Larian Studios Forums Baldur's Gate III Baldur's Gate III - Suggestions & Feedback Karmic Dice Enabled makes Enemies hit too much

As far as I understood, the "karmic dice" setting should "smooth things out". As there is no ingame description, I expected that this just means that the overall roll distribution stays the same, but it tries to break lucky/unlucky streaks. However, it seems to consistently make enemies hit more than they should, effectively working as a significant AC penalty.

Explanation posted on Reddit: https://www.reddit.com/r/BaldursGate3/comments/zwqaem/psa_having_the_karmic_dice_setting_turned_on/

Data: https://docs.google.com/spreadsheet...JpkP9anhPshxjVinoXwKdYByYhQkhIxm/pubhtml

Suggestions made by the community:

• Disable karmic dice by default
• Add more finegrained settings (karmic dice only out of combat, or both)
• Add more explanation what the effect of this setting is
• Rework the karmic dice rolls

As far as the data indicates, this setting mainly makes the fights faster and more unforgiving, which is especially punishing for newer players.

+1

Very interesting results. I can't remember ever changing the Karmic Dice setting, so that could explain why some fights seemed much harder than they should've. I was planning on giving a Ranger a go with the level 5 cap anyway, so maybe I'll get to that today with Karmic Dice turned off.

they are the old loaded dice and yes Larian changed the script... i don't use it myself so havn't tested it yet

What I don't understand is why are Karmic Dice the default setting? I don't want to use Karmic Dice, it's basically cheating. I wouldn't have even known Karmic Dice was enabled if I hadn't browsed the different settings earlier.

Yes, karmic dice is a very awkward feature. It is on by default, it has no explanation anywhere, it is actually a buff for enemies and it brakes balance by triggering crits twice as often. It should absolutely not be on by default.

Karmic Dice is for the weak!! Embrace those consecutive 1 rolls!!

Sounds like it might be the opposite!

Definitely agree that the impact of the setting should be clearly explained and it should be off by default.

Also, if the OP is right that it makes enemies do more damage that doesn’t sound like intended behaviour. I’ve never played around with the feature but assumed it made no difference on average to the attack and damage rolls, only reduced the chance of a run of low or high results.

My understanding of karmic dice was that one of its purposes was to reduce misses. At the start of EA, the steam forum especially was filled with posts expressing frustration about missing a lot. Because of this, to me it makes sense to have it on by default. More options are always good, though, if some want the option to have it on just for dialogue. There will also be difficulty options in the final game for people to adjust as needed.

Huh. Interesting. As OP notes, they "only" get 150 rolls for each AC, so there's some uncertainty.

Specifically, there seem to be 3 conclusions:
1.) Karmic Dice result in an average roll of ~12.5, higher than the expected roll of 10.5. This is fairly significant, as there are ~650 rolls in each category (Karmic vs non-Karmic) = small uncertainty.
2.) Karmic Dice result in more crits. This also seems significant; in every AC category, Karmic Dice crit more often than non-Karmic Dice
3.) Karmic Dice result in more crits against higher-AC characters.

However, there are some issues with the 3rd conclusion.
1.) Both "AC 23, Karmic" and "AC 23, Non-Karmic" roll sets show anomalously high crit chances.
- For AC 15-21, the Karmic Dice crit chance is 7-10%, but the AC 23 rate jumps up to 15%
- For AC 15-21, the non-Karmic Dice crit chance is 2-3%, but the AC 23 rate jumps up to 8.5%.
Both Karmic and Non-Karmic sets show a jump of ~6%, indicating it's not necessarily true that Karmic Dice alone result in being crit more often at higher AC values.

2.) With only ~150 rolls in each section, the # of instances of any single result is much less significant. A chi-squared test rule of thumb is the expected occurrence of each result should be greater than 5, and 150/20=7.5 which is approaching that limit. The expected 7.5 rolls could easily randomly vary to 10 or even 15.

Edit: Also, theoretically any roll bonuses should be applied to PCs & companions, so you'll roll higher against enemies.

What a horrible idea for a feature in the first place.

Tampering with dice roll results means a 50% hit chance on screen is not actually 50%. Might be a bit difficult to explain such a feature.

How about increasing proficiency bonus instead if the goal is to break missing streaks. The information on screen would at least be accurate.

I totally agree @mrfuji3 - I was just a bit tired of testing/writing down results manually ;-) .

I thought the data might be large enough to encourage someone at Larian to look into it with either better tools (or just look at the code directly) instead of spending more hours, but I also had some doubts which tempted me to collect more data, to reduce the uncertainty.

That's fair. Kudos to recording all those results and doing this analysis!

I've done a bit more analysis using your raw data.

Combining all the Karmic Die Rolls (735 total rolls)

• The average is 12.33 +/- 0.22, inconsistent with the expected average of 10.5 at >3-sigma.
• 18s, 19s, and 20s appear too often (1.5, 1.75, and 2x more frequent than expected for 3, 4.7, and 6.3-sigma differences, respectively.)
• 2s through 9s don't appear often enough (roughly 60-70% as frequent as expected, for ~2-sigma differences).
• Somehow, natural 1s occur ~exactly as often as expected, 1/20th of the time. Perhaps Karmic Dice doesn't re-roll 1s...? This would tie in to the larger issue in BG3 where natural 1's override everything.
• Natural 20s occur about as often as they would in an Advantage system (~10%), but this is not true for other values.
• The full distribution of rolls is inconsistent with a uniform distribution, with a chi^2 value of 111. WAY larger than the critical value of 30.144 for 19 degrees of freedom
• Individually, we also reject the hypothesis that each "vs X AC" set of rolls is consistent with uniform, as each has a chi^2 value greater than 30.144.

Combining all the non-Karmic Die Rolls (634 total rolls)

• The average is 10.42 +/- 0.23, consistent with the expected average of 10.5.
• All values appear within ~25% (<1.4-sigma) of their expected frequency, with most falling within 15% (<0.8-sigma).
• The full distribution of rolls is consistent with a uniform distribution, with a chi^2 value of 12.18. Way smaller than the critical value of 30.144 for 19 degrees of freedom
• Individually, we fail to reject each "vs X AC" distribution as being distinct from a uniform distribution.
• The higher frequency of natural 20s in the "vs 23 AC" data might just be within expected variation (less than 2-sigma difference; e.g., within the 95% confidence interval).

Because this caught my eye, I did a quick 200 Attk NON-Karmic test with a PC with a very high AC 26, just to see if this would bring out any subtle "Non-Karmic" boosts for enemies.

This was against Erna & Yerle in the Temple, mixing it up with ranged and melee attacks. They couldn't hit AC 26, even on a Nat 20, so it went pretty quick.

Just at a cursory glance, I didn't see anything unexpected.

200 Rolls
8 - Nat 20s
8 - Nat 1s

Interesting, I hope Larian sees this because it feels like it is either a bug or a completely unintended byproduct of their reworked loaded dice into karmic dice. As it was explained to us when it was implemented the idea is to have favorable dice odds in conversations and skill checks rather than affect combat balance in such an extreme way.

I don't get it. Why can't they crit?

I'm not sure. If it helps to know, none of them could add more than +4 to their Attk.

Assuming they did hit on a Nat 20 (otherwise your game is bugged, and thus the data is possibly unreliable) that's good to know, thanks!

Combining your 200 data points with the 150 high-AC data points from OP, it seems like there isn't an excess of Nat 20s against high-AC characters using non-Karmic dice. 350 data points still isn't great for a d20, but it's probably good enough.

Edit: just saw your post. Hmmm, you might want to report that as a Bug?? A Nat 20 should be a critical hit, regardless of AC.

I'll definitely do some more investigating.

I did an even quicker Non-Karmic test this morning against an enemy who could hit an AC 26 - the Ogress Grunkkoh in the Blighted Village.

After you kill Buthir, she gets a Rage boost, and I was getting smacked down pretty hard.

30 Attks
4 - Nat 20s
4 - 19s
2 - 18s

Obviously, a meager 30 data points aren't going to tell us anything, but it may be worth pursuing data on this to see if the numbers are skewing for higher-level/story-relevant NPCs, even with Non-Karmic dice.

Maybe, but as you say, 30 data points is nothing. 100 rolls (each value is expected to occur 5+ times) is the bare minimum to say anything approaching conclusive.

I'd be surprised if Larian coded certain enemies to use different rng systems given all the troubles they've had with rng since EA release. But hey, it's Larian, so you never know. If you do record those rolls, I'd be interested in the findings and/or happy to do some stats on them.