Wish List: targetable spells

[Daynarr]

Mages will make good choices in spells (although you can tell them what to cast) and targets as well. For example, my pretender from before (level 9 fire) wasn’t shooting fire darts at enemy who were engaged in close combat with my troops, but at enemy that was behind them, or if there was no such enemy he would use more accurate spells, like incinerate to avoid hitting his own troops. All in all i think I saw him make only 1 friendly kill so far while he killed about 500 enemies total.

Oh yeah, it's true AI will ignore commands that will be useless in battle (like casting arrow fend if there is no enemy ranged troops on the field and he will also restrain himself from using gems if the battle is easy to win.

[Endoperez]

Oh the joy! This is better than what I dared to dream! Anyways, some questions for you about the mage AI:

IF a mage has skill in multiple paths, how does he decide what to spells to use? Does the fatique the spell causes taken into account? What about storm/ rain and their effects on the fire magic?
If the mage has gems and no (intelligent) orders, does he summon or does he cast? From what this depends?
Is there any situation in which the mage casts a wrong spell more often than not?

The Last one interests me most. Does the AI have any faults? And will the players be able to use them? Like getting the mages to use the gems with Call of the wind...

[Nerfix]

I think they fixed the "nuke spells at trivial enemies" syndrome.

[Mortifer]

Awesome.
'Spellcasting AI' wasnt really good, Im glad that it will be tweaked and fixed.

[Psitticine]

I've found spellcasting AI to be quite good overall. In fact, I think it might be the strongest part of the tactical AI in Dom II.

The AI is conservative when it comes to using gems against weaker opponents, but doesn't hesitate to bring out the big guns when threatened. It is also very good about casting appropriate defensive spells.

Sometimes I'll wonder for second why it chose a seemingly pointless spell, like Holy Avenger by/on a priest that'll probably not get attacked, but then I realize the priest in question really didn't have anything more useful to cast that round anyway.

I've noticed a tendency to lean on conjuration, but that might just be because that's what I have researched for them most often. Bit of a contaminated sample there, if you know what I mean.

It is also very good at targetting. The only problem I've seen is when I was playing the flame-retardent Abyssians and had one group of Militia along for the ride. (Hey, they volunteered!) In one particular battle, my most powerful mage used some fire spells that didn't act as any threat to my regular guys, but were a bit hard on those soft-skinned Militia. I think that decision may have been weighted by the overall desperation of the battle: we were badly outnumbered and the enemy Pretender God was present on the field of combat. (Our guys took it on the chin in the end; it wasn't pretty!)

Other than that incident, the AI seems to do a pretty good job of trying not to toast your guys while frying the enemy. The "arrow-in-the-back" syndrome has been greatly reduced, especially with one particular beta-patch, and is now rather rare. I only run into problems when either I've given orders that force the situation (oopsie) or things have gotten either truly desperate or utterly chaotic.

The large spells are used fairly rarely because of their cost. Sometimes, I'd like to see them used more often, just for the pure joy of watching them go, but I can't argue with the conservation of gems. It can be a pain to keep your mages stocked with them if they are pushed into constantly using them up.

The portion of the combat AI that I'd like to see improved is in the "mad rush towards the enemy" department. Keep it together, lads! If Illwinter were to turn their hand towards AI improvements some day, I'd suggest they start with that and wait until Last to see if there's anything to tweak with the spell-casting.

Oh, Last minute thought: I have seen the spell-casting AI conjure things in back that can't seem to make it up to the front line in time to be a part of the battle. That particular problem is the only thing I've found I don't like, but it might be a part of the fact I lean too heavily on Conjuration research. They may just not have anything better to do at those points beause I haven't researched them the spell options, but I can't say for sure. (It'll take more playing to find out . . . yeah, that's a good excuse to play more . . . )

[Wendigo]

Originally posted by Mortifer:
Awesome.
'Spellcasting AI' wasnt really good, Im glad that it will be tweaked and fixed.

I actually considered the Dom I spellcasting AI pretty good, and it often surprised me with excellent spell choices of its own well past the 5 scripted spells: like Decay vs Iron Dragons, Blade Wind vs Seasonal Spirits, Star Fires vs Combat Pretenders, Nether Darts vs HI...always fun also when one of your Seihktonur runs into a Combat pretender unexpectedly & stamps a big fat curse on the offender's chest before dying a glorious death.

The only problems I had with it was its casual disregard of friendly fire issues (most notably when a caster comitted suicide by casting an area spell vs a close enemy), and the gem wasting/gem draining complex...but on the whole I rate the spell AI pretty high.

[Pocus]

Originally posted by Psitticine:
Oh, Last minute thought: I have seen the spell-casting AI conjure things in back that can't seem to make it up to the front line in time to be a part of the battle. That particular problem is the only thing I've found I don't like, but it might be a part of the fact I lean too heavily on Conjuration research. They may just not have anything better to do at those points beause I haven't researched them the spell options, but I can't say for sure. (It'll take more playing to find out . . . yeah, that's a good excuse to play more . . . )

there is a similar syndroma in doms I, when you want your mages to cast safely spells. Generally, for summons, you have to move them forward a bit, and take some risks with enemy fire.

another issue too:

will the AI use gems when the combat can be won otherwise? I encourage you especially to cast call of the wild / call of the wind on a big enemy army (or target one of your own with an horror, phantasmal army, or any other unaligned spell). Will they waste gems in big spells (even if scripted) and will they try to handle the menace without too much spending?