Talk:Ability Drain

Targeting Allies

Responding to @Anything Random:

There are many other instances of descriptions using the words target, creature, entity, and character imprecisely, so I wouldn't uniformly consider that a bug. In this case, the code explicitly checks for an enemy, and it's reasonable for players to expect an offensive feature doesn't work on allies.

It would also be just as reasonable (if not more reasonable) for players to expect an offensive feature to work on allies. Many, if not most, offensive features do work on allies, including weapon attacks and cantrips (e.g. Fire Bolt, Ray of Frost, Eldritch Blast) which are among the most common actions taken in the game. Exceptions that can't target allies (e.g. Hold Person) at least usually and really should specify that you must target an "enemy" in their descriptions. Even very similar spells, like Wrathful Smite and Compelled Duel, are consistent with their descriptions and behavior about whether they can affect any "target" or "an enemy".

Furthermore, even if inconsistent usage is widespread, that doesn't necessarily make incorrect usages correct. Especially when the terms are already given specific meanings in D&D rules that many readers and players would understand.

So I still disagree, i.e. I think that this disagreement between the description and behavior belongs in the bug section. Quaris (talk) 12:45, 1 October 2026 (CEST)

By my own personal rubric, this seems like it's 100% intentional behaviour on the part of the devs, and reasonably expectable on the part of the players, call it 50% expected. So, overall, I find it leans more towards an understandable hand-wave than a bug.

I asked another frequent editor about this particular change and they seemed to be on the same page with me. If you want to seek more consensus, feel free to ask other editors on discord. Anything Random (talk) 23:13, 1 October 2026 (CEST)

When you say "this seems like it's 100% intentional behaviour": Do you mean the condition only applying to non-allies is intended? Or that the disagreement between the description saying any "target" and the behavior only applying to non-allies is intended?

I agree it's likely the former was intentional; it's the latter that I think is unintentional. Do you agree with that much?

I would rather reach consensus via either of us being persuaded based on the merit of arguments, as opposed to multiple people "voting". It is not possible to "seek consensus" by only speaking with other people. Quaris (talk) 01:27, 2 October 2026 (CEST)

I meant that the the coder who wrote the feature took the time to write and not Ally() at the end of the Conditions. So the former, I suppose. It is likely that the person who wrote the (English) description was not the coder and may not have even been a technical person. The description may also have been based on the initial design from before the feature was even implemented. It's hard to say, but even if you take the position that the wording was not intended based on how the feature actually works, I don't read it as a strict disagreement. That would be like if the feature said it only targeted enemies, but in fact worked on allies. 'Targets' is ambiguous. Another point that recently occurred to me: would you consider this something that should be fixed in a patch? In the context of BG3, I suppose the question is more so if you would expect this to be changed upon downloading a bugfix mod, since there are effectively no more patches coming. That seems, to me, a good heuristic for where to draw the line for a bug.

And, I didn't mean to say that we put it to a vote. The purpose of asking other editors is to get more perspectives and possible arguments than just two people can muster (If you have some time to spare, take a read of the Wikipedia policy on this topic, I try to pull from these whenever situations arise on this wiki where policies aren't clearly outlined or established). Very often I have found third options or novel arguments that swayed me when discussing "edit war" pages (not that I would describe our interaction as such) with other editors. There are also just some people who I know have a deeper game knowledge than me (and pretty much anyone). On the topic of game code, if someone like Joow weighted in on the discussion, I would probably take their opinion as near-truth, since they have a much deeper knowledge of the codebase than me (This issue is quit minor though, so I wouldn't ping them or anything). Anything Random (talk) 02:25, 2 October 2026 (CEST)

Ok, so we agree the description specifying "target" was (likely) unintentional, but we disagree on the interpretation of "target".

The D&D 5e 2014 rules explicitly use "target" to describe creatures who are affected by spells or other effects, regardless of whether the spell has beneficial or harmful effects, and sometimes even use the term "target" to mean yourself. For example, see the D&D 5e 2014 player's handbook, Chapter 10: Spellcasting > Casting a Spell > Targets.

Baldur's Gate 3 also widely uses the term "target" when describing any creature who are affected by spells, regardless of whether the spell has beneficial or harmful effects, and even sometimes for spells who can only affect allies. For example, the UI when casting multi-creature spells calls the creatures (or objects) that are affected by the spell "targets", even for spells like Bless that can only affect allies, or spells like Hold Person that can only affect enemies.

The term "target" is intentionally ambiguous about whether it describes an ally, enemy, or even whether it's describing a creature at all (it could be describing an object). It would not be correct to interpret "target" as only meaning enemies.

Re "would you consider this something that should be fixed in a patch?" (or bugfixing mod): Yes. For example, the Item and Spell Bug Fixes mod includes "Type 2 fixes: After determining Larian's intent for an item/spell, changing its description..." and has made description-only changes before, such as correcting the description of Cloak of Shadows. (This is assuming the description of Ability Drain is what would be changed, but a bugfix mod changing the behavior to match the description doesn't seem unreasonable to me either.)

Stating that someone else, even an expert, agrees with your stance is "putting it to a vote" though, which is not preferable to building consensus through discussion. Even the wikipedia guidelines that you linked yourself state that getting a third party opinion should only be done after discussions have failed, not as the primary way to resolve discussions. I find it very frustrating to have someone use solely an appeal to authority to shut down a disagreement, and I suspect many other people would too. Quaris (talk) 22:59, 2 October 2026 (CEST)

We have 5 whole pages detailing the significant D&D 5e rule changes in BG3, and honestly those pages are far from complete, even just on other wiki pages people add notes all the time going over this or that change from tabletop. Because of that, I never find an appeal to D&D to be especially convincing. Regardless, I agree that that's how the word 'target' should be used, and even matches the dictionary definition. But that doesn't change the fact that it's not used consistently like that in BG3.

Here's a list of actions I could find that can only target/trigger from enemies and use the noun 'target' in their tooltips:

•  Great Weapon Master: Bonus Attack (only triggers if you kill an enemy)
•  Mantle of Majesty: Command (refers to a 'creature' and a 'target' but only works on enemies, same goes for the other sub-spells)
•  Crushing Flight (Owlbear Wild Shape)
•  Mind Blast
•  Lupine Slash
•  Stage Fright
•  Arrow of Many Targets (Despite the name, only works on enemies and objects)
•  Bend Luck: Ability Check Penalty
•  Ichorous Corrosion
•  Coldbrim Chill

Here's some that use a similarly ambiguous/inaccurate term, like 'creatures':

•  Displace
•  Sleep
•  See Invisibility
•  Burrow (Badger) (Only knocks enemies Prone)
•  Force Tunnel
•  Bane (spell)
•  Let the Slaughter Begin (I think allies can technically get the condition, but are immune to the Dazed effect)
•  Confusion
•  Hypnotic Gaze
•  Planar Binding
•  Slow (Someone edited the word 'enemies' onto the wiki, it doesn't say that in-game)
•  Blindness
•  Vicious Mockery
•  Charm Person (Just says 'humanoid')
•  Guardian of Faith (Again, wiki is wrong, in-game it says 'entities')
•  Cull the Weak
•  Executioner (Requires you to kill an enemy)
•  Hellish Rebuke
•  Lathander's Light (Only affects enemy fiends and undead)
•  Howl of the Dead
•  Radiance of the Dawn
•  Destructive Wave (Says 'creatures' in-game)

And here's some that don't mention any particular noun for what you hit, but only work with enemies

•  Arcane Vehemence
•  Radiant Shockwave
•  Splinter Shadow (Must be attacked by an enemy)
•  Sweetplum Gales
•  Dominate Beast
•  Time Thoroughly Bided
•  Synergetic Moments (Requires a condition to be inflicted on an enemy)

This list is not exhaustive, but I think I've made my point. You can understand why I'm hesitant to set a precedent here: it would have a large rippling effect across the wiki. I don't think there's any issue with adding targeting information to a page, but to consider an unambiguously offensive effect to be bugged because it doesn't work on allies feels like a bridge too far to me. Anything Random (talk) 11:06, 3 October 2026 (CEST)

The reason I believe the D&D rules are relevant is that it impacts how many players will interpret descriptions; it would hold sway in the "reasonably expectable on the part of the players, call it 50% expected" factor that you were considering before. This would only/mainly be true when Baldur's Gate 3 does not clearly communicate to players its own different definition of shared vocabulary, and I believe that is the case here.

Re other abilities with similar descriptions and behavior: Doesn't the inconsistent use of "target" (or "creature"/etc., as opposed to "enemy"/"hostile"/"foe"/etc.) being widespread just show that it is more likely that players will become confused or misled by the term? The list of spells/abilities/etc you provide shows players are less likely to interpret "target" as including allies. But I don't think you contest that there are also many spells and abilities that say they work against any "target" or "creature" or just don't specify they only apply to enemies, and they do work on allies (for example, weapon attacks and most cantrips, which are the most commonly used and therefore probably also the most scrutinized abilities in the game); and that evidence shows that players are less likely to believe "target" only means enemies. In other words, I don't think the evidence of the same description being widespread helps make a case for the description being clear (as opposed to misleading).

This is a fairly minor point, but one distinction between Ability Drain and some other actions listed (such as Vicious Mockery) is whether the player is able to learn whether it can affect allies, enemies, or both prior to using it. With Vicious Mockery, the UI tells you when a creature you've selected is an ally and therefore not a valid target. With Ability Drain, the description must do the heavy lifting because the same UI guardrail isn't there to help out (since some, if not most, types of attacks can be made against any creature as opposed to only enemies).

Re "You can understand why I'm hesitant to set a precedent here: it would have a large rippling effect across the wiki." No, I don't understand; what makes you hesitant to apply whatever decision is made here to pages for similar abilities too?

Re "...to consider an unambiguously offensive effect to be bugged because it doesn't work on allies feels like a bridge too far to me." I think you misunderstand what I'm arguing, because I do not believe this standard to be generally true, nor have I (meant to) claim this during this discussion. To try to clarify, I believe that this unambiguously offensive effect that only works on allies has a misleading description, because that description states the effect applies to any attack, when it should instead specify that it only applies to attacks against enemies.

Let's say a tooltip of a magic item claims its effect triggers on a weapon attack. But the effect only triggers on melee weapon attacks, and does not trigger on ranged weapon attacks or thrown weapon attacks. Would you say that description is misleading, or correct? What if ~33% of all magic items in the game did the same, ~33% specify their effect triggers on melee weapon attacks (and did exactly that), and ~33% specify their effect triggers on weapon attacks (and triggered on melee, ranged, and thrown weapon attacks); would those proportions affect your answer? Quaris (talk) 04:27, 4 October 2026 (CEST)

The majority of BG3 players have never played tabletop 5e. Even without accounting for demographics, BG3 has sold roughly double what the 5e PHB has (based on public estimates). That's why I don't subscribe to the idea of a shared vocabulary. We should assume that the people who play the game, and by proxy, who browse the wiki, are coming in with no outside knowledge.

The point I was trying to make with that list was not that the game is consistent with its tooltips or use of language, but rather that an experienced player would consider Ability Drain's behaviour to be within the expected range, and that an experienced player knows not to put too much weight into the exact wording of tooltips.

Whether the tooltip is clear or misleading is not the crux of what determines behaviour to be a bug. In our prior conversation, I said that one of the two criteria I consider when making a bug determination is, "How likely is it that an experienced player would be caught off guard by this behaviour?" Examining that consideration, if there is a similar implementation (in this case, the wording of tooltips) across many different parts of the game, then it is unlikely to be a bug (unless it is truly, unambiguously wrong), because players familiar with the game should be able to expect it. To be clear, if there's a single thing (item/passive/action) that affects the behaviour of many different things, then that could still be seen as a bug.

Looking back at it, I was being far too curt when I wrote the ending of my last message, since I was quite tired after digging up that list, but what I meant when I said that I would be hesitant to set a precedent was that large changes to the wiki should not be made without a large consensus. I'll grant you that people have flaunted this guideline in the past, but for something this subjective, I would hesitate to make such a determination without consulting others. And if you make a change that's so blatantly against the wiki's precedent without changing those other pages, then you're just creating inconsistency and confusion.

The hypothetical you give about a magic item is not far off from how the game actually is.  Gloves of the Growling Underdog is the example that comes to mind, but there are all kinds of features with poorly-written tooltips on list of features and items that affect attack rolls. That's besides the point though, because the question was never if a tooltip is misleading or not, but whether or not something is a bug. I would not consider any of those three scenarios to be bugs. In most cases, the exact wording of a tooltip just doesn't bear enough weight to be considered a bug. It would require something more unambiguously wrong. Maybe if a tooltip said a feature affects spell attacks and it also affected weapon attacks, that might rise to the magnitude of a bug, depending on the context. Anything Random (talk) 12:34, 6 October 2026 (CEST)

Also, to your last point, "putting it to a vote" would be to rule based on the majority opinion, which I am not in favour of. The point I was making about an expert weighing in was secondary to the main point of bringing in other perspectives, and was mainly to give a simple example of how I could be swayed to the other side of the argument. Anything Random (talk) 12:04, 3 October 2026 (CEST)

Then can you explain what the other person's perspective is that you're trying to bring into the discussion? Just mentioning that they agree with you is not helpful.

Re "...to give a simple example of how I could be swayed to the other side of the argument." Then what was that example of how you could be swayed? By simply the other person's expertise being trusted more than mine? Quaris (talk) 04:42, 4 October 2026 (CEST)

How would I know what others' perspectives are if I haven't heard any? If I already knew them, there would be no point in asking, now would there? Also I don't know who you're talking about that agreed with me, as far as I know, no one else has seen this discussion. I mean, you specifically asked not to bring it to wider attention.

The point of asking an expert is to trust their expertise. I personally think that Rydiak can be overzealous with his bug classifications sometimes, but if he's adamant on his position after a discussion, then I would give him the grace of having final say, both because he's an expert on the topic and because he's proven that he takes others' feedback into consideration when making changes to the wiki. Anything Random (talk) 12:43, 6 October 2026 (CEST)