We use Cookies to help personalize and improve Roll20. For more information on our use of non-essential Cookies, visit our Privacy Policy here.
Accept
Advertisement Create a free account

algorithm broken?

For the last month the rolls seem to be broken, im getting upwards of 3 crit fails and rarely above a 9 on a d20. how can it be so consistent giving me bad rolls and critical fails? did they break the system recently?
I suggest you look at the rolls in total....I felt the same way about a month ago then I pulled out the rolls from one of my sessions and it averaged&nbsp; out to 10.36 (a little low the average should be 10.5). But looking back at my rolls, I had high rolls on things that did not matter very much and low rolls on most the important things. So you should be able to see everything Also here is the site with all the rolls <a href="https://app.roll20.net/home/quantum" rel="nofollow">https://app.roll20.net/home/quantum</a>
Yes very fair rolls, 4 d20s and 3 crit fails...
1593440546
That's far too small a sample to tell anything.
1593441806
The Aaron
Forum Champion
API Scripter
You're falling into the Gambler's Fallacy . The odds of rolling a 1 thee times in a row is only 1:8,000 (0.0125%).&nbsp; The odds of rolling three 1s and some other number in four rolls is better at 19/16,000 (0.11875%).&nbsp; By comparison, your odds of winning a conventional 6 ball lottery are 1:13,983,816 (0.0000071511238%), but people still play. A given explicit sequence has low odds, but the odds of your next roll being a 1 are always 1:20 (5%).&nbsp; Dice are independent.&nbsp; Many people believe Random means evenly distributed.&nbsp; It does not.&nbsp; A deck of cards is evenly distributed. If you keep drawing from it, you will get each card a single time.&nbsp; Dice have an equal probability each time you roll them of producing any of their possible values.&nbsp; Because the probability is evenly distributed, it is statistically probable that they will eventually produce a nearly even distribution across millions of samples—but they are not guaranteed to.
1593450128

Edited 1593450201
Andreas J.
Forum Champion
Sheet Author
It's bad luck and confirmation bias. You know it's pretty likely that someone rolls 3 ones in a row every now and then, considering the large amount of rolls being made. And if you'd track your rolls over a longer period of time, you'd likely notice the results evens out. Anyway, here I the general rebuttal I've posted the last few times someone brought this up, as this have been discussed more than a few times: No, the dice rolls are not broken Human intuition on randomness is flawed, and here is an attempt to explain why Roll20 has a page to show how random the rolls are. The thing about randomness that people don't often get, is that streaks" of "good or "bad" luck is to be expected. Truly random doesn't feel random. Steve(Original poster) says: But there is something bugged about the way the rolls are distributed over a short period, like 200 rolls That is unlikely, but expected to happen to someone, at times, given the large number of rolls made on Roll20. You know how there exist people who have got hit by lightning more than once, or have won the lottery more than once? It's unlikely to happen to any given person, but the chance of it happening to someone , given large enough pool of samples, is extremely large. The chance of something happening even at small scale is something we don't have an accurate intuition about, for example the Birthday paradox , which is the statistical oddity that with a random group of 23 people, it's a 50% chance at last two people share a birthday. Math &amp; Science behind randomness Gambler's Fallacy - wikipedia Law of Truly Large Numbers - wikipedia - "With a large enough number of samples, any outrageous (i.e. unlikely in any single sample) thing is likely to be observed." Independence of Events Probability, Random Events, and the Mathematics of Gambling - 20 pages Then there is also the fact that when the results are given instantly, it feels less random than if we see numbers flutter a bit before they appear. Slot machines and gambling games likely uses this trick of slowly showing the result to feel more random and increase the anticipation for the result. Psychology behind randomness (and why we're bad at perceiving it) Randomness is Random Numberphile video on how humans think randomness looks like Apple Made Their Shuffle Feature Less Random, to Make It More Random - DaxThink.com Humans have hard time with randomness - Daily Mail article Why random shuffle feels far from random - Independent UK article The Near Miss Effect - Wired - why seeing near misses feels better(and why physical dice might feel more random) Why random hero selection doesn't feel random - LoL wikia rambling If you watch the " Randomness is Random " video, it demonstrate quite well how we think streaks of "good" or "bad" luck are longer and more prevalent than if we'd try to make a random sequence of numbers ourselves. That's why some random shuffle features on music playlists aren't always completely random, as they might prevent longer streaks of songs from appearing in their original order, or bias the system to try not to have too many songs from the same artist/album to play back to back. Previous Forum Threads on the topic This come up every now and then, and is repeatedly explained every single time. Here is a fairly comprehensive list of previous discussions on the topic: Algorithm broken? (this thread) Dice roller issues? (Previous Thread) Roll20 RNG Testing Dice Are Not Rolling Randomly Random Rolling numbers maybe not so random? Roll20 Dice are not True random (Pro forums) How random is quantum roll random number generator Dice results broken? Spring Update: Meh (users comments how pseudorandom where good enough even before Quantum Roll, and think it's overkill) Multi-dice rolls averaging out? Three Critical Failures In A Row! Dice Rolls - Come in Three's fix die roller Questions about Quantum Roll "Quantum" rolls aren't. Dice roll randomizer (6 years ago, from before Quantum Roll feature, but explain why normally created randomness is already pretty good) Random dice roller (6 years ago, same as above) TL;DR: Humans have a bad intuition on what is or isn't random, and this have been debunked every time it's been discussed here.
1593454759
I guess I should buy a lottery ticket. I played yesterday My first attack roll was a 6, followed by seven 1's in a row then a 2, then a 4 and finally on my last roll of the combat I got a 16. Needless to say I wanted to break my computer.
1593464377
keithcurtis
Forum Champion
Marketplace Creator
API Scripter
We had two sessions in a row recently that started with double crits and ended in double fumbles. Weirdness happens.
1593471952
I just had to I made a program to analyse patterns that I copied from 10.000 d20-rolls (it looks like you have to roll 1000 ten times)&nbsp; I did find 2 patterns 20, 16, 12, 16, 7, 19 and&nbsp; 3, 6, 6, 7, 11, 3 Both occurred twice, so If anyone wants to&nbsp; look for hidden patterns... (there also were a few shorter ones&nbsp; 9, 9, 13, 12, 15&nbsp; 12, 17, 9, 16, 16 3, 6, 4, 10, 16 19, 18, 4, 2, 17&nbsp; 20, 16, 12, 16, 7&nbsp; 16, 12, 16, 7, 19 19, 8, 7, 20, 9&nbsp; 11, 12, 5, 11, 15&nbsp; 8, 19, 6, 8, 20&nbsp; 4, 15, 20, 7, 11&nbsp; 8, 13, 10, 18, 7&nbsp; 18, 7, 8, 13, 17&nbsp; 9, 13, 20, 7, 17&nbsp; 3, 6, 6, 7, 11&nbsp; 6, 6, 7, 11, 3 Also twice, I checked them against the text file....could feed more numbers into the program but it is getting late in my part of the world... )
1593472177
GiGs
Pro
Sheet Author
API Scripter
The thing about randomness, is patterns will &nbsp;occur. Having a couple of 6-roll series occur twice in 10,000 rolls, when you only have a range of 20 possible values, is pretty damn random.&nbsp;
1593472762
Yeah I feel like I would need to feed more numbers into the application to get a better result.
1593504514

Edited 1593504544
Ziechael
Forum Champion
Sheet Author
API Scripter
I like to run my own average result tests whenever these threads occur, based on 999d20s per roll the average is easily within the expected bounds for standard deviation, when analysing such a small set of data, from the supposed average roll for a d20: