Jump to content
Bearwere

Finishing off the opponents

Recommended Posts

So I was reading the readme, and here's the quote from it:

One general principle of this component (and of most of the AI components) is that enemies usually will not waste time finishing off helpless (paralysed, asleep, feared, etc) characters. With very few exceptions, BG opponents do not see themselves as there to soften up players so that later opponents can do better. They're fighting to win (albeit we as players know they're unlikely to) and so they're not going to kill someone paralysed who's out of the fight already.

Basically, it contradicts my combat experience. If I manage to stun/paralyze/sleep some enemy, the first thing I do is direct all my melee damagers to that enemy and kill him. I'm pretty sure I'm not alone in this. When spell time has run out, the opponent would raise and be a great nuisance again. Why let it happen? I do fight to win, and it means that all enemies must be killed.

Obviously it doesn't apply to chaos/fear/etc, because it doesn't help your fighters. But avoiding stunned opponents is a suboptimal choice.

Share this post


Link to post

From a "realisitic" perspective, AI behaviour actually makes sense. If you were fighting 2 people, and one of them becomes immobile, would you not focus your attention on the one who's still active and trying to kill you?

Some aTweaks monsters (Ghouls and the like, for example) have the opossite behaviour and will target their helpless targets. This also makes sense, they're predators.

What SCS does I think is making things less frustrating - it's very tiresome to pick up stuff from dead people, carry it over to the temples, re-equip character etc. This is far less troublesome in BG2 since your priests can raise dead, but in BG1 it's a pain to do, and if AI would target helpless targets it would happen a lot - even Remove paralyisis is an expensive spell for BG1 standards.

Share this post


Link to post

Kreso is right that this is partly about frustration and partly about in-game realism.

 

However, I'm also not convinced that it would be tactically optimal to kill the helpless in any case. Combat in the IE is generally pretty fast. Most disabling spells last at least five rounds and many last at least ten. Most battles (at least at low/mid levels) are over or at least decisively resolved before spells have time to wear off.

Share this post


Link to post
From a "realisitic" perspective, AI behaviour actually makes sense. If you were fighting 2 people, and one of them becomes immobile, would you not focus your attention on the one who's still active and trying to kill you?

Sure I wouldn't. I can kill him in basically one blow. If I don't, the next round I'll be having 2 people fight me again.

 

Very simple question - do you or do you not prioritize stunned enemies with your fighters?

 

Frustration I can understand. Under that argument you can ditch just about any mod that makes the game harder. Realism is what the player does. And I do eliminate helpless enemies first.

Edited by Bearwere

Share this post


Link to post

Very simple question - do you or do you not prioritize stunned enemies with your fighters?

No. As DavidW mentioned, Hold type spells last for very long, not to mention PW:Stun, Emotion and similar spells.

Of course, it's a bit more complicated:

- generally, you'll want enemy mages dead first

- if you can't remove their protections/invisibility etc. then focus clerics and/or thieves

- fighters you should generally disable

- if mages are still too evasive/protected, focus the disabled fighters while debuffing the mage

- if not, focus your warriors on mages, target disabled fighters with your mages/clerics/possibly summons (in between spell removal casting) since THAC0 becomes irrelevant.

That's how I play it at least. There's no point in wasting time attacking a stunned fighter with 100+ HP while enemy cleric is casting Flamestrike, or wasting time when ToB mages cast Imprisonment/ADHW.

Share this post


Link to post

No. As DavidW mentioned, Hold type spells last for very long, not to mention PW:Stun, Emotion and similar spells.

Of course, it's a bit more complicated:

- generally, you'll want enemy mages dead first

- if you can't remove their protections/invisibility etc. then focus clerics and/or thieves

- fighters you should generally disable

- if mages are still too evasive/protected, focus the disabled fighters while debuffing the mage

- if not, focus your warriors on mages, target disabled fighters with your mages/clerics/possibly summons (in between spell removal casting) since THAC0 becomes irrelevant.

That's how I play it at least. There's no point in wasting time attacking a stunned fighter with 100+ HP while enemy cleric is casting Flamestrike, or wasting time when ToB mages cast Imprisonment/ADHW.

100+HP. In ToB. That's like 2-3 hits. 1/3 of a round. And if they're not made, the next round these hits will be delivered on you. OK.

What if it isn't warriors who are disabled? What if it's mages, or clerics? What if it isn't ToB?

 

Kreso is right that this is partly about frustration and partly about in-game realism.

 

However, I'm also not convinced that it would be tactically optimal to kill the helpless in any case. Combat in the IE is generally pretty fast. Most disabling spells last at least five rounds and many last at least ten. Most battles (at least at low/mid levels) are over or at least decisively resolved before spells have time to wear off.

Is it not effective or is it too effective? Pick one.

In-game realism argument is not valid. Just throw some dice to see it. Killing stunned is the best way to win.

As for the duration, Command lasts 1 round, and it's effective through all BG1 at least.

 

 

What I mean to say, there sure are times when you'd better try to interrupt enemy's mage spellcasting (especially on the higher levels). But there also are times when you'd better just destroy what you can with your sword, while you can. And it doesn't look like SCS takes it into account.

Edited by Bearwere

Share this post


Link to post

My previous post is generalization, ofc. If there are 2 oponnents left, and I'm sure the one who's not disabled will not endanger any lives, sure, I'll kill the disabled one. If somehow an enemy mage gets disabled instantly, I'll probably (again, this is situational and complete theorycrafting) kill him asap. Depending on who's left fighting, perhaps I won't.

The main point here is that a disabled oponnent is essentialy dead the moment he fails his save or goes under HP treshhold for PW Stun (if you have that spell), therefore wasting time on attacking him is pointless. It's not 2 or 3 seconds, it's more since you still must move in range and make the attack. This is of course, much more prominent in BG1, where things such as Hold Person equal Finger of Death in BG2 by their effectivness. In late BG2, mages won't care about who's disabled or not, they'll simply ADHW/Fireball/Chain Lightning everyone. In ToB, nobody will care about disables anyway since Whirlwinds, Planetars and Comets are available (Hold and similar spells just doesn't compete anymore), and disables are much easier to resist, with notable exeptions such as Holy Word and the like.

In BG1, point that you don't need to focus disabled oponnents is completely valid - if remaining active oponnents hit you few times, you die. And they do hit at least 5% of time, regardless of AC.

A perfect example of this is Bandit Camp battle in BG1. You can disable >10 oponnents. If you don't focus the remaing bunch, but waste time on those disabled (and it takes more than 3 seconds to kill one of them with swords), they'll kill you easilly even if you had -15 AC, by the sheer power of d20 dice.

In BG2/ToB, point is simply non-existent, since mages change their role from casters of Sleep/identify/fireball to Aoe damage/debuffers, while clerics become (more or less) dedicated healers/protectors, whilst in BG1 they both excell at disabling.

Share this post


Link to post

Very simple question - do you or do you not prioritize stunned enemies with your fighters?

Very simple answer - no, in general I do not, for the same reasons as Kreso.

 

Frustration I can understand. Under that argument you can ditch just about any mod that makes the game harder. Realism is what the player does. And I do eliminate helpless enemies first.

That's not what I mean by realism. I mean, "not immersion-breaking". Players can do all manner of unrealistic things. (And in any case they do lots of different things, as witness the fact that your strategies aren't the same as mine or Kreso's.

 

I also don't agree that frustration and difficulty increase are synonymous. A major design goal for SCS is to minimise increased frustration.

 

 

100+HP. In ToB. That's like 2-3 hits. 1/3 of a round. And if they're not made, the next round these hits will be delivered on you. OK.

I have some time for the view that in ToB - when disables are shorter and damage is higher - it often makes sense to kill the helpless quickly. To some extent SCS betrays its origins as a mod developed originally for BG1, then for SoA, and finally for ToB. Even there, though, I think the probability of kill-the-helpless leading to something immersion-breakingly stupid remains too high. (Avoiding immersion-breaking stupidity is SCS's highest priority, much higher than being tactically optimal.) And even there, 1 round is short for a disabling hit, and in any case (if we're talking tactical optimisation and putting aside immersion) by ToB death is actually quite a mild "disabling" effect compared to some - killing a player can actually benefit them!

 

Kreso is right that this is partly about frustration and partly about in-game realism.

 

However, I'm also not convinced that it would be tactically optimal to kill the helpless in any case. Combat in the IE is generally pretty fast. Most disabling spells last at least five rounds and many last at least ten. Most battles (at least at low/mid levels) are over or at least decisively resolved before spells have time to wear off.

Is it not effective or is it too effective? Pick one.

I'll pick both, thanks. Killing the helpless is less effective at achieving victory outright, more effective at causing PC casualties in the course of a losing battle.

 

In-game realism argument is not valid. Just throw some dice to see it.

As I noted above, what I mean by "realism" has very little to do with the game mechanics.

 

As for the duration, Command lasts 1 round, and it's effective through all BG1 at least.

And for exactly that reason, Command is a specific exception - if a cleric casts Command at a melee opponent, they'll prioritise that opponent for the next round. (It's not technically viable to have other opponents do likewise). But only a very small fraction of disabling events have durations that short. (Note that we're talking about NPC-caused disabling effects here, and in SCS Command gets relatively little use compared with longer-duration disabling attacks.)

 

What I mean to say, there sure are times when you'd better try to interrupt enemy's mage spellcasting (especially on the higher levels). But there also are times when you'd better just destroy what you can with your sword, while you can. And it doesn't look like SCS takes it into account.

That's a different argument. Of course the best strategy isn't "never kill the helpless" or "always kill the helpless"; it's "make a careful tactical assessment of which makes sense at any given moment, do that, and update the assessment on a round-by-round basis". But that is wildly beyond anything that the BG2 scripting system can handle. Ultimately, AI scripting in BG2 has to consist of relatively simple rules of thumb. So far you're not persuading me that changing this particular rule of thumb is advisable for any of realism, tactical challenge, or gameplay experience.

 

(Incidentally, this is a perennial point of discussion about SCS - someone makes this point every year or two. So far there's been a fairly strong majority preference for something like the present behaviour - not that I do these things democratically in any case, but it's good to keep track.)

Share this post


Link to post

So I was reading the readme, and here's the quote from it:

One general principle of this component (and of most of the AI components) is that enemies usually will not waste time finishing off helpless (paralysed, asleep, feared, etc) characters. With very few exceptions, BG opponents do not see themselves as there to soften up players so that later opponents can do better. They're fighting to win (albeit we as players know they're unlikely to) and so they're not going to kill someone paralysed who's out of the fight already.

Basically, it contradicts my combat experience. If I manage to stun/paralyze/sleep some enemy, the first thing I do is direct all my melee damagers to that enemy and kill him. I'm pretty sure I'm not alone in this. When spell time has run out, the opponent would raise and be a great nuisance again.

 

 

 

Not always. If I manage to stun/paralyze/sleep the fighter, I will leave him be (for now) and go after the mage. In like 70% cases. Usually, concious mage is much greater nuisance than disabled fighter. really depends on the actual battle situation and on which enemy is more dangerous.

Share this post


Link to post

In the same topic, I really like your easy of use script that comes with SCS. There's only one thing that's irritating about it, the fact that characters don't hit helpless targets if I don't force them.

 

Mostly harmless, but if I for example get web off and all enemies get caught, then my characters just stand still doing nothing unless I micro them to attack targets. Wish there was an option in that script to not ignore helpless targets.

Share this post


Link to post

The main point here is that a disabled oponnent is essentialy dead the moment he fails his save or goes under HP treshhold for PW Stun (if you have that spell), therefore wasting time on attacking him is pointless. It's not 2 or 3 seconds, it's more since you still must move in range and make the attack.

That if you don't heal. Or Dispel. Or something else. And missile attacks also get the bonuses.

If the opponent is essentially dead, then the change would do essentially nothing.

 

That's not what I mean by realism. I mean, "not immersion-breaking". Players can do all manner of unrealistic things. (And in any case they do lots of different things, as witness the fact that your strategies aren't the same as mine or Kreso's.

 

I also don't agree that frustration and difficulty increase are synonymous. A major design goal for SCS is to minimise increased frustration.

 

I have some time for the view that in ToB - when disables are shorter and damage is higher - it often makes sense to kill the helpless quickly. To some extent SCS betrays its origins as a mod developed originally for BG1, then for SoA, and finally for ToB. Even there, though, I think the probability of kill-the-helpless leading to something immersion-breakingly stupid remains too high. (Avoiding immersion-breaking stupidity is SCS's highest priority, much higher than being tactically optimal.) And even there, 1 round is short for a disabling hit, and in any case (if we're talking tactical optimisation and putting aside immersion) by ToB death is actually quite a mild "disabling" effect compared to some - killing a player can actually benefit them!

I don't get your separation of "tactics" and "immersion". The best available tactics is the least immersion-breaking, as I see it (leaving aside specifically added variations).

 

And for exactly that reason, Command is a specific exception - if a cleric casts Command at a melee opponent, they'll prioritise that opponent for the next round. (It's not technically viable to have other opponents do likewise). But only a very small fraction of disabling events have durations that short. (Note that we're talking about NPC-caused disabling effects here, and in SCS Command gets relatively little use compared with longer-duration disabling attacks.)

So there are exceptions, after all? Any more of them?

1) I can name more short-duration spells (chromatic orb comes to mind)

2) a disable can be dispelled

3) there are spells that allow a save every round (web, for example)

 

That's a different argument. Of course the best strategy isn't "never kill the helpless" or "always kill the helpless"; it's "make a careful tactical assessment of which makes sense at any given moment, do that, and update the assessment on a round-by-round basis". But that is wildly beyond anything that the BG2 scripting system can handle. Ultimately, AI scripting in BG2 has to consist of relatively simple rules of thumb. So far you're not persuading me that changing this particular rule of thumb is advisable for any of realism, tactical challenge, or gameplay experience.

Well, you did a pretty good job with casters, and I think fighters do deserve to be thrown a bone, if only with better targeting. Of course, if it's just not possible with BG2, then there's nothing to talk about.

 

Not always. If I manage to stun/paralyze/sleep the fighter, I will leave him be (for now) and go after the mage. In like 70% cases. Usually, concious mage is much greater nuisance than disabled fighter. really depends on the actual battle situation and on which enemy is more dangerous.

I don't mean always, but pretty often, at the very least. So if it's not the fighter, but the mage? Will you still avoid him and target on the healthy fighter? SCS will. That's my point.

Edited by Bearwere

Share this post


Link to post

.... then the change would do essentially nothing.

....see the point?

Share this post


Link to post

If the opponent is essentially dead, then the change would do essentially nothing.

It would waste time.

 

I don't get your separation of "tactics" and "immersion". The best available tactics is the least immersion-breaking, as I see it (leaving aside specifically added variations).

No, for two reasons.

 

(i) Consider the following:

Script A has an 70% chance of picking the best tactic, a 10% chance of picking an okay tactic, and a 20% chance of picking a tactic so stupid that it's obvious only a computer would do it.

Script B has a 30% chance of picking the best tactic and a 70% chance of picking an okay tactic, but never picks an immersion-breakingly stupid tactic.

 

SCS would go with B over A. But probably A is tactically better.

 

(ii) It's a role-playing game, not chess. It's perfectly plausible that creatures engaged in melee wouldn't always do the optimal thing. Indeed, because of limitations of the D&D combat system, what's optimal in D&D isn't always realistic - basically D&D decouples attack and defence in a way that doesn't track the way real melee works.

 

And for exactly that reason, Command is a specific exception - if a cleric casts Command at a melee opponent, they'll prioritise that opponent for the next round. (It's not technically viable to have other opponents do likewise). But only a very small fraction of disabling events have durations that short. (Note that we're talking about NPC-caused disabling effects here, and in SCS Command gets relatively little use compared with longer-duration disabling attacks.)

So there are exceptions, after all?

"After all?" I don't recall denying that there were any in the first place; indeed, you quote my readme as noting explicitly that there are exceptions.

Any more of them?

Offhand: mind flayers prioritise helpless targets, as much for RP reasons as anything else; meleers won't leave a nearby helpless target to seek out a faraway vulnerable target; spiders prioritise targets caught in their webs; demons (iirc) occasionally teleport to vulnerable targets.

 

1) I can name more short-duration spells (chromatic orb comes to mind)

It's not really possible to distinguish between different sources of the same effect. ("not really possible" means "not possible at all" pre-ToBEx, "possible in principle but with an unacceptable script-length overhead" with ToBEx".)

2) a disable can be dispelled

That wouldn't be an exception; it applies in general. (I'm also happy with players having some chance to dispel disabling effects; it adds combat variety.)

3) there are spells that allow a save every round (web, for example)

In general, attacking people in webs is contra-indicated because you can get stuck in them too. (Spiders, with web immunity, are a hand-coded exception.)

 

I think fighters do deserve to be thrown a bone, if only with better targeting. Of course, if it's just not possible with BG2, then there's nothing to talk about.

I can only repeat my previous comment: I can do "always attack the helpless" or "never attack the helpless". I can't do "attack the helpless only if it's tactically optimal" because that's beyond the current state of the art in AI, let alone the pitiful resources of IE scripting. I can do slightly more nuanced variants ("attack the helpless only if XYZ occurs") and I'm happy in principle to hear suggestions for XYZ, but a typical SCS combat script for a fighter is 3000-4000 lines long (compared to 27 lines in baseline BG2) and I don't really want to make it any longer!

Edited by DavidW

Share this post


Link to post

I can only repeat my previous comment: I can do "always attack the helpless" or "never attack the helpless". I can't do "attack the helpless only if it's tactically optimal" because that's beyond the current state of the art in AI, let alone the pitiful resources of IE scripting. I can do slightly more nuanced variants ("attack the helpless only if XYZ occurs") and I'm happy in principle to hear suggestions for XYZ, but a typical SCS combat script for a fighter is 3000-4000 lines long (compared to 27 lines in baseline BG2) and I don't really want to make it any longer!

Well, that sums it all. If it can't be done, then it can't be done.

Share this post


Link to post
Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...