|
enthusiast
|
enthusiast
Joined: Feb 2021
|
Now, I'm not a maths person, so maybe someone who is could help me. How does rolling a 5 (a 1/4 chance to miss) equate to a 95% chance to succeed? Even with 2 dice rolls. If you look at this webpage it gives the likelihood of a result for straight d20 rolls, with Disadvantage, and with Advantage. If you look in the box for Advantage on a 5+ you see the probability is 96%. Advantage/Disadvantage Probability
Last edited by Elessaria666; 16/03/21 09:29 PM.
|
|
|
|
addict
|
addict
Joined: Jan 2021
|
This is a good start @Saberem, but you need more than 100 rolls to determine if there is actually a statistical difference between BG3 rolls and randomness. Your non-weighted results produce a Chi^2 of 20.8, and your weighted-results produce a Chi^2 of 24.8. Both of these are below the 95% confidence value of 30.14, and thus we can't conclude that either of these distributions is different than a totally random distribution.
When I add this to @DragonSnooz's weighted-dataset (608 total rolls), the dataset is still indistinguishable from an even distribution (Chi^2 of 16.2, less than the 95% confidence value of 30.14). There are only 2 values that differ from expectation at >3-sigma: 6 (18 rolls out of an expected 30, for a 5-sigma difference) and 17 (40 rolls when we expect 30, for a 3-sigma difference).
...which is a bit odd to me, tbh. This is the weighted-roll dataset, and thus it should be different than an even distribution. Maybe I'm doing the stats wrong? @DragonSnooz, the "Normal Attack Rolls (Player)" data you posted earlier in this page was for weighted rolls, right? Yes, all the data in that post is for weighted rolls. If weighted dice only function to reduce streaks, then it should become an even distribution faster than Larian's true random. Here is the chart from patch 3 (unweighted).
Last edited by DragonSnooz; 16/03/21 09:42 PM.
|
|
|
|
journeyman
|
journeyman
Joined: Mar 2021
|
If I said I had a seed for your goodness of fit, would you salt my validation? Here's the catch! I did the same thing you're doing with a real deck of cards. And I was really actually unlucky for a whole night. Everyone else was using the same deck of cards. And how could they know where I was going to sit?
|
|
|
|
journeyman
|
journeyman
Joined: Mar 2021
|
@DragonSnooz , I thought about this, and I'm going to help you. If you think the dice rolls are non-random, it could be something personal to you. And if that's the case, you're the center of attention for the magical effect. And therefore your response to it matters for all people interested in the product of that effect. P.S. Things Math Professors Never Say!
|
|
|
|
addict
|
addict
Joined: Jan 2021
|
I'm confused, what makes you think I'm in the opinion that dice rolls weren't random? The weighted dice setting is a form of pseudo-RNG. But it still has randomness to it.
|
|
|
|
journeyman
|
journeyman
Joined: Mar 2021
|
There is no room for opinion when the correct answer is that true randomness does not exist. The more I learn about how randomness is a presumption, the more capable I am in knowing what will happen if I do one thing or the other, and the more confused the rest of people will be about why the dice rolls seem non-random. All randomness is stupid. You're replying to Wofenring. I'm Gristly-Knuckle. Well met.
|
|
|
|
veteran
|
veteran
Joined: Sep 2020
|
Yes, all the data in that post is for weighted rolls. If weighted dice only function to reduce streaks, then it should become an even distribution faster than Larian's true random. Here is the chart from patch 3 (unweighted).Ah. I was of the mindset that weighted dice made it more likely to roll a middling number, not reduce streaks. Thanks for the unweighted chart. It's Chi^2 is 23.5 which is still below the 95% confidence value of 30.14. The most extreme values in that chart are 6 (4 times) and 17 (16 times): The chance of rolling at least one number only 4 times is 3.5%. The chance of rolling at least one number 16 times is 0.8%. Neither of these are too incredibly unlikely. BUT! If we combine your unweighted chart with @Saberem's unweighted data, we get a Chi^2 of 32.26 which is GREATER than the 95% confidence value of 31.14. Thus, we can conclude that Larian's unweighted dice rng is NOT random. The most extreme values were 6 (7 times), with a ~1.8% chance of happening. and 17 (rolled 28 times), with a ~0.02% chance of happening. This is pretty unlikely. Math
|
|
|
|
addict
|
addict
Joined: Jan 2021
|
There is no room for opinion when the correct answer is that true randomness does not exist. The more I learn about how randomness is a presumption, the more capable I am in knowing what will happen if I do one thing or the other, and the more confused the rest of people will be about why the dice rolls seem non-random. All randomness is stupid. You're replying to Wofenring. I'm Gristly-Knuckle. Well met. I still have no idea what point you're trying to make. I think we're all in the understanding that RNG is not true random. It's just easier to refer to Larian's RNG as such relative to weighted dice. Larian stated they were testing a form of true RNG in early access to see the players' response. Currently we have a weighted dice option and we're are reviewing how random either are. Sure nomenclature was juggled when comparing different forms of RNG, but I'm still confused by your posts. Yes, all the data in that post is for weighted rolls. If weighted dice only function to reduce streaks, then it should become an even distribution faster than Larian's true random. Here is the chart from patch 3 (unweighted).Ah. I was of the mindset that weighted dice made it more likely to roll a middling number, not reduce streaks. Thanks for the unweighted chart. It's Chi^2 is 23.5 which is still below the 95% confidence value of 30.14. The most extreme values in that chart are 6 (4 times) and 17 (16 times): The chance of rolling at least one number only 4 times is 3.5%. The chance of rolling at least one number 16 times is 0.8%. Neither of these are too incredibly unlikely. BUT! If we combine your unweighted chart with @Saberem's unweighted data, we get a Chi^2 of 32.26 which is GREATER than the 95% confidence value of 31.14. Thus, we can conclude that Larian's unweighted dice rng is NOT random. The most extreme values were 6 (7 times), with a ~1.8% chance of happening. and 17 (rolled 28 times), with a ~0.02% chance of happening. This is pretty unlikely. Math I definitely prefer weighted dice as RNG, I was getting tired of seeing 5, 10, or 17 xD
|
|
|
|
journeyman
|
journeyman
Joined: Mar 2021
|
The point I'm trying to make is that I have my own dice. Larian should just let us type in our own die results. If you thought I was trying to make any other point, then you are wrong. That is the right answer.
|
|
|
|
addict
|
addict
Joined: Jan 2021
|
|
|
|
|
apprentice
|
apprentice
Joined: Oct 2020
|
Hotfix #10 "From now on, loaded dice will only bend RNG in the rolling character's favour. That means you will not be made to miss to make up for a lucky streak of hits." So my understanding that means the target roll affects the outcome of our rolls. I was curious to see how this change affected our roll distribution. Here is my comparison against the same low & high AC target: Hotfix #10 Weighted Dice Target 4 (AC 9 - 5 Proficiency 85% chance to hit)Number of rolls 3 or below: 7 Number of rolls 4 or above: 93 Longest Streak 3 or below: 1 roll Number of rolls 11 or below: 45 Number of rolls 12 or above: 55 Longest Streak 11 or below: 7 rolls Hotfix #10 Weighted Dice Target 12 (AC 16 - 4 Proficiency 45% chance to hit)Number of rolls 3 or below: 8 Number of rolls 4 or above: 92 Longest Streak 3 or below: 2 rolls Number of rolls 11 or below: 42 Number of rolls 12 or above: 58 Longest Streak 11 or below: 2 rolls Comparison Chart: https://imgur.com/a/BXTd7NII could have been extremely lucky when I tested on the higher AC target but the overall distribution was similar to the low AC target, except my rolls below 12 streak was reduced from 7 to 2 which means I never missed more than twice in a row even against the higher AC target. The chart shows that against low AC there is a more even spread of rolls above 4 because any roll above 4 should be unaffected. Against higher AC, the distribution is more above 12 which seems to confirm that the change in Hotfix #10 is doing its job and rolling more hits than misses.
|
|
|
|
enthusiast
|
enthusiast
Joined: Oct 2020
|
The point I'm trying to make is that I have my own dice. Larian should just let us type in our own die results. If you thought I was trying to make any other point, then you are wrong. That is the right answer. Some of your replies actually made me chuckle in combination with your portrait, it's quite funny imagining the figure in your portrait having eloquent meta-magical philosophical monologues on statistical theory. Anyhow, just wanted to add a +1 to the option to add your own dice results. I plan to play BG with the gf and it would be such a nice and social little extra to be able to roll our own dice in the sofa in front of our screen. Def needs to be optional though, no need to needlessly anger any part of the fanbase who might find this sacrilege.
Last edited by SerraSerra; 15/04/21 09:45 PM.
|
|
|
|
addict
|
addict
Joined: Jan 2021
|
Hotfix #10 "From now on, loaded dice will only bend RNG in the rolling character's favour. That means you will not be made to miss to make up for a lucky streak of hits."
So my understanding that means the target roll affects the outcome of our rolls. I was curious to see how this change affected our roll distribution. Originally the weighted dice just had added logic to reduce streaks for all outcomes [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]. The recent hotfix changed that logic to ignore streaks above a threshold, (let's assume it's 10). So [11-20] are allowed to streak but other outputs are not, this will undeniably bias dice rolls in the player's favor. Which your chart shows. We introduced loaded dice in Patch 4 to try to smooth out the extremes of the dice-rolling bell curve. Even with this change, we noticed in your feedback that the RNG wasn't feeling fun for you. We've seen the dice described as being harsh, cursed, rigged and someone said the RNG was downright evil. We want you to have the best experience possible when playing the game and so the changes in today’s hotfix are here to help with your hit chance, if you’re playing with loaded dice. Original Post
|
|
|
|
|