December 13th, 2010, 17:59
Posts: 901
Threads: 28
Joined: Oct 2008
WhiteMage Wrote:Kyrub, although your sharp positions may lead you to the desired effect against players who are roughly at the same level as you are, international grand masters would also counter it against you nearly 100% of the time, not only the AI. Absolutely. But note, that it is based on the nature of chess - it is a game with relatively low complexity, because there is a very fixed number of possible moves on a board. Less branching, easier counting, less surprises. With more branching, I could give more headache to my brother, a great strategist - and a much better chess player than me. (Due to low branching and deterministic movement, I like to compare MoO to chess. This game could have an outstanding, unbeatable AI.) - All this is just a side note, I love babbling about chess and games nature (or the games theory) in general.
Quote:First let’s fix the game, then the rules, then the smart AI and then the personality AI.
Perfect. I'm back to work.
Posts: 28
Threads: 1
Joined: Feb 2011
In my current game as Custom Wizard replacing Rjak with 10 death books and Conjurer I had cast "Evil Presence" against the capitals of my enemies. 
A while later I conquered the capital of S´ssra on Myrror. :
Another while later I noticed by chance that the "Evil Presence" was still there when I looked at the city screen 
Evil presence is described as not having any effect when the target is a wizard who owns death spell books.
Shouldn´t the spell then eliminate itself automatically? Especially if you have YOURSELF become the target of your own, then useless, spell? 
Or is that not possible?
Posts: 973
Threads: 20
Joined: Oct 2010
ConjurerDragon Wrote:In my current game as Custom Wizard replacing Rjak with 10 death books and Conjurer I had cast "Evil Presence" against the capitals of my enemies.
A while later I conquered the capital of S´ssra on Myrror. :
Another while later I noticed by chance that the "Evil Presence" was still there when I looked at the city screen
Evil presence is described as not having any effect when the target is a wizard who owns death spell books.
Shouldn´t the spell then eliminate itself automatically? Especially if you have YOURSELF become the target of your own, then useless, spell?
Or is that not possible?
Actually the spell description says "The spell cannot be cast on a city whose controlling wizard understands the elements of death magic."
In other words, since you have death books yourself, you cannot cast it on your own cities, but once cast it doesn't disappear automatically, nor does it lose its effect.
And I'd hardly call the spell useless on your own city. I'd call it "counter productive"  . As yet I can't think of a reason how it could be desirable, but there's probably a scenario somehow that makes it useful. Perhaps to get rid of a city without losing fame or something like that. Either way it's a good thing you can simply deactivate the spell yourself. 
--I like ILSe
Posts: 357
Threads: 22
Joined: Oct 2010
ConjureDragon, please don't use enter instead of full stop. (.)
There are cases where not dispelling Evil Presence or a similar bad spell could be good in a long run. For example you conquer a city, but you know you will lose it soon. It may be less expensive to keep the spell going instead of casting it again.
Just razing the city would result in a Fame loss. But what if you want to recapture the city, but don't have any forces nearby right now ?
Admittedly, current behavior is more trouble than it's worth. Such hypothetical situations are very rare. They would be more common if AI was better. Yet players who know about it can easily avoid it, and I wouldn't risk demotivating kyrub by asking him for such a trivial change. Movement bugs are much more annoying (for example if a unit stops on a square with a stack which is out of movement points, the unit loses all of its movement points). Or cities not counting as roads if a boat is the final destination.
Posts: 7
Threads: 0
Joined: Apr 2011
Posts: 28
Threads: 1
Joined: Feb 2011
Nad Wrote:Okay. I don't really like this feature in MoO either.  The problem is that you can turtle in MoO after researching the first planetary shield and solid missiles (e.g. planetary 5 + merculite), but I don't see an adequate concept in MoM. Spell wards and other high level spells take way too long if being available at all. Furthermore, apart from scouting magic spirits which raze unguarded outposts I don't see a way of flagging land (systems in MoO with scouting ships) appropriately. I don't mind battles between smaller armies, but attacking the fortress is something else. I'm a peaceful player who doesn't do this without former provokation. Especially a 11-book starter will burn you with his strong summons... I'd love to see a more peaceful game from time to time, perhaps it's on personal issue.
That´s no problem - if you want a more peaceful game take the "charismatic" pick and small landmasses. Suddenly you will have far less problems and wars with the other wizards.
October 21st, 2011, 16:35
Posts: 901
Threads: 28
Joined: Oct 2008
I played a first spectator (AI only) game after a six month void. Huh. There are several mismanaged concepts:
- After the fix, that all but prevents clumsy unit bumping, AIs almost never declare war on themselves. Rarely, I have seen them come close, cancelling their pact... only to reinstate it few turns later (I had no chaotic wizard in the game). This is very bad, it gives the game plastic feel. It has to be changed. (similar experience from others?)
- War bears are everywhere. The AI spams teddies in untold quantities. Shocking, it's a left-over of mine in the code. You could have told me (I mean, sorry, ehm).
- AI tends to spend incredible amount of time creating artifacts. That's not bad by itself, only it should do it much later in the game. Very important fix.
- I think I overdid the emphasis on the AI player objective. I wanted the AI wizards to play differently. But in my game, I had AIs totally ignoring the all-important Marketplace building until very late in the city development. This must change.
- Mercennaries seem fine right now. I am undecided on fortress defense. Either it changes very little, or it is dysfunctional. Don't understand the problem. Not good. Items, I couldn't tell in the spectator scenario.
- The amount of battles when AI loses its hero in the (automatic) combat is shocking. I actually noticed it because I used the automatic combat to quickly defend my own fortress. In about 12 battles, I lost my hero (1 from 9 units) 4 times, with usually about 6 units remaining. I looked in the code, and it seems there is a prevention mechanism (to keep the heroes) that is completely broken, making the hero more likely to be killed. This is on very good way to be fixed, it will improve AI heroes.
- AI is not extremely good at coping with only one continent (Huge). It was programmed to operate on several ones. Still, this is not a tragedy, the game is challenging as it is. In every case I see very little I can do about it, this is coded too deep in AI structure.
What do you think? Any other points?
October 21st, 2011, 23:20
(This post was last modified: October 21st, 2011, 23:43 by momfan.)
Posts: 121
Threads: 2
Joined: Dec 2010
kyrub Wrote:I played a first spectator (AI only) game after a six month void. Huh. There are several mismanaged concepts:
- After the fix, that all but prevents clumsy unit bumping, AIs almost never declare war on themselves...
- War bears are everywhere. The AI spams teddies in untold quantities...
- AI tends to spend incredible amount of time creating artifacts...
- I think I overdid the emphasis on the AI player objective. I wanted the AI wizards to play differently. But in my game, I had AIs totally ignoring the all-important Marketplace building until very late in the city development...
- Mercennaries seem fine right now. I am undecided on fortress defense. Either it changes very little, or it is dysfunctional. Don't understand the problem. Not good. Items, I couldn't tell in the spectator scenario.
- The amount of battles when AI loses its hero in the (automatic) combat is shocking...
- AI is not extremely good at coping with only one continent (Huge). It was programmed to operate on several ones...
What do you think? Any other points?
AI vs AI:
In 1.40k I see AIs at war with each other and cities changing hands, and Corruption being cast on lands around the AI cities. Quite gratifying actually. So it was fine in 1.40k, they didn't all gang up on the player immediately. I think it helps if at least one AI is chaotic or has Death books.
War bear spam:
Not obvious to me, but I don't see many summoned creatures in general. In my recent 1.40k games I remember seeing guardian spirits, war bears, sprites and basilisks but not a whole menagerie. In vanilla I remember seeing storm giants, sky drakes, angels etc.
The scarcity of summoned creatures could be linked to the Create Artifact obsession, but it's not visible without Detect Magic, which I only cast in the late game, and only if I also have Spell Blast.
Perhaps you can look in the code surrounding how the AI chooses a creature to summon? Maybe the code is broken and jumps to war bears too frequently, or the AI makes some optimization calculation and war bears fall right on the margin somehow. Presumably there is a code sequence like the following:
- Casting Spell? Yes/No
- If not casting spell, start new spell? Yes/No
- If Yes, what type: Summon, City, Enchantment, Unit, Special etc?
- If summon, what creature?
The code for "what creature" is probably where you should start looking. It's possible that the Nature spell book is given priority for example, so war bears get chosen even when guardian spirits, hell hounds, nagas, ghouls etc are also available and more suited to the situation.
Marketplace:
All AI cities seem to have marketplaces built by the time I take them, unless it's early in the game where they only have a granary and barracks and maybe a smithy. Could also be obscured at Impossible level due to fast AI city development - the "wrong" sequence is not visible because the AI city develops so quickly.
AI heroes:
It is trivial to kill AI heroes. First, they do not carry good items, second they charge into battle even when outnumbered, especially the mages(!). In one recent game I had a level 1 Malleus charge my troops (1x swordsmen + 1x cavalry) after using up his magic ranged attacks. Malleus did run away when facing overwhelming odds though (2x Basilisk).
Sad to hear automatic combat is even worse.
Huge Landmass:
Probably the game was envisioned by the designers as a world of water with land on it, not a world of land with water on it. So maybe we should respect that choice. Huge landmass only adds to the AI advantage anyway since they get to build even more cities. I'm not sure it improves gameplay.
(additional inputs)
Lairs:
Maybe the Magic power option at game creation could affect lairs too? I find it a bit daft that with Powerful magic the nodes are guarded by drakes and wyrms, while the lairs are guarded by war bears or hell hounds. Not saying there shouldn't be weak lairs, but their strength (and loot) should scale together with the nodes and towers.
AI combat strategy:
I think this has been mentioned before, about the AI advancing its troops at the speed of the slowest unit. This makes them easy targets for archers, when in fact the fast units should close in ASAP.
Maybe the AI can have some simple routines for non-heroes:
Melee Units:
- charge!
- use up all movement to close in and fight.
Ranged Units:
- shoot until ammo = zero
- if no ammo, wait until all melee units are dead
- if all melee units are dead, charge!
This will put melee units into action quickly, and keep ranged units out of trouble. Once the melee units are dead, the ranged units have nothing to lose by fighting too. Of course, each round the AI should also check whether it should flee.
Volcanoes:
According to the README.TXT for vanilla, volcanoes are supposed to have a 2% chance to revert to mountains each turn, with a 5% chance of providing a mineral vein. I have raised dozens, if not hundreds, of volcanoes in vanilla and 1.40j/k, and most turn into mountains eventually or are converted to hills by Gaia's Blessing. I have NEVER had one give me minerals. Relatively minor, but it would make Raise Volcano more useful instead of just being a tundra-conversion tool.
October 22nd, 2011, 01:05
Posts: 525
Threads: 6
Joined: Dec 2010
momfan Wrote:Volcanoes:
According to the README.TXT for vanilla, volcanoes are supposed to have a 2% chance to revert to mountains each turn, with a 5% chance of providing a mineral vein. I have raised dozens, if not hundreds, of volcanoes in vanilla and 1.40j/k, and most turn into mountains eventually or are converted to hills by Gaia's Blessing. I have NEVER had one give me minerals. Relatively minor, but it would make Raise Volcano more useful instead of just being a tundra-conversion tool.
The native chances are : voulcanoes provide mineral vein 20%. May be this is test version. And people, who play with this feature (volcano can provide mineral vein) says this is not good. Becouse the vein appear when the volcano raises, not when it revert to the mountain. And bad ,becouse its very like for player to raise fields of adamantium, coal ore, gem mine near the capital and others own cities. They even try get 10 iron mine to build paladins for free.
October 22nd, 2011, 01:38
Posts: 121
Threads: 2
Joined: Dec 2010
Asfex Wrote:The native chances are : voulcanoes provide mineral vein 20%. May be this is test version. And people, who play with this feature (volcano can provide mineral vein) says this is not good. Becouse the vein appear when the volcano raises, not when it revert to the mountain. And bad ,becouse its very like for player to raise fields of adamantium, coal ore, gem mine near the capital and others own cities. They even try get 10 iron mine to build paladins for free.
The official README.TXT for 1.31 says (verbatim quote):
- Volcanoes have a 2% chance to revert to mountains each turn (with a 5% chance of providing a mineral vein).
So a 20% chance means that version you are playing has been modified.
If it works as described (5% chance of minerals, only after the 2% chance of reverting to mountain) it costs too much for players to abuse. I can see how 20% chance of minerals immediately upon raising the volcano would change player behaviour significantly, especially with the benefit of save-reload.
|
