting for monster difficulty.   This isn't a big thing.  Dense feels best to me right now.

- Some of the content did get repetitive, but it was great when I got new quests or saw something new.  This is easy to fix and I'm sure a much lower priority than getting the game solid.

- I'm pretty sure there is a memory leak or some other bug that crashes the game after being on for a long time.  I played for a while on Friday and left it on the pc until evening.  It was crashed by the time I came back to it.  I had some other games get slow or crash after having them up for 6-9 hours.  Overall, stability was good.

 

Great progress!!!  This was the first time that I've really had fun with the game.  Having the ability to tune the challenge to play ability is a key success factor for this game.  Too easy or too hard will make folks quit playing.  As long as players can tune the game settings to get a good challenge, they'll want to play.

 

on May 07, 2012

Finished a game on hard yesterday.  bloody hell, 2x AI Sov health!!  This was actually good, because I wasn't able to opurtunistic and do a quick Sov kill followed up by beeline to their capital (which keeps the Soc out of action due to continuously dieing).

AI expanded about right.  Built good armies (hell a lot of them!!) It was choosing good places to build its cities.  Still doesn't design it's Sov skill sets very well.  The reason I ended up winning is due to a poor strategic choice by Verga (my next door nabour, I was Pariden, so not a good nabour to have...).  Verga declared war on me, but Verga was no where near my border, so my champ with it little army and my Sov with its little army were able to slowly push forward.  I lost a city due to a bug (I need to post this next), but took it back and forced Verga to surrender (no cities left) and that's how I found he was no where near me.

If Verga was near my border, I was in big doggy doo doo.

I'm about to post some tis for the strategic AI and the tactical AI that should help.

PS: if I wasn't playing Pariden, this would have been a harder game to win due to no insta outpost spell.....

EDIT: I won the game on hard due to a very close but decisive battle.  Tactical AI did all the right things, but I burned about 200 mana (path of mage + adept cloak) which got me over the line.  2 high defense tanks protected my Sov (who could have taken a beating anyway) and the tanks kept getting healed when they were 1/4 health, the rest of the spells I cast were damage spells and haste and stoneskin.

on May 07, 2012

1. Monsters were a threat.  Wandering assassin beasts and umbers in particular would take out a garrisoned city creating no go zones until I was strong enough to deal with them.  I liked this but did not like the fact the AI fractions could build in these areas without risk of attack, even next to a powerful lair that would immediately activate and kill my stack should I dare to take one of those cities.

2. Early on the AI was more concerned with city spamming, grabbing goodies and levelling their sovereign, so didn't threaten me.  Later on they were a threat in that I had to fully garrison my cities and prepare a number of killer stacks to deal with reprisals before going to war.  The AI knew where to attack me and send stacks to cause trouble behind my lines so I had to be ready for this.

3. Normal

4. Custom sovereign, summoner, extra mana and research, life and fire magic, lucky and initiative

5. There was no point in building my own cities, they get killed by wandering mobs and by the time I had cleared and built troops to garrison with that AI had claimed all the good building spots.  So I concentrated on troops, shards and mines and took AI cities, one fraction at a time.  Sovereign was a good healer/ ranged magic paired with a tank hero (leather clad + shield) to soak damage and get healed.  Summoned wargs (1 to stack plus 1 in battle) for quick damage.  Tank troops as per hero.  Stone skin and regeneration on front line troops.  Picked up obscuring fog asap.  Keeping troops alive with heals and minimising damage was key and they just got stronger as they leveled.  When 3 AI's left, two of which were at war I took out the third to give me the game.  The other 2 AI's only wanted to trade with me and continue their war against each other when I was the biggest threat.  Got bored after that so went for spell of summoning.

6. Tipping point was when I took out the 3rd AI's sovereign.  I let his stack attack and capture my heavily defended city weakening his stack, all while my sovereign was nearby and then retook the city easily killing the AI sovereign. AI sent other stacks but all along the same road which I only had to block with a killing stack.  Once dealt with I just had to mop up the AI cities and keep their sovereign down.  I used  a second hero stack with cavalry to chase down raiders and recapture lost cities.  I garrisoned my capital well with an administrator hero which could come out to deal with raiders.  The AI did a good job of trying to get raiders behind my lines to destroy resources and attack cities in my rear and I had to prepare for this before going to war.

Comments: This isn't a game for builder/colonisers, you can't compete with the AI due to monster one way hostility towards the player.  This will alienate players who prefer that sort of game.  Perhaps the difficultly selector could be changed to select between a builder or a war like game or a bit of both to suit all?  AI research roads very early showing me the way to their next city and giving me a highway to conquer along. 

 

on May 07, 2012

 

1. God, yes. But a good threat! So much fun to play with the monsters. I would like the bandits to spawn a little less frequently (I had 3 camps blocking all 3 ways from my city to the rest of the world) but they were dealt with rather quickly, so it wouldn't be a major issue either way.

2. Fairly low difficulty on a huge world, so I only met Gilden and Pariden. Pariden had the annoying habit of running around without a city for the first 100 or so turns, so they are mostly out of it now. Gilden has been expanding quickly and has already begun to wage war in Curgen's Tomb, much more effectively than I.

3. Easy

4. Amaca - traits emphasized were Research (the one that gives +10% research) and Tarthian longbows.

5. Early on, I got a good champion and found an item that lets me cast Blindness on everyone, while she has slow. We started near 3 shards and quickly found 2 more just around a nearby mountain, and my Sovereign had the mana-generating bonus, so we had a ton of mana to play with. Enchantments were thrown around almost constantly, my main army had a ton of buffs, my main city had a ton of buffs. It was around then that I managed to generate 30 gold per turn on 0% taxation.

6. Still working on it <:

on May 07, 2012

Hi

Good work, i think it is coming along well. I enjoyed this game although it is still a bit buggy (but it is beta)

1. The monster threat was hard (i think they were a little too aggressive) but do able. The spawning was good, but did make it tough at the beginning, since i had a couple of Drakes that were spawning stuff, i couldn't kill the Drakes and the small Drakes took out my second city a couple of times. The real problem was that losing cities early makes the game very hard when the nine tiles around the city become wasteland. I would strongly suggest that if a city is lost to monsters then the tiles do not get destroyed.

2. I though the AI was quite good. It is better than it was and generally it did alright. But as you can see from the saves, there is no way i should have won. Altar was on 600 power and had 13 settlements. I was on 175 power and had three (although only 2 were any good). and yet i went through Altar very easily with two stacks. Mostly this was because hardly any of his troops had armor, and the computer doesn't seem to get the importance of leveling troops. ( couple of the guys i had, had 275 hit points, which is a lot).

3. Challenging

4. Gilden, medium map, 6 players: me, Alter and the rest were Empire.

5. I did notice Magnars sov get stuck for quite a while in one place. Altar had a big stack of Pioneers sat there, but generally the AI did OK. I think the default for troops should be that they can upgrade weapons and armor. Also Altar seemed pretty lucky with the monsters, quite few did not attack him. But they all went after me, especially the big winter guy who decided to destroy most of the cities that i took from Altar. That did feel a little unfair (to say the least!!) and explains why so much of the map at the end is wasteland.

https://www.dropbox.com/s/gy4sfledzyx60g6/Is%20it%20Won.EleSav

https://www.dropbox.com/s/c21dfsicc364gpc/Yes%20its%20won.EleSav

on May 07, 2012


Ok, I played another game on Challenging, large world as Kraxis. 

Great gameE  mind you, I got crushed by the Gilden in the end, but it was a legitimate crushing.  It didn't feel cheesed, just out played. 

The main issue to start was a nasty hunter spider spawn that was south of my starting city.  I dropped a second city...delayed because of the roaming spider groups that came from that spawn, and the city grew fast and was a strong second city.  however, I have to stay near it and continually defend, and hordes of spiders would come up and attack my city every so often.   This, ultimately, gimped me.

However, as the game progressed, two factions, Tarth and Gilden, started roaming around me, settling and then declared war.  My mistake?  I turtled up in my cities, which allowed them both to settle all around me.  Finally, , and it was beautiful, Gilden came with two stacks to my capital, one with his two heroes and three spearment.  I beat them with 2 HP left on my hero.  But I knew it already.  The other stack was waiting right next to my city and immediately attacked.  5 stacks of spearmen.   They then immediately moved down my roads to my southern town and took it.

 

I should have attacked their small wandering units, especially the pioneers.  I should have declared war.  And I love it!  Also, I should have left the southern spider laid area alone, and gone north and built my second city there.  That is where Gilden settled and crushed me from.

edit: I used to own pretty easily on Challenging, and this is two games I have been crushed in, the first by a wandering umberdroth at turn 70ish, and now, by a pain in the rear monster pit, and some scheming AI Sovs.  It can be sometimes frustrating, but the monsters seem right.  It feels like a dangeerous world, and I need to start thinking a lot more about what I attack, when, do I stay and defend, how quickly do I build an army?  Good work!

on May 07, 2012

Ausland

 
AI keeps sending units through my territory.  I purposely tried to bottleneck a section of land, and every other AI thinks they have the right to walk right through my area of control.  Not only is it frustrating to keep telling them to "get off me land", but sometimes they refuse to even talk to you so you CAN say this.  There should be a way to say stay away from me or else it is an act of war.  Period.

I agree with you 100%.  This shouldn't be a game of "lets see if the player notices that someone is sneaking across his land this turn!".  AI should have to ask permission to cross.  The only other thing I think you can do is fill a bottlekneck with cheap slaves (as magnar). 

on May 07, 2012

Played on Hard, with a medium map, dense monsters, magic and resources, fast pacing.

 

Well, I actually won a game. That makes a change - normally I stop playing long before the end due to there not being any challenge.

Accidentally won via the quest of mastery - there were some dragons to kill, and my heroes did just fine, although it would have been an easier fight had I had mana available for it.

 

I was playing Pariden with a channeler with beast mastery. Some of the key stages seem to be about taming larger and larger spiders - getting one with 140~ hp seemed very nice, and said spider survived all game.

The ability to create outposts with a spell was very key to this game - there were stages where I had 5+ of each shard.

At some point, Altar declared war on me, and destroyed a lot of my outposts, although I was taking one to two cities from him per turn. This process was interrupted by accidentally winning with the quest victory.

on May 07, 2012

Playing the game on hardest (ridicules) with all the options being on norm.

Unsurprisingly, I failed to beat the game. This was done due to two factors:

1) Player units (computer) grant a very small amount of exp. Beating on an earth elemental is easier, and probably give a better exp reward than a small army of mediocre units. I am not sure if it's intended, but the whole feeling is that it's way harder to level champions compared to earlier versions.

2) Computer spams units like there's no tomorrow... Just because I crushed 5 stacks of 5 units each, doesn't mean the 10th one won't start killing my units. This is probably due to difficulty, and the reason I didn't test the easier settings.

 

Also, a couple of notes which were strange, and probably are bugs:

Certain monsters (mostly drakes) ignore my settlements... until 50+ turns when they decide to attack them, for no reason at all. A bit more consistency for all the monsters will be nice.

Unit stacking- It's quiet annoying killing a stack of monsters, only to find there are more stacks below it. I've had to attack the same tile 3 times in order to clean it from monsters once.

Turns not ending- A few times I pressed the end turn (or it was autopressed? not sure) and the game got stuck. While everything was viable (I could cast spells on enemy units), I couldn't end my turn (grayed out). Saving and loading fixed it, though.

on May 07, 2012

I have noticed that when the end turn doesnt work, its usually because some background AI process is running. Wait half a min and suddenly you can move on.

on May 07, 2012

The Ai is kicking my ass, early game on many of the games I have played. Most recent playing Altir , war was declared by 5 factions on me before 100 turns on EASY, you read that right EASY. I like a good challenge but If I wanted to have no fun and feel pain I would just smash my hand with a hammer.  Other wise I think things are coming along nicely !!

 

 

I think I found my problem, Autoplay still does not make the best choices for you. New game I am playing, large map all factions, seems to be going much better so far

on May 07, 2012

I played a Challenging game as Pariden, large map, dense everything, and beat the smack out of this build. Arcane Monolith's are awesome, and are the lazy player's way of expanding. with that said had a lot of fun with it. Neither the monsters or AI felt very challenging though, maybe because of how I spammed Arcane Monoliths and cleared monsters.

However, I next I played a tiny map 1v1 as Kraxis... against Gilden. Hard difficulty, dense everything, and let me tell you... the one opponent focusing on me is a whole different ballgame than a map loaded with a bunch of AI players duking it out. I tried to rush him early but honestly got pwned by monsters on the way... and then the monsters coming to attack my homeland made me beeline back to base. In the meantime Gilden recovered from early rush and is on the offensive. I haven't been this challenged in a long time. Game ain't over yet.

I'm really curious how a 1v1 on a large map on ridiculous would turn out. Brad's one faction AI free to focus on the player... will have to give that a go if I somehow win this hard game on a tiny map.

Really enjoying this build! Haven't come across the random Umberdroth or Drake roflstomping my areas yet.

on May 07, 2012


I've just returned to elemental after playing Wom for a couple of weeks then uninstalling it. I've been curious about the direction FE has taken so I decided to try out the beta to see how I liked it. I started with beta three and have played a bit with .913. I've only had time to play a couple of quick games with .913 so far but here is my feedback for what it's worth.

1) Monsters felt about right to me. There were weak ones in the immediate vicinity of my starting area and then stronger ones farther out. Perhaps more of a "medium" layer would be perfect but all in all I thought the monster distribution was good. The easy monsters were easy to clear and the hard ones tended to stop progress/expansion until I was strong enough to clear em.

2) AI players were only a threat in the sense that they came after me from multiple directions once they declared war. Thier armies weren't that difficult to kill but I had to pay attention to them destroying all my resources and cities.

3) Played on hard difficulty

4) Played 2 games so far in .913. First game was Gilden, summoner (just cause I wanted to try it out) large map, random map type moderate everything. Took water and earth level two so I could get the summon elemental spells from both spell lines. Can't remember the other settings. In the second game I played with the following settings:

Custom faction (Spartans), adepts, enchaters, flesh bound tome, no ranged weapons, Tarth race

Custom Soveriegn - earth disciple, water disciple, summoner, attunement and brilliant, weakness=cruel

Map - large, random map type

Options - Hard difficulty, normal pacing, moderate everything, 9 opponents

5) In my first game as Gilden I benefitted from a good starting location in the top corner which only allowed enemy AI's to come at me from one direction so I didn't have to worry much about defending my cities/resources from multiple directions. As a result, it was pretty easy to clear my territory with my summoned units and defeat enemy threats once they did declare war. I could then comfortably progress up the research tree to make some stronger units and them take my enemies cities. The terrain must have been mountainous which made the game easier since enemies could only approach from certain directions. I stopped playing after awhile because I seemed to have a pretty big lead in the power rankings over the next most powerful kingdom (Pariden) and I wanted to try another combination

In the second game I planned on being a summoner again so that I could get the shadow warg spell which would effectively give me two troops per battle if necassary early on without me having to train any troops in my cities. I took earth and water for the elemental summon spells as I leveled up to add to my summoned army. Took adepts so that I could immediately start harvesting shards to add to my mana supply, took enchaters so that I could capture resources using the arcane monolith spell so that I wouldn't have to use production lines to produce pioneers, and took flesh bound tome specifically for the Cull the Weak spell which I planned to use to build up my mona supply and heal myself during battle by sacrificing my summoned units. I also took cruel to give myself an extra point for picking traits and because I planned on not hiring any other hero's (ay least at the beginnig) and going at it with just my soverign and summoned unit stack.

In this game I found that my experience with monsters was pretty much the same as the first game. The easy monsters were easy to handle and clear and then as I got further out I encountered the odd "medium" monster and then the strong ones which stopped my progress until I was strong enough to take em on (basically whenever I got to level four earth and gained the summon earth elemental spell). On thing that almost screwed me was the cull the weak spell. In the description it says that it gives you 40 mana and life but the description is either wrong or the spell is bugged because whenever i tried it, i got the mana but not the health which almost killed me during a battle or two before I realised that. What was great about this spell as a summoner and what may make it overpowered is that it is a great way to get mana built up very quickly. Summoning a shadow warg during battle only costs you 22 mana but sacrificeing it gives you 40 so you end up with a net gain of 18 mana. Once I got  the summon ice elemental spell, it only cost me 15 mana to summon it during tactical battles and you get 40 for sacrificeing it, a net gain of 25 mana. By summoning creatures and then sacrificing them before the end of tactical battles, my mana pool was always bigger after battles then before and grew very quickly.

With the AI I found the same kind of thing as with my first game. The armies that were thrown at me weren't that hard to defeat but they came at me from many directions. As the map in the second game was much more open than the first game, I found that my initial rapid expansion with arcane monoliths were getting destroyed and my one stack wasn't enough to defend everything. So armies not hard to defeat but they threatened my cities/resources which prevented expansion without dedicating resources to train garrisons. The solution I found may be a bit of an exploit using the cull the weak spell. I ended up training alot of pioneers (the unit that took the least time to train), and sending them into battle with my soverign and elemental. Once in battle I would immidiately use cull the weak on all pioneers to build up a huge mana supply quickly and then after the battle I would use my level 3 earth - raise mountains spell to erect a wall of mountains around my territory so that enemies could only come at me from one direction. This has allowed me to eliminate threats to my cities and resources from multidirectional enemy attack and still only use one (or two at the most), stack of units to defeat enemies and expand my empire.  The game is on season 189 and I'm third in the power rankings and quite sure that I can now comfortably win now that I don't have to worry about attack from multiple directions.

6) Unfortunately I don't know how to do that

Have a good day and keep up the good work

12 PagesFirst 7 8 9 10 11  Last