821 replies to this topic

[Andrea C.]

You might as well be requesting this: [CUT]

Hi Galactygon,

yes, it really looks like something that could easily achieve what I was asking for It would be awesome if it were implemented. I'm no modder myself, but I love mods that bring BG closer to the PnP rules of AD&D.

[Ascension64]

Another wish, possibility to equip a buckler with a two handed weapon (bow, crossbow, staff..) as stated in the description.

Similar was requested before. Not happening, because the engine relies too much on checking the shield slot.

Can you clarify this? If for example SOMEITM.itm has 1 melee attack ability and 2 abilities that use charges, how does this impact the attack animation for BOW01.itm?

Oh, if SOMEITM.itm was a weapon then it would use the attack for SOMEITM.ITM. However, if SOMEITM.itm was an amulet with a charge ability that forced SEQ_ATTACK, and you had a BOW01.ITM equipped, the engine will try to look for an attack (not shoot) animation for the bow, which doesn't exist.

This came up when I wanted the modified ranged Spiritual Hammer/Mordenkainen's Sword/Black Blade of Disaster to use the SEQ_READY animation to reflect that the spiritual hammer was telekinetically wielded as per AD&D. Even if I remove the hammer animation from the paperdoll, zero out the hack/thrust/backhand animations, and add the relevant lines to ITEMANIM.2da, the character still swings in the air.

This is very dangerous to change, because the engine actually relies on the correct animation sequence to play during attacks to work out things like which part of the round it is in, etc, when actions are processed, etc. Knowing this, are separating ability indices in ITEMANIM.2DA still useful if restricted to non-attack abilities?

Charname's death can be handled by manually adjusting all plot cutscenes, etc. And indeed, I'd love it too. But - there're mods, that may have their own Player1 action references. In other words, it'll be white-, not blacklisting compatible mods.
Still, if it doesn't take much work, I suppose it could work out for expert users, well familiar with game and tools. Or TCs.

Perhaps a modding-only tweak. I imagine something similar operates in the Arenas game type.

Is it possible to get sneak attack like in 3E (and the option in IWD) where thief's attack from side or back is 1d6 (on up) depending on level.

Could you explain what you mean by "on up" in the text in bold?

Edited by Ascension64, 28 February 2011 - 04:48 AM.

[Ascension64]

Here's one I'm keen on:
The engine rolls saving throws and calculates luck bonuses before the first feature block of the extended header is applied, rather than before the first feature block where the saving throw or damage reduction is calculated. This prevents us from applying effects that give bonuses to saving throws and luck bonuses on the fly (ie. in the same extended header). This is not the case with damage/magic resistances, where they are calculated when first looked for, so this is some engine inconsistency.

If the engine could take saving throw modifiers/luck modifiers into account when placed in the same feature block, but right before the order in which the rest of the effects are applied, then it would be possible to implement AD&D saving throws vs. damage types / wisdom modifiers to saving throws against charm-related effects.

-Galactygon

Yeah, so finally had a look at this. The saving throws are pre-rolled every time the creature is "refreshed" or "AIUpdate" (every tick), which is completely separate from effect application itself. While you can still give on-the-fly saving throw bonuses (did not look at luck), the same roll is used against it. This isn't really an engine inconsistency per se because the engine needs the saving throw values before an effect is applied, whereas something like magic resistance is conditional, and has effect-specific application.

So, for example, you have a block that applies -2 save vs death penalty and then instant death if not save vs death. The target's current save vs death is 13. The target rolls 15. The target won't die because 13 + 2 >= 15.

Edited by Ascension64, 28 February 2011 - 05:18 AM.

[Galactygon]

Oh, if SOMEITM.itm was a weapon then it would use the attack for SOMEITM.ITM. However, if SOMEITM.itm was an amulet with a charge ability that forced SEQ_ATTACK, and you had a BOW01.ITM equipped, the engine will try to look for an attack (not shoot) animation for the bow, which doesn't exist.

Got it. This could be circumvented by disabling all the weapon animations while the animation in ITEMANIM.2da is playing. Just like when characters cast their spells, the weapon animations disappear temporarily.

This is very dangerous to change, because the engine actually relies on the correct animation sequence to play during attacks to work out things like which part of the round it is in, etc, when actions are processed, etc.

Whenever activating an item with a special ability, doesn't the activation animation in ITEMANIM.2da interrupt and override the previous animation without regard for which frame it's at?

Knowing this, are separating ability indices in ITEMANIM.2DA still useful if restricted to non-attack abilities?

Yes.

Yeah, so finally had a look at this. The saving throws are pre-rolled every time the creature is "refreshed" or "AIUpdate" (every tick), which is completely separate from effect application itself.

Got it.

While you can still give on-the-fly saving throw bonuses (did not look at luck), the same roll is used against it.

Fair enough. But the on-the-fly bonuses/penalties are not applied to that roll, so that succeeding effects compare with the character's original saving throw rather than the modified one. I have tested this by giving horrendous bonuses/penalties at ranges of +-50.

This isn't really an engine inconsistency per se because the engine needs the saving throw values before an effect is applied, whereas something like magic resistance is conditional, and has effect-specific application.

How so? Isn't the magic resistance roll rolled at every AI update, just like saving throws? How is MR effect specific when saving throws are not, even if the preceding effects do not require saving throws?

So, for example, you have a block that applies -2 save vs death penalty and then instant death if not save vs death. The target's current save vs death is 13. The target rolls 15. The target won't die because 13 + 2 >= 15.

Perhaps it's just me, but I can't seem to understand how this theoretically prevents the saving throw from being updated.

-Galactygon

[smeagolheart]

Is it possible to get sneak attack like in 3E (and the option in IWD) where thief's attack from side or back is 1d6 (on up) depending on level.

Could you explain what you mean by "on up" in the text in bold?

Here you go:
(source:
http://www.planetbal.../backstab.shtml )

Sneak Attack and Crippling Strike were introduced with Heart of Winter:

Sneak Attack: This is an optional ability that you can replace backstab with. This option can be enabled in your config program. When a thief uses Sneak Attack and attacks an enemy from the rear or flank, they can do additional damage but one time only. This additional damage is +1d6 every four levels. At level thirty this would be 8d6. Critical hits do not double damage. The thief does not need to be hidden or moving silently to use Sneak Attack.

Crippling Strike: In addition to Sneak Attack, at level five the thief learns the ability of Crippling Strike. When a Sneak Attack is successful, the thief cripples the enemy reducing their hit and damage rolls. This effect is in addition to the Sneak Attack damage. The enemies regain normal hit and damage rolls one turn after the successful Crippling Strike. At level five this causes the enemy to have -1 to hit and damage rolls. Every four levels after, the thief gains minus one to this ability. At level thirty, an enemy hit with Crippling Strike would have -7 to hit and damage rolls.

This was in IWD (and NWN 1+).

Edited by smeagolheart, 28 February 2011 - 09:41 AM.

[DrAzTiK]

Hi Ascension,

Could it be possible to not delete active memorized spells for spellcasters wicth have level drained. It is so boring to reasign spell after cure level drain.

ty

Edited by DrAzTiK, 28 February 2011 - 10:55 AM.

[GeN1e]

Looking at the code, BioWare never meant for things like Spell Immunity to be contingency-able (see CONTINGX.2DA).

I would have guessed the CONTINGX was made to store 'abnormal' spells, those that couldn't fit into the only three available contingency targeting options. Not vice versa.

But if you say it's too much work, then so be it.

[the bigg]

NB: this is a theoretical request: I don't need it for modding or playing, I'm just reporting this potentially interesting factoid.

The algorithm used for applying the effect percentage (0x12 and 0x13 in effect structure in itm/spl files) is completely out of whack, and runs counter to what most modding guides state. I can identify three problems:

1) percentages are both inclusive (effect is applied if prob2 <= rand <= prob1, whereas modding guides suggest that this is prob2 <= rand < prob1).
2) the random number aren't calculated once (when it starts applying the effect block), but every time the engine feels like it wants to recalculate (it never recalculates chances if old_prob1 == new_prob1 && old_prob2 == new_prob2, or if new_prob2 <= old_rand <= new_prob1)
3) probably related to the algorithm in 2, effects are applied more often than what the percentages state.

Test cases:

COPY_EXISTING ~amul04.itm~ ~override~
  FOR (i = 0; i < 101; ++i) BEGIN
    FOR (j = 0; j < 3; ++j) BEGIN
      LPF ADD_ITEM_EQEFFECT
        INT_VAR
          opcode = 139 // display string
          target = 1
          timing = 2
          parameter1 = RESOLVE_STR_REF ( ~Result %i%, run %j%~ )
          probability1 = i + 1
          probability2 = i
      END
    END
  END
  
COPY_EXISTING ~amul05.itm~ ~override~
  FOR (j = 0; j < 3; ++j) BEGIN
    FOR (i = 0; i < 101; ++i) BEGIN
      LPF ADD_ITEM_EQEFFECT
        INT_VAR
          opcode = 139
          target = 1
          timing = 2
          parameter1 = RESOLVE_STR_REF ( ~Result %i%, run %j%~ )
          probability1 = i + 1
          probability2 = i
      END
    END
  END

Tests are run by cheating in and equipping amul04 or amul05. In a ideal world, both cases should both print exactly three strings: "Result X, run 1" to "Result X, run 3".
With amul04 (which groups effects with the same probability), run 1 through 3 are always printed together (because it doesn't reroll if two effects with same probabilities are adjacent), but it prints lines for both X and X + 1 (because the check is inclusive at both ends, and doesn't reroll if the current random number results in the effect being applied); moreover, sometimes more than the six messages are printed (prob X and X+1, run 1, 2 and 3), because of favorable reroll results (I haven't observed a single instance of nothing being printed for a batch).
With amul05, the X/X+1 pairing is still observed, but having result X at run 1 doesn't guarantee that result X turns up at run 2 (because the random number has been rerolled in the meantime). Also in this case, at least one (X, X+1) pair is always printed for all three runs (at least during my experimentation).

[the bigg]

The RESPONSE #X mechanism in BCS files is similarly broken.

Experiment:
[codebox]
response_value = 1 // 10, 100...
<<<<<<<< a.baf
IF
True()
THEN
RESPONSE #%response_value%
ActionOverride(Player1,DisplayString(Player1,~%response_value%0~))
Continue()
RESPONSE #%response_value%
ActionOverride(Player1,DisplayString(Player1,~%response_value%1~))
Continue()
END
>>>>>>>>
COPY - ~a.baf~ ~a.baf~
EVALUATE_BUFFER

EXTEND_TOP ~baldur.bcs~ ~a.baf~[/codebox]

Theoretical results: branch #2 should be taken ~50 of the time.

Actual results:
[code]
RESPONSE # Branch #2 is chosen this % of times:
1 0
10 46
100 46
1000 10[/code]
The # of times branch2 is chosen can be obtained by using 'log dialogue bar' in ToBEx, using sed to extract the printed number from the log lines, and then computing sum(values) - num_runs * 10 * %response_value%

[Ascension64]

Looking at the code, BioWare never meant for things like Spell Immunity to be contingency-able (see CONTINGX.2DA).

I would have guessed the CONTINGX was made to store 'abnormal' spells, those that couldn't fit into the only three available contingency targeting options. Not vice versa.

But if you say it's too much work, then so be it.

Oops, I accidentally deleted my original post because I clicked the wrong button. Anyway, I am actually in error and have a plan to implement this - not easy, but doable and easier than I thought it would be.

CONTINGX.2DA is a strict list of spells to exclude from being selected in the contingency table, sorted by level (column). It doesn't care about targetting.

[Ascension64]

Whenever activating an item with a special ability, doesn't the activation animation in ITEMANIM.2da interrupt and override the previous animation without regard for which frame it's at?

I believe so, but I still say it's dangerous and deserves a WeiDU equivalent of DO NOT USE posted all over it.

Fair enough. But the on-the-fly bonuses/penalties are not applied to that roll, so that succeeding effects compare with the character's original saving throw rather than the modified one. I have tested this by giving horrendous bonuses/penalties at ranges of +-50.

I don't think the engine likes radical saving throws and tries to limit it. I seem to recall the engine just gives a -20 if a spurious number is given for a saving throw. Have you tried something more rational, like set to 0 or 20?

How so? Isn't the magic resistance roll rolled at every AI update, just like saving throws? How is MR effect specific when saving throws are not, even if the preceding effects do not require saving throws?

Sorry, I'll clarify my muddliness. The following are calculated every AI update.

char rndEffSaveDeath; //6282h, 1D20
	char rndEffSaveWand; //6283h, 1D20
	char rndEffSavePolymorph; //6284h, 1D20
	char rndEffSaveBreath; //6285h, 1D20
	char rndEffSaveSpell; //6286h, 1D20
	char rndEffResistMagic; //6287h, 1D100
	char rndDoEff; //6288h, 1D100, test against effect prob1 and prob2

Perhaps it's just me, but I can't seem to understand how this theoretically prevents the saving throw from being updated.

Effect application occurs between an AI update. Therefore, while an effect is being applied, the saving throw itself is always the same (so if the last AI udpate rolled a 15, the throw will always be 15 until the next update). The target throw to beat can change though using effect 32-37 or whichever they were. (so you can have a base target of 13, and get a -2 penalty, making the target a 15). But the actual roll of 15 will always stay the same.

Edited by Ascension64, 28 February 2011 - 05:42 PM.

[Galactygon]

I believe so, but I still say it's dangerous and deserves a WeiDU equivalent of DO NOT USE posted all over it.

What I've described is carrying vanilla behaviour (ie spellcasting) over to attack animations. I'm only intending to use this to disable the attacking animation of select items.

I don't think the engine likes radical saving throws and tries to limit it. I seem to recall the engine just gives a -20 if a spurious number is given for a saving throw. Have you tried something more rational, like set to 0 or 20?

I haven't tested setting saving throws to 20, but this isn't the issue - I was able to get +-50 to work with other timing modes. So this purely has to do with the timing of the effect application, so that the .cre's saving throw stat is updated to the new value before the next effect in the effects list is applied.

char rndEffSaveDeath; //6282h, 1D20
	char rndEffSaveWand; //6283h, 1D20
	char rndEffSavePolymorph; //6284h, 1D20
	char rndEffSaveBreath; //6285h, 1D20
	char rndEffSaveSpell; //6286h, 1D20
	char rndEffResistMagic; //6287h, 1D100
	char rndDoEff; //6288h, 1D100, test against effect prob1 and prob2

Effect application occurs between an AI update. Therefore, while an effect is being applied, the saving throw itself is always the same (so if the last AI udpate rolled a 15, the throw will always be 15 until the next update). The target throw to beat can change though using effect 32-37 or whichever they were. (so you can have a base target of 13, and get a -2 penalty, making the target a 15). But the actual roll of 15 will always stay the same.

What you're describing is what behavior I'm desiring. I'm trying to change the saving throw to beat. Attached is a copy of SPWI304 (Fireball) with 2 effects with a duration of zero: A save vs spells modifier of -50 and a luck modifier of -50. The intended behavior: all .cres always fail their saving throws, and always suffer 60 points of damage. I was unable to get either part to work.

-Galactygon

Attached Files

•  SPWI304.spl   442bytes   147 downloads

[Maanape]

As a dream request, I'd like to know if there's any way to make the party's sight go totally dark when in places unlitted as Ulcaster School or Bodhi's Lair and wilderness areas at Underdark. This would make it much more thrilling and could make room to spells that deal with light/darkness, and would make infravision and divination spells much more useful. This could grant advantage to drows and creatures able to cast darkness spells, as well as kitted priests of Shar. Also, non-living or cold blooded cratures would be undetectable when at complete dark. Torches could be made available items, and they could grant some limited range of sight. On the other hand, carrying a torch would introduce new concerns, as the use of the off-hand to hold it instead of a shield or a second weapon, and the very fact that the party could run out of torches when in a deep dungeon. Well, that's an idea, don't know if it's even possible...

[Salk]

It's very possible I am suggesting something silly (it wouldn't be the first time) but I would like to ask: would it be possible to sync the attack animations so that they are played only when a real blow is landed?

Edited by Salk, 02 March 2011 - 10:53 PM.

[Ascension64]

It's very possible I am suggesting something silly (it wouldn't be the first time) but I would like to ask: would it be possible to sync the attack animations so that they are played only when a real blow is landed?

I think this is, but you don't need TobEx for that. Someone would just need to edit the RNDBASE*.BMP files and change the various pixels denoting the round.

[DavidWallace]

As a dream request, I'd like to know if there's any way to make the party's sight go totally dark when in places unlitted as Ulcaster School or Bodhi's Lair and wilderness areas at Underdark. This would make it much more thrilling and could make room to spells that deal with light/darkness, and would make infravision and divination spells much more useful. This could grant advantage to drows and creatures able to cast darkness spells, as well as kitted priests of Shar. Also, non-living or cold blooded cratures would be undetectable when at complete dark. Torches could be made available items, and they could grant some limited range of sight. On the other hand, carrying a torch would introduce new concerns, as the use of the off-hand to hold it instead of a shield or a second weapon, and the very fact that the party could run out of torches when in a deep dungeon. Well, that's an idea, don't know if it's even possible...

I think this could probably be done without changing the game engine. It would be a matter of having the area script apply a Blindness effect every few seconds if some condition wasn't met.

[Ascension64]

I trust you also saw my request here? I know you tend to keep track of everything, Ascension64 - just wondering if it's feasible.

I think it's feasible. I'm still nutting through the contingency/select spell stuff to be able to mind anything else at the moment.

[Ascension64]

Effect application occurs between an AI update. Therefore, while an effect is being applied, the saving throw itself is always the same (so if the last AI udpate rolled a 15, the throw will always be 15 until the next update). The target throw to beat can change though using effect 32-37 or whichever they were. (so you can have a base target of 13, and get a -2 penalty, making the target a 15). But the actual roll of 15 will always stay the same.

What you're describing is what behavior I'm desiring. I'm trying to change the saving throw to beat. Attached is a copy of SPWI304 (Fireball) with 2 effects with a duration of zero: A save vs spells modifier of -50 and a luck modifier of -50. The intended behavior: all .cres always fail their saving throws, and always suffer 60 points of damage. I was unable to get either part to work.

-Galactygon

Yep, I see. The incremental changes are applied to a sum CDerivedStats that is applied to the main CDerivedStats after all the effects are applied. This is to prevent order of operations issues. It is possible to make the changes directly onto the main CDerivedStats at the disadvantage that order of operations issues would be present (so, a set or % applied after +/- will overwrite the latter)

[pacek]

I'm sure its been brought up before, but what's the possibility of making sorcs and monks kittable?