Larian Studios Forums Baldur's Gate III Baldur's Gate III - General Miss! Miss! CRITICAL MISS! Miss! Miss! Miss! Miss!

Just read this piece of a very insightful series of Shacknews articles:

https://www.shacknews.com/article/1...y-engine-era-of-rpgs?page=24#detail-view
I also didn't know that Pillars Of Eternity 2 initially got rid of the grazes -- you don't merely hit or miss on Pillars, you can also graze, which softens the "frustration" for some, apparently. Pillars is purely dice based, no softening or "stabilizing" as such, although it's a Dice 100 rather than a 20, technically.







On a somewhat related note, Sawyer's current project is probably the one I'm most curious about. It's small-scale, thus doesn't have the problem of big audience expectations -- and it's apparently also "non-violent", whatever that means in the end.

Okay, I recorded a small series of rolls for attacks that require saving throws. Although the amount of data is not satisfactory, the picture seems to become clear that the "unreliable" Sacred Flame is not as unreliable as it feels.

Link: https://docs.google.com/spreadsheets/d/1YzEY9Tr7LDgg_k3tP8ra6AeazxRH1z969Qx8dIBgvaA/edit?usp=sharing

Also I think it is possible to critical miss and hurt yourself? I'm pretty sure this happened to me when I was trying to cast that poison beam spell and knocked myself unconscious?

Did anyone every play Rolemaster? You only did most of your damage through high rolls, and you had to look at a big chart telling you how well or how poorly you succeeded. So a normal yet hostile squirrel who made a good roll could eviscerate a fully armored player or the player could get such a massively bad roll they could fall over and impale themselves on a fork. It was awesome. xD

sacred flame never misses for me i have a 100% hit rate with this excellent cleric cantrip. but then i never cast it on someone who isn't busy taking a nice nap. Cant pass a dex save if you can't move.

Nerf Sacred Flame! Overpowered cleric cantrips breaking the game!

I still think that numbers are not based on dice rolls.

E.g. when rolling 2d6 you would expect rolling 6-8 most of the time, but in BG 3 it feels like a random value from 2-12, and actually it seems to prefer low rolls.
4d4 healing potion? enjoy your 4-5 HP healed.

I've uninstalled the EA for the time being. In parts because I've had enough for now. But it also seems that the game is going to change significantly anyway even on the dice rolls once it's released fully... and it remains to be seen if any of those changes are actual improvements. Generally, with such a huge EA audience, I'm a little sceptical. To quote Larian, and that was recorded here all along the thread, "what you're seeing right now is as random as computers can possibly output. It's just true randomness, and we're seeing that it works for some players and some players would like more options and more control over what's happening."

I hope they keep in mind those players also for which it works.

On a somewhat related note, the Football Manager guys have introduced Expected Goals into their game this season. Not only is Football Manager a very popular stats heavy game in general. Expected Goals is a metric that applies every chance a value of how likely it was going to be scored. A penalty, which on average is scored ~76% of the time thus gets and xG of 0.76. Can't wait until somebody posts how the game were broken as he'd just missed three 95% tap-ins or half their penalties two seasons on the bounce. And the only way for that not to happen to anybody would be scripting things.. which actuall football is *anything* but even with the likes of Messi taking kicks.

https://www.transfermarkt.com/fc-barcelona/elfmeterstatistik/verein/131/plus/0?reldata=%262014
https://www.transfermarkt.com/fc-barcelona/elfmeterstatistik/verein/131/plus/0?reldata=%262015

I think the real issue is that dice rolling isn't actually really random.. The science boffins have been working hard and they deemed real life dice to be not actually random.


https://insidescience.org/news/dice-rolls-are-not-completely-random

Now for some this might not be news, but for many this could be why game rng is so "broken" as playing with real dice is not the same and virtual.

Did you even read the article you posted? Here are the relevant parts.


Or how about this bit here?

Now, as a matter of fact, I was actually familiar with this particular fact before being linked this article. I also knew for all intents and purposes, it does not matter for us. Its like when someone tells you that they can "feel" an improvement of 1% chance. No, they really, really, can't. That isn't how the human mind works. The way the human brain interprets probability is it tends to swing things towards extremes. If something falls into the 0-40% range, your mind expects results closer to the 0% point. If something falls into the 60-100% range, it tends to expect results closer to 100%. The problem is, this just is not how probability works. The human expectation of probability is flawed, not probability itself.

If you want to test or measure probability to ensure that the results are fair, the best way to do it is, well, to measure it. As some people have done in this thread. Guess what their measurements turned out? The probability within the game looks more or less to be random. Here is the thing with random chance, random chance does not mean the results would be uniform. You would expect within a random series for there to be some repetitions and within a large enough sample, there will even be strings of numbers that don't look random. What does the human brain do with this? It forgets all the times when you are "lucky," when the dice swings your way and puts that all down to chance, then it remembers all the times that you are unlucky and creates the false impression within your head that the dice are unfair.

If you want something to help convince yourself of this, the digits of Pi are for all intents and purposes random (To be more accurate, its unproven whether they are normally distributed or not, but the expectation is that they are). Here is a website where you can enter pretty much any number combination you like and it will tell you how many digits of Pi it will take to get there. Pi is infinite, they will pretty much all fall somewhere within it.

Yep... that's pretty much it. And this wouldn't change with loaded dice. It might reduce the chances of an "unlucky" series, but those misses on a high hit chance attack, which might even be the crucial decider of loosing or winning a fight, will still feel unfair...

Unless they would code it the way the Gamasutra articles suggests...

Every roll above 90% to be an auto hit.

One thing I'm curious is whether the loaded dice will only apply for the player or also the opponents. At least it then wouldn't affect difficulty as much. (Though arguably, it does: AI no matter how good will never plan as well as a human player. Therefore, actual RNG that keeps the human's plans from going to perfection every time, arguably gives the AI a bit of a better fighting chance).

There are lots of ways to, "solve" the problem, all of which involve eliminating the randomness. For example, if you added a counter. Lets say you enter combat with 3 enemies. The first thing you do is you generate a random integer between 0 and 100 for all 3 of them. Lets say you generated the following:

• 46
• 68
• 44

Now, lets say you have a 35% chance of hitting the enemy. What you do is, if you try to attack 1 of those enemies, lets say you attack the 1st 1, you add the percentage as a number (35) to that result. In this case, the numbers would change as follows:

• 81
• 68
• 44

If when you perform this process, the resulting number exceeds 100, you subtract 100 from the result and you hit the target. If the resulting number is below 100, you miss the target. Either way, you save the new incremented result. So, an example of a hit would be if you targeted the enemy with the number 68 here, it would add 35 to the result and you would end up with 103. After subtracting 100 from that, the counters would be left with:

• 81
• 3
• 44

All of this would be hidden from the player of course. On their side, what you would display is, if the counter rolled over 100, you generate a number within the successful range and if the counter did not roll over, you generate a number within the failed range. You are technically hitting the expected percentage of the time, but there is nothing random about when you hit and miss.

On the subject of AI never being as good as a human, I respectfully disagree, it just depends how much you want to invest into your AI :P You can look at Alphastar vs professional SC 2 players, or AlphaGo vs professional Go players, or Open AI vs professional Dota players, or the many chess AIs better than any chess player for reference.

There is a simple solutions to misses, which players supposedly find not fun.

Difficulty levels: story +8, easy +4, normal +2, hard +0 to all PC attack & ability check rolls and spell DC increased by that value.

Want to hit more often? Don't play on hard.

There is no need for artificial and free of charge high ground and backstab advantage with difficulty levels like above. If such advantage is not implemented, it should also improve chance to hit when other sources of advantage are applied, e.g.
In story mode d20 + 8 is most often better than 2d20(drop low), but 2d20(drop low) + 8 is better yet.

@DanteYoda
If you don't agree with Sharp, record your whole game (each and every roll) and check for yourself stuff is random enough and it's your brain that fails to remember all the events properly.

https://www.ign.com/articles/baldurs-gate-3-early-access-changes-player-data-feedback-larian


Sounds like it's going to be optional, phew.

I think the question is not whether or not RNG is unfair - it's literally RNG and the numbers do not discriminate - but whether or not pure RNG translates into a fun digital gaming experience. When it comes to pen and paper you're sourcing enjoyment not just from the gameplay but from the whole shared experience with others. This may go lengths towards offsetting the disappointment / frustration with a case of bad rolls. However, when you're sitting in front of a PC on your own playing a game there's nothing really supplementing your experience other than progression, which is hindered by bad rolls and will just lead to a frustrating experience. This is why I think it's good that, as mentioned above, there will be an option to configure how this works.

Exactly this!! In a tabletop game we have the benefit of the GM making our "critical miss" rolls cinematic and believable. I had a character in a tabletop game once that had maxed athletics, and yet every damn time I tried to pull off a feat of spectacular gymnastic agility I'd roll ones and land on my ass. In the tabletop context, my GM could explain it as "your raging hangover made you roll your ankle on landing" rather than just "you fail".

In games like this, we don't have a GM explaining away RNG to buffer suspension of disbelief, so constant unlucky rolls just appear as "miss miss miss". In artistic terms to flavour the mechanics you are probably dealing with "enemy got too close so your blow lacked power and the knife skidded off his armour" or "the arrow caught a strong breeze before it hit so it was a glancing blow" or "the acid sluiced off his manly brawny bugbear chest because he's just that tough".

It doesn't always mean "you suck and miss completely even though you are an elite warrior who teaches gaggles of squires how to use a sword every day", it means "target soaked the blow", "your dagger caught on a buckle of his armour but didn't penetrate" etc.

Bloody annoying but that's how D&D works.

Upside; if you want an "Authentic D&D Experience"tm, just let the fails fly! The cinematic fails are pretty entertaining tbh. There's no rule that says you *have* to win every roll, and in D&D the fails are as important as the crit successes.

RNG is RNG... i am playing tabletop rpg with friends, one day i was up in a tree, firing on orks while my compagnons was fighting. I had 85% chance to hit... i shot more than 10 arrows in this fight, at the end of it i calculated my hit ratio... it was 50%... my friends was crazy to see how my roll dices could be so bad lol.

And in BG3 it happen often to miss 3 successive 80% hit attaks ahah... :'(

I thought the graze mechanic in pillars was awful. It just makes combat needlessly complicated and hard to read.

In BG3, I just really want to see the hits that don't do damage actually get deflected from armor and weapons.

Where is the clash of steel ??? Even NWN had blocking animations and sparks flying in 2002 and it's real time.

We Demand More Sexy Clash Of Steel

Yes i read it, basically anyone who has ever rolled dice knows you can "game" them to a certain extent.

Just because you don't believe it doesn't mean its not true.

I never said BG3 wasn't random, i said its too random, the dice in game is not realistic to rolling an actual dice on a table where humans have complete control. Also why are we only rolling D20 in chat why can't we roll dice of all types in combat as well.