Larian Studios Forums Baldur's Gate III Baldur's Gate III - Suggestions & Feedback Might be thinking difficulty level wrong

Sitting around the firepit last night, our discussion turned to DnD and Baldurs Gate 3.

As we talked about the game and our replays a comment was made. "You know what makes replays so easy anymore? The fights are all scripted, so we know exactly what we will fight and where we need to go."

Everyone is talking about a more advanced difficulty level. What if we had a way that all but very specific encounters, became random? Random in enemy location at start? Random at specific enemy compositions. Random chance for a "hero level" enemy combatant or two in the opposition.

When you can do 400+ average damage in a single turn I don't think "random" elements in encounters would change much.
You can also throw potion, help characters laying on the ground and "revive" them so easily.... game over is mostly impossible in the game.

That said, you can try this : https://www.nexusmods.com/baldursgate3/mods/3255?tab=description

I am not saying that there aren't things that could be improved with the AI, or that there aren't things that could be balanced better. But this statement alone accounts for a lot of the complaints in my view.

There's a ton to be said about knowing what the proper critical path is, what is coming, what resources to use and what skills fit best. Even what gear to select. If you know that a given fight is easier if you are X level, or if you have Karlach in your party instead of Lae'zel. Or that certain potions are only useful in certain situations so you don't make too many of any one type. All of these things add up to making the game easier.

I would argue that familiarity with the game is potentially more powerful than optimization. But to give credit where credit is due, if there are exploits where someone can abuse abilities such that a player can regularly do hundreds of points of damage a round, that needs to be addressed as well. But again, this is hardly unheard of. BG1 and 2 had their exploits as well. Heck, if you played a Sorcerer summoner in either game on the hardest difficulty, it was actually easier (if you knew what you were doing) than standard mode.

I think that the Devs should put some effort into improving some of the grosser balance issues. But I also think that there is a certain amount of hyperbolic behavior from the community on the topic. But that's simply my opinion.

For me the game is enjoyable in a different manner when you now what's coming. You can turn into your own DM in a way. Shape things according to your RP to make it harder on you. Use your own meta knowledge against yourself. It's a mentality shift from a blind first playthrough, but well worth it if you want to keep playing.

Combining that mentality with some self imposed restrictions and/or mods will keep thing interesting.

The things I don't like about the scripting (regarding difficulty) is that I know stuff like which one is Mayrina, and I know where the invisible duergar went. Overall, though, I just don't see how a lot of the scenarios could be randomized.

I really wish the AI could learn. If it had a way to process all the millions of combats and optimize itself to beat the players, then the game would be far more challenging. I don't know how many years away from something like that we are, but it doesn't seem like an unreasonable approach. isn't that what AI is supposed to be good at?

I don’t really have input about the difficulty level, but man I wish I had people to sit around a firepit with and talk about Baldurs Gate haha!

For example implementation of a better AI like chagpd whould be cool and whould solve so many problems.
It whould be playing aganst a chess ai. Actually that's hard.

But the main issue that is preventing larian to do this is steam.
From steam all ai powered games are banned. An on steam they sold approximately 11 mill + copies. I doubt that they will risk this.
Mybe with the new Microsoft changes, the industry will have more open mind regarding an ai implementation.
That doesn't mean that they need to kick out ppl from the dev team. But they have two choices work on it or implement ai.
They choose non of them so technically implementing an ia will not rob any ones job.

I really think it's too early to call for specific changes like this, and when they start to increase the difficulty they should start with the obvious stuff - like not being able to long rest or switch party members everywhere, not being able to fast travel out of certain areas, not being able to kill off a section of places like grymforge or moonrise without aggroing the entire building, forcing you to walk into an ambush if you failed the perception check, balancing game items...etc.

Just by looking at the patches, they are really in the mode of fixing the experience so conversations happen at the proper times and certain events are harder to miss or avoid. Right now it's like going to see a movie that only has a rough edit and is in need of some reshooting. Increasing the difficulty is a privileged change - and something they could deliver on way later when the hype has died down to make players want to pick it back up after taking a break.

And even if they do all that, if they don't deliver on a bunch of time saving QOL improvements, I doubt people will be playing long enough to care. Too much time is lost in the inventory screen. It's like 1:1 with actual playing time for me which is ridiculous, I'm eventually just going to get tired of it and forget all the wacky builds and party comps I want to try.

I have to tell you it is the high light of the week. Sit with my best friend since high school and his sons plus occasionally others, all gamers. We talk politics, PCs, General Gaming and of course DnD. I introduced my best friend to DnD in 1979, we have been playing together in one form or another since. MANY tabletop and computer games. Then of course he got his boys playing with us and so on

So once a week; we grab a nice cigar, glass of brandy and then sit around just chilling for 2 to 3 hours.


I am not sure the Steam limitations are in the manner we are discussing. Steam is trying to stop AI "created" games. Games using an AI to power the NPCs or for randomization of none AI created material, would likely be allowed.

Scripted fights are boring for a 2nd time, but there are much bigger problems with difficulty and replayability.

The biggest one being Larian's stubborn reluctance to make the D&D ruleset work properly, i.e. the "op is fun" philosophy. Crazy Illithid powers, magic items, abundant exploits and of course, unlimited Long Rest. They destroy D&D's resource management and bounded accuracy concepts with reckless abandon. Combat becomes a dumbed down smash fest where damage is insanely high, reviving downed characters is trivial and a Long Rest is available anytime you want full powers back.

What the game needs is a Core Rules difficulty mode that imposes restrictions of the 5e system in a meaningful way. Embrace how the system the game uses is designed, not break parts of it. BG3 uses 5e, but the devs didn't really want to because they wanted a flashy frag fest over a smart tactical experience. That shows. And because of this shift in play style and how much damage martial classes inflict, high level magic like Chain Lightning now feels weak and underwhelming.

Conversely, increasing enemy stats and HP is not a real new difficulty mode. It just means you now have to use the most OP builds and resort to Illithid powers. The game would still essentially play the same. So a difficulty mode like that actually limits your options instead of adding anything meaningful. D&D-like restrictions would be better for challenge and changing how the game plays, but I'm sure Larian disagree. Then again, objectively, if you already have 3 difficulty modes that just fiddle with stats and modifiers, why wouldn't you want to add a difficulty mode that does something more profound and actually changes how the game plays?

The amount of companions is a huge hindrance to replayability. I'm sick of them after just one almost finished playthrough. Well, not sick but replaying their stories is not appealing and using the same characters in combat is not a fresh take tactically. Adding just a few more companions would have created a massive increase of different party compositions. And no, respeccing Shadowheart into an Open Hand Monk doesn't help, it's just stupid.

All that said, the game also desperately needs random elements. Random encounters and randomness inside encounters. Random events. The world needs to feel real and unpredictable rather than a movie where you get to choose path A or B.

Well said @1Varangian - but Larian ...can only do Larian, I don't think they are prepared to make the unpopular(?)/uncool choices required to improve combat.

I made an overhaul new thread about Difficulty here for greater visibility because it's a real wall of text that I think is quite exaustive: https://forums.larian.com/ubbthreads.php?ubb=showflat&Number=922161#Post922161

I think there's plenty to talk about.

The game is wildly successful because of it's other merits so it's unlikely they feel any kind of pressure to spend resources on creating a mode where you can't just spam the rest button endlessly. They couldn't figure that one out in 6+ years of development so maybe it will never happen.

Then again it's a D&D 5e game, asking it to play like one is only reasonable. And that would answer the need for a higher difficulty setting at the same time.

I think this game operates very much within the Baldur's Gate milieu, meaning it presupposes Meta knowledge on the player's part. The first time out on any fight is the surprise, second time you tease out a plan, third time you mop the floor, fourth time you try something else just for kicks etc.

BG is effectively Groundhog's Day, but with a D&D spin.

The game needs to remain challenging or at least engaing even when one already knows what's coming, or what's what and what's where. All the encounters are scipted out and the drops are set, cause that's Baldur's Gate 1 mode. You could design everything to be random but that's more like Diablo 1 mode right? Just a very different approach there.

I've played BG1 a thousand times, and enjoying a replay isn't really dependant on the surprise factor for its staying power there. It's something else that makes that worthwhile.

That said, I'm not totally taken with Tactician mode currently, even though it's now my default. I also find most self-imposed ways to increase difficulty are simply adding more time sinks into the equation rather than making a given scenario more challenging. It encourages the player to be even more Meta about everything and then to take it all 4 times as slow. I like the approach that has the Developer providing the playscales and awarding points or achievements based on what they think will provide repeat entertainment and to have some clear defaults for difficulty tiers, but I think they should cater to different playstyles rather than a one size fits all.

For example, I enjoy Tactical Combat, but dislike story gating via Conversation rolls. So my version of "Explorer" would only buff that stuff for the player. Again on a first out, it might be entertaining to lean into dice chaos for convos, but on a replay I often already know what's what there, so what I really want is to explore the thing I didn't see last time. Explorer mode allowing us to back up a conversation and choose a different branch would reduce uncessary save/reloads. Then the game could autosave anytime a character joins initiative and a battle is launched. Explorer mode done that way might not affect enemy buffs/nerfs at all but just grant something like Tides of Chaos for convo rolls, and then Enemy challenge scripting could be its own separate aspect of the 'difficulty' scale.

Again I will go with my default rule of thumb that 6 is better. If they want to have a 4th difficulty tier, I'd suggest that having 6 would be better than 4 - just to remain internally consistent with myself lol

Couldn't have said it better myself!

Although, there are some gimmicky ways to wipe the floor with high lvl spells, but but but! Martial classes can just keep on going, meanwhile sorcs/wizards/spell oriented druids can only use their high lvl spells so often, then they need to long rest/use up their resources to refresh spell slots.

Another thing is, it's easy to beat pretty much everything without using any tadpole powers on tactician too, without ever consuming any tadpoles and just the early game is a little bit harder, but once you start snowballing (all thanks to OP items and somewhat questionable multiclass options) you just don't stop. And yep, the illithid powers entirely break the game. It's just so damn silly.

What annoys me the most about the way tactician is implemented currently, is that enemies don't exert any extra or smarter behaviors. They will try to rush squishies first and that's about it (and it's easy to abuse that behavior to funnel them through some nasty surface spells like cloudkill). I just wish enemies had access to the same spells AND abilities as player characters, but nooo. Nearly 2 full runs on tactician (roleplaying runs too btw) and I have never seen an enemy cast any lvl 6 spell, or 5, or 4. I only once saw a fireball. 99% of the time it's just counterspell.

Edit: What I miss the most in BG3, in comparison to the originals is day/night cycle. Especially the different behaviors of NPCS and different enemy types bound to the time of day. But as BG3 is structured currently, I don't think we'll ever get day/night cycle.

I think the idea to have Balanced key of "Core" is the best approach too, because then you can go either direction from there. Meaning if you want to make the game more challenging for your vets or simpler for your new players, you can do that more easily when the foundation is sorta set in the touchstone that way. Players who are transitioning from the table top to BG3 or vice versa, could have that mode that they can go to in order to learn how it's all been translated. Then venture from that basis into ways to make that experience more fun for the given player, whether that's making it easier or harder, or just better suited to whatever group is playing right?

Examples that come to mind immediately would be anything in an S tier category on the Spell list currently. Stuff like Haste and Spirit Guardians and Counter Spell at the first big power spike in BG3, cause those just run roughshod over other things and set the whole play pattern across the board. They're also harder to nerf once you get everyone all used to using them in particular ways with limited downsides at the base level of play, since they then define optimal play and optimal builds. Other spells which do similar things, but where the overlap here just gets lost in the shuffle, as one thing becomes way more mechanically useful at the expense of everything else. Like Counter Spell is easily the best in class right now, because it takes the opponent off the table on the fly, but then you also know what they're casting and at what level, and the practical result in-game becomes more like table tennis and sorta lackluster. Pre-buffing Haste has a blitz effect across the board, cause it affects just about everything else in the playpattern with the action stack. Granted that is super BG in it's way, with that need for speed (boots) and so I kinda let that slide, but again the idea of stacking that with Twinned Meta and like before you know it's melting all the glue that holds the whole thing together lol.

I'd like to have some sort of mode that gives a de-buffed core rules spin on any big change that was made. The whole idea of borrowing that type of grading tells us how much video games have caught up to the TT experience, and this game is trying to position itself as exactly that sort of hybrid. I guess maybe in the same way BG1 did that for RTS of the era. I think it's working cause cinemascope is a good hook and the basic system of combat and the core mechanics of D&D are so transparent. The transparency is probably more important than anything else, cause that's what you get with a dice game and shared systems and the thing that differentiates a D&D pathfinder vampire white wolf sort of game from a more custom system. Faerun as a setting does something similar for the narrative cause it's kitchen sink, but also transparent, meaning anyone can just look up those sourcebooks or wikis to figure out the broader context.

ps. I also really like the idea of randomization in structure for a game based on dice. Right now the randomization is pretty much positional and based on PC/NPC movement before joining initiative. I think they could randomize encounters and that would work, and be in keeping with the BG1 spirit, but it'd have to really be done right. I think things like starting location could probably be randomized or available party comps or strongholds and things of that sort, basically randomizing sequence rather randomizing the total number/kind even if that departs from the BG formula somewhat. Waylaid by enemies was a hallmark too, so they could try for that I think with some grace to take a few liberties on how it's re-imagined for a BG3. All that stuff could probably work, but it might only be for players who vibe on that sort of run. Usually I think that would work better as a gauntlet style dungeon crawler or something more Icewind Dale 3 if that ever were to happen, but I could see it here. Just needs to be handled with care so it doesn't spoil the groundhog day thing overmuch.

I going on my second run with a spore druid urge.
I inserted on the top of my rules what was missing from Zentu's Dwarf rules. So like no death /no save scuming/ no tadpole / limited resting /no respecing /no multi classing /no elixir/ solo/ only use lockpick one try if it fails it fails.
All in all to add no elixirs aren't really an issue actually its just makes the game slower. Also lockpick makes the game slower.i mean I can still trow down the chest from a high roof.
Currently I had zero issues.

I think the biggest issue that makes this game easy are wendors. They are constantly refreshing they stock on resting and on lvl up that makes healing potions unlimited. Also I find them too cheap. After act 1 its not even essential to use the vendors that's mean that the difficulty curve are falling down dramatically.
At act 3 a defensive play style cannot be even hitted any more. And this happens all because of critical immunity. Another ridiculous problem.

All other things are just cherry on the cake because if I can do it there is at least 1 million people that can do it as well. Because I am not a good gamer for sure I am not the top 10 %.
I am just playing w/o looking for guides or broken items.

Greetings adventurous friends and lovers of the Forgotten Realms, I've been passionate about this universe for over 20 years and a big fan of the old Baldur's gate.
I'd like to thank Larian for the passion perceptible in his jewel and I hope we'll have more content later. I'll do a post about it later.

However, as indicated in the title, I'd like to add my 2 cents to the question of the game's difficulty. Let's face it, with some playing experience and a good understanding of the rules, it quickly becomes a walk in the park.
I've finished the game in Tactical mode with numerous difficulty mods, which gives me some idea of what needs to be corrected/improved.

I think we'll have an Honor Mod like Divinity 2, but I don't think this will represent an improvement in the intrinsic challenge; it's more a bonus that can be activated, rather than an improvement in the fundamentals.
My comments therefore concern an improvement in difficulty for a MOD HIGHER than Tactician, even if certain elements can be used in the lower difficulties.

In my view, there are 3 main areas of focus.

Let's go!!

I/ Enemies.

- Well, it's a sensitive subject, but he's the ogre in the middle of the corridor, and I can't ignore him because he's drooling. There's clearly room for manoeuvre in the AI, particularly in its movements in relation to attacks of opportunity/ the management of area control spells (moving inside, bypassing, assessing the threat).
The rest is honest (targeting weak PJs, coordination), but the first battles of Act 1 are much more polished.

- A simple, basic improvement: better stats (HP, saving throws, CA, actions) as already offered by mods.

- New abilities, gifts, items and consumables for this difficulty level. For example, all Bhaal assassins and rogues should have the sneak attack. Every opponent should have the abilities of their class as a player, including gifts. Similarly, they should systematically have consumables (inducts, poisons, potions, elixirs, bombs) that they use in combat.

- Additional enemies during encounters in higher difficulty modes, to make combat even tougher.

- Additional phases for the main bosses. The fight against Ketheric is, frankly, a good model. The fight against Orin, on the other hand, is rather disappointing. I want to feel both our bodies covered in sweat, I want to see her shapely legs and murderous hands trying to eviscerate me. I want...
Okey I digress, basically a phase where we face her in her humanoid form in killer mode with unheard-of DPS.

I'll give you another example of possible improvement with the fight against Gortash:
Frankly, the fighting mechanics are good compared with Orin, but there's just one thing missing to make this confrontation really stand up to the test.
I'm thinking of 2-3 modifications to make this fight truly legendary:

- Remove the defective shield system in tactician+.
- Add Baine disciples, there are currently two consuls and a black gauntlet. I'm thinking two consuls + two black gauntlets and a black fist as captain. It would be logical for Gartosh to have a praetorian guard worthy of the name.
- During phase 2, make the Manifestations of Tyranny active, why not with a debuff or control spell (domination, for example).


II/ Character Building.

You're still here? Okey I'm past the "2 cents" stage, I'm actually doing a full bank transfer.

The other real problem lies ahead of the confrontations, in the very construction of our characters. In all honesty, they're demigods, and at level 10 you can fraternize with Divine Chosen without batting an eyelid, and have a mind-boggling array of magic. My aim is absolutely not to frustrate, but to rebalance the walls of the system within which we can evolve and do as we please. Because at the moment, frankly, it's open bar, including with regard to DD5 rules.

- There are some real bugs to be ironed out, such as the absurd bare-hand and projection damage, or the fact that a priest can only learn mage spells at level 1.
Solution: a mage/priest multiclass must not be able to cast mage spells beyond the abilities of a mage of the same level.

- Unlimited use of scrolls by all classes. So it really hurts me as a wizard when I see a Barbarian with 8 Int casting a level 6 spell between two burps.
Solution: A class that doesn't master spells can't use scrolls (perhaps include a GIFT that would allow this up to level 3 scrolls, for example, but with a Stat condition for DD).
An exception is made, of course, for thief or warrior specialties giving access to magic, who could obtain this ability in addition to the rest of their gameplay.

How would the spell's DD be calculated? Simply like a normal spell: 8+associated char (SAG / INT / CHA) + mastery bonus ONLY if you have access to this level of spell.
It seems logical to me that a young magic student just out of his magical duffel bag party would have difficulty throwing Arrêt du Temps.

- I know I'm going to make people cringe because we used to make crazy combinations, but Multiclassing without conditions has become crazy. We need to reinstate feature conditions as in the official rules.
Nothing insurmountable really, but it would already have a great control filter. I would remind you, of course, that this option only applies to the MAX difficulty.

- By the same token, there should be conditions for Gifts, either carac or race/class. This would make it possible to add missing gifts and give more unique personality to our choices. (Incidentally, I have a clear preference for unique racial bonuses as in EA, rather than the levelling Tasha rule we have, but that's secondary and another subject).

- The CA bonus when using a shield should not be applied passively when equipped with the ranged weapon.

- Regulation of the possibility of equipping magic items, at the very least the inability to combine items whose bonuses are cumulative or play on the same aspect.

- A time limit on the use of elixirs and potions, not a "duration until the next long rest".

- ILLITHID POWER ARE TOO STRONG, EVEN IN A ROLEPLAY POINT OF VIEW.

A major shortcoming of the game is the total absence of consequences for reputation, morality or any kind of limitation.
In fact, it's all about the powers that be, without any quid pro quo. This is a major problem, and a system for limiting the number of powers according to carac or rolls or associated malus, for example, would be really useful.


III/ Gameplay.

Yes, I know, it's long. It sounds like a lecture by an arch-mage historian... Good logic because it defines me very well!
The last point concerns the gameplay itself.
- There's clearly a problem with XP, I know it's debatable but the progression is too fast and the accumulation of this point with the rest makes the game like a walk in the park. What's more, it's strange to reach almost max level at the start of act 3. Personally, I finished the game with a surplus of unused XP. This also makes completing quests a bit empty. So either increase the levels or divide the amount. The ideal would obviously be to have more levels laugh and thus smooth out the game with, for example, a max level of 16. Okey Larian, it's my Christmas Wish wink.
This would also make it possible to add DLC during the course of the adventure, even if it means reworking and readjusting certain high-level battles.

- The price of magic items is absolutely trivial, especially some MUST HAVE items sometimes selling for 80 or 140, while some armors are several thousand golden coins.

- Accentuate the limitations of Rest with a requirement of 120 provisions (as proposed by a Mod) or possible access only to teleportation points.

- Well, there's clearly a problem with two spells:

1/ Haste is totally abused, at the very least you should limit the number of turns.

2/ The greatest abuse concerns the Sphere of Invulnerability, which is objectively a God Mod and does not correspond to the spell in its original version. It should be a protection against spells, not a zone of invincibility where you can act on the outside world.
At the very least, you should be unable to impact the outside, the number of turns should be much lower (2 turns max) but ideally it should be like its paper version. A spell-absorbing zone with why not a quantity of spell damage hp that can be absorbed before it disappears, either at the end of its duration or through exhaustion of hp.

- Finally, higher-level summoning spells deserve an adjustment, either in the form of a duration limit or a concentration roll per turn for the summoning of Exteriors endowed with consciousness, such as Angels or Demons. Walking around Baldur with an unlimited Deva as if you were taking a poodle out for a piss is quite...surprising.


So this time my thesis is finished, I'm looking forward to your opinions and if you have any additional points to add. Hopefully a Larian member or Moderator will stop by wink

See you soon in the Kingdoms, I'm going to make a post almost as complete on a third-list of cut content.

Archmage Florian DaKatarn