Larian Studios Forums Baldur's Gate III Baldur's Gate III - Technical & Gameplay Problems Ratio of Failed rolls is absurd/Flawed RNG

The ratio of failed rolls in this game is absolutely absurd. Tracking the number of failed rolls I have had, you are 75% guaranteed to fail a roll if it requires a 5 or lower. I mean for gods sake I failed a roll that required a 2... a stinking 2... and I roll a 1.

I've started choosing skills I have no proficiency in because I will succeed more often. The random number generator is not really that random or balanced. I roll 6 or lower 75% of the time and my skills don't seem to even make a difference.

That sounds perfectly right. If you require a 5 or lower (so, a 1, 2, 3, 4 or 5), you have a 25% chance of success, and thus a 75% chance of failing.


That happens. About 5% of the time.


Skills only affect your roll target. Say, if the Difficulty Class (DC) of a skill check is 11, and you have proficiency in the skill (giving you a +2) and 14 in the associated Ability Score (giving you a +2), you have a total bonus of +4. Instead of computing 1d20+4 and checking whether this is greater than or equal to 11, the game rolls 1d20 and checks whether this is greater than or equal to 11 - 4 = 7. This 7 is the roll target. Skill doesn't affect it : the roll target has already factored in your skills.

Now, if you must roll 7 or more to succeed, you fail if you roll 6 or less, which you have a 30% of experiencing. If you have data to support the fact that you roll 6 or less 75% of the time, I'd be happy to see it.

Random number generators are not super difficult programs to write. So I think a serious amount of data would be required to test the random number generator they are using and reject the hypothesis that it's a good one.

Does this game have some kind of log file that I can track rolls? I'll be happy to start tracking and provide all the data you want.

Bottom right corner in game, there is an expandable combat log that shows most of the rolls that take place and the modifiers that are added/subtracted.

This is what I am talking about. I started my game over and made the exact same choices as before. I even put my stats in there for you to see. 80% failure rate on rolls doesn't make for a very enjoyable game. I will continue to keep track. The log in the the bottom right corner does not give rolls when talking to people. I also have critical miss and misses quite frequently.



Roll reason Skill Target Rolled
Investigating egg Investigation 9 2 Failed
Remove Brain Investigation 9 8 Failed
Remove Brain Dexterity 7 2 Failed
Pod on ship Arcana 6 7 Succeed
Dying Midflayer Persuasion 15 8 Failed
Free Le'zel Deception 10 5 Failed
Astarion - Roll Away Dexterity 7 4 Failed
Scared Boar Animal Handling 6 13 Succeed
Druid Camp fight Persuasion 15 5 Failed
Talk to Rat Animal Handling 6 2 Failed

Race: Wood Elf Class: Ranger
Str 12 Skills
Dex 17 Magebreaker
Con 13 Outlander
Int 8 Wasteland Wanderer: Fire
Wis 15
Cha 10

The die can be fickle sometimes friend. My party has had sessions that our DM designed as some easy combat for us, but turned into characters nearly dying because we couldn't roll... well to save our lives that night. On the flip side, sometimes in boss fights that were designed to push our limits with an intense chance for a TPK, we rolled multiple criticals and wacked it around like a rag doll.

Sometimes you just gotta pray to Tymora and hope for the best.

You're being clever but you know what he means. If the minimum you must roll is five or lower, he's saying it feels like the stats are 75% for it's 75% against.

I combat the RNG gawd with my "Finger of Reload".

I don't know the stats, but it does seem like there are some encounters the chances of failing the roll is built in, regardless of what you "need" to roll.

Don't get me wrong. I actually don't mind taking my losses with the die. There is a qualitative difference between this and PnP gaming. The DM can make the losing roll just as interesting (or even funny). Or even when you are playing with other human players, there is a way to have each other's back. There is a human flexibility that the game does not (cannot?) have. There is a way that the roll isn't the final say, even when it is.

I don't know if that made sense.

That said, I'll give Larian props that failing most throws isn't game altering. Or rather making the roll is more game enhancing. Now if he can make losing the roll also game enhancing.

Joe

As an actual Masters of Statistics, let's do this.

1. You must have 100 or more d20 rolls. Not the pass/fail value, the actual roll result. For practical purposes, the actual count needs to be a multiple of 20, so 100, 120, 140, etc. rolls.

2. You must pick them in an unbiased manner, i.e., you can't wait until you have a set of several low (or high) rolls in a row and then start keeping track. Or sample some, then ignore some, then start sampling again. Or run multiple tests and choose the results you want. There are LOTS of other ways to consciously or unconsciously bias the results, also, too many to go into here. Suffice it to say that if you are actually looking for a specific result, you are more likely to find it.

3. Go to https://www.icalcu.com/stat/chisqtest.html

4. The expected values for 100 rolls would be 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 in their format, or 5 of each number. Or all 6s for 120 rolls, all 7s for 140 rolls, etc.

5. The actual value would be the count of 1s, followed by the count of 2s, followed by the count of 3s, etc. So something like 4 8 3 5 6 3 5 6 10 4 4 2 0 6 5 8 6 4 4 7 would indicate 4 1s, 8 2s, 3 3s, etc.

6. Stack the rows as it shows in the site example. (Or if you want to use my sample data, just copy and paste from below.) Calculate.
5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
4 8 3 5 6 3 5 6 10 4 4 2 0 6 5 8 6 4 4 7

7. The P-value is the likelihood that the actual rolls come from a completely unbiased die. By convention, if it is 0.05 or less, you have significant evidence that the rolls are biased.

8. For an interesting thought exercise, how do you think the numbers above come out? Write it down. Then do the calculation. Surprised at the result?
9. Stop griping until you do the actual test.

There are lots of other tests you can do looking for various things, but the number of ways to do things wrong far exceeds the number of ways to do things right.

I think everything is correct. What I think is wrong is the percentage associated with chances to hit based on the modifiers. I think its actually lower. Not sure what calculation there are using but its off a bit. 50% to hit, is not really 50% to hit but more like 25%.

To be fair, the rolls are not truly random - and your computer lacks the ability to generate genuinely random numbers. You need a quantum computer for that - they run in the millions of dollars right now and are incapable of handling a program of this size. Your rolls are based on the internal measure of time on your computer's clock - so it is your fault for rolling at the wrong millisecond.

Can you give an example to support this? If possible, linking a screenshot of the percentage, your modifiers, and the enemy AC?

It's pretty simple to convert a # needed on the d20 (even considering advantage/disadvantage) to a percentage. I'd be shocked if there was an error in that calculation; it's much more likely that Larian implemented some weighted/weird rng than this.

AFAIK, the only percentage displayed incorrectly is when the games says you have a 100% chance to hit. This is due to rounding 0.9975% to 100%, as opposed to an actual calculation error...

Funny things - dice.

Wife and I were fighting the Hyaena Gang.
We weren't exactly well prepared and got slaughtered.
Reload
Manage to turn the "leader". He/It helps, or tries to help. Get absolutely slaughtered by some bow wielding former friends. We die as well.
Reload
Manage to turn the "leader" He/It helps out. And he whacks them all. Because of constant shifting "who gets to fight next" wife and I get to attack ONCE in the whole fight.

And I have had plenty of rolls where I'm thinking - "hmm this is because we have to fail on acount of the story line".
On the other hand I have had SUCCESS in four or five successive rolls and after a Reload I failed all.

Definitely makes for an interesting game. I actually really love that I can't be certain to win a roll, and always have to make allowances for the possibility of failing.


------------------------------------------------------------------
If Superman is so smart, why does he proudly stick his chest out and deflect the flying bullits, but still duck when the criminal throws the empty gun at him??

Let's see how tables come out... Answer: I ~!@#$%^&*()_+ hate that web browsers automatically collapse multiple spaces to single ones. I had to use spacers per Baraz's suggestion.

Target Flat - Advantage - Disadvantage
1 ---- 100% - 100.00% --- 100.00%
2 ----- 95% -- 99.75% ---- 90.25%
3 ----- 90% -- 99.00% ---- 81.00%
4 ----- 85% -- 97.75% ---- 72.25%
5 ----- 80% -- 96.00% ---- 64.00%
6 ----- 75% -- 93.75% ---- 56.25%
7 ----- 70% -- 91.00% ---- 49.00%
8 ----- 65% -- 87.75% ---- 42.25%
9 ----- 60% -- 84.00% ---- 36.00%
10 ---- 55% -- 79.75% ---- 30.25%
11 ---- 50% -- 75.00% ---- 25.00%
12 ---- 45% -- 69.75% ---- 20.25%
13 ---- 40% -- 64.00% ---- 16.00%
14 ---- 35% -- 57.75% ---- 12.25%
15 ---- 30% -- 51.00% ----- 9.00%
16 ---- 25% -- 43.75% ----- 6.25%
17 ---- 20% -- 36.00% ----- 4.00%
18 ---- 15% -- 27.75% ----- 2.25%
19 ---- 10% -- 19.00% ----- 1.00%
20 ----- 5% --- 9.75% ----- 0.25%

How to use this? Example: Your enemy has a 12 AC. You have +4 to hit. That means you need to roll an 8 (=12-4) to hit. Find the '8' in the first column.
---Without advantage or disadvantage (flat), you hit 65.0% of the time.
---With advantage, you hit 87.75% of the time.
---With disadvantage, you hit 42.25% of the time.

For targetted spells (roll to hit, but no save allowed), it works exactly the same, just substitute your spell +to hit.

Example for spells with a save: You have a 13 spell save DC. The spell saves with DEX and your opponent's DEX is 15, which means +2 to save. They have to roll an 11 (=13-2) to save. Find the '11' in the first column.
---Without advantage or disadvantage (flat), they save 50% of the time.
---With THEIR advantage, they save 75% of the time.
---With THEIR disadvantage, they save 25% of the time.
Reverse the last two percentages if you are counting YOUR advantage, so in this case it would be 25% save with your advantage and 75% save with your disadvantage.

The plusses and spell save DC are available on your full-screen character sheet. The enemy's AC and stats are available by examining them (right-click).

The display rounds that to the nearest percent, so back-calculations are a bit more tricky.

- You could add separators like __ -- ||
- The % in the first column all end with .00 which you can remove to simplify the view.

Target Flat Advantage Disadvantage
1 ----- 100 % | 100.00 % | 100.00 %

It's luck.

Sometimes I wasn't able to beat easy mobs, next time I was able to get rid of phase spiders and matriach spidy.

Sometimes I fail persuasion rolls with target 5 or less (I recall lossing to a target 2 ), sometimes I win persuasion rolls against the unthinkable (I recall a winning roll against a 19).

I get the frustration (specialy when in a fight bad luck means a lot of miss or 1 damage point) but I think is just bad luck [I'm a master to natural 1s rolls. I remember my first tabletop play with Vampires The Masquerade, my first scene was meant tobe be: awake, explore your lair start exploring the neirghbourhood, it ended with the lair burned down, the master of my vampire pissed of, ambulances and firetrucks at the door and my character running from the police and the mob. All the result of an epic trail of 6 consecutives natural 1s ]

The one roll issue I have is the illithid wisdom ones that require you get to 1. That's a failure by the DnD rules. I got a 1 on one of those and passed - why bother rolling at all if there is no chance of failure? I'm not going to get more content if I roll a 20 which is the only reason a DM might try something like that.

Well, it's not forever, you'll have to pay for it with your brain. >D If you use a worm, you already conditionally get more content, but pay for it.

@VeronicaTash

To-hits of 1 are automatic failures; skill rolls of 1 are not. That was clarified by WotC. Same for automatic successes: to-hits of 20 hit; skill rolls of 20 do not work if the DC is greater than 20.

My patch 3 campaign now brought me back to the blighted village well and the ettercap + phase spider battle.
Party was completely wiped out two times in a row now, by not being able to hit anything, except for Gale's magic missiles.
I'll retry a third time another time maybe. Pffffff...... (I already won this fight in earlier games, it's really rotten rolls here)
You can't really do much to prepare for this battle, the spiders are always in the advantage at the beginning.
And then continually missing with hit-probs between 56% and 75% is quite frustrating. A real-life DM would probably find some way to save the campaign.