Are dice rolls rigged?

[Millenia]

I seem to roll very low on a 20d VERY often compared to higher numbers.

[Erei]

The random generator number is kinda bad. Which, sadly is common in videogames. The latest pathfinder also have the same issue.
Poor random gen number often give the same numbers several time in a row, or a very close number.

Also, do note it's harder to notice when it's some random number (like a 12 for example) which is repeated over a 1 or a 20. Which mean, the problem is even bigger if you already notice repetition in the 1 and 20.

It is not biased toward or against the player though. You'll notice regularly the fight go "one way" or the other. With one side getting low roll several time in a row, or high roll. You just notice the low roll because it's easier to note the bad, and because you can miss even on a 2 or 3, while a 18 or 19 won't likely make a difference (even though they are both high).

I had situation many times where the odd to have something like that statiscally was incredibly low. For example, 3 20 in a row. And they were not rare occurence (as you'd expect).

I also noticed usually it's 3-4 numbers that get "borked" then the RNG fix itself. At least for a time.

[Grumpy Old Dude]

If you check in Settings and opt NOT to preserve the RNG seed; you will not get the sequence of die rolls over and over.

[dataseer21]

It is probable. However I think what you might be noticing is if the random number generator causes a .1 shift in the mean. Not saying that is the number but a simple .1 shift on a d20 would be something that would be felt. I honestly believe though that there is bias. However I believe that it is something small enough that if you know how to optimize your characters you can overcome it quite easily. I remember being on roll20 and a guy in the game for a campaign that was a year long went into the logs and calculated the average of each players rolls over a whole campaign. There was bias. I believe that there is bias in games like this but it can be overcome. If your character is optimized you can see the bias and laugh because even though you keep rolling 4 4 times in a row you can still hit the creature because of all your bonuses. I took the same approach with kingmaker and make characters that hit monsters on a roll of a 2 on the dice so even getting above average number of 1s doesn't seem that discouraging.

[Grumpy Old Dude]

"I believe..."..."I feel like..."...."I think..."...

sorry, meaningless drivel without relevant data sets to support.

[Cartesian Duelist]

Originally posted by Grumpy Old Dude:

"I believe..."..."I feel like..."...."I think..."...

sorry, meaningless drivel without relevant data sets to support.

You'd think you'd understand by now that nobody cares what you think. And quit asking people for data that you haven't taken the time to collect either.

[Aimee]

Originally posted by Grumpy Old Dude:

Are you saying that from memory, or do you have recorded data points to support your conclusion?

I have a pretty good memory for better or worse.
And even if i have data to provide, people like you wouldend even appreciate it anyways.
So why would i put that effort in for someone like you?

[Grumpy Old Dude]

Maybe, contrary to what the courtesan above alleges, because over a year ago I DID collect the data and disproved the allegation of bias die rolls. No, I did not retain the data sheets to refute this same claim over and over again. Why would I? This question has been settled, long ago.

Here, is a recorded approx 40 hr play through of the Crown campaign done shortly after the game came out of Early Access. Record the die rolls, they show; then evaluate them. Sorry, but you'll find that there is no innate bias to them.

https://youtube.com/playlist?list=PLm_iFrgsdkTlRSfkUI0BhLxIX8sVKOMGr

[Tinball]

Originally posted by Cartesian Duelist:

Originally posted by Grumpy Old Dude:

"I believe..."..."I feel like..."...."I think..."...

sorry, meaningless drivel without relevant data sets to support.

You'd think you'd understand by now that nobody cares what you think. And quit asking people for data that you haven't taken the time to collect either.

No - he is correct. Anecdotal evidence is weak when complaining about RNG. Show us the stats!

[Erei]

As I said, I didn't notice a bias toward one side, but the roll ARE wonky. Yesterday I had 3 1 in a row, which would be very rare, but I had 3 20 the day before. I don't remember how to caculate that, but I used an online tool and apparently the odds to have the same dice roll from a 20d 3 time in a row is 0.0125% chance. While statiscally it can happen, it's not a regular occurence at all, and the odd of it happening more than once in such a small period of time is incredibly low.

Those repeating rolls both affected the enemy side, so clearly there is no bias (one time they made 3 crit miss, the other 3 crit hit). But clearly, the RNG is wonky. And those were memorable dice rolls due to their nature, I did notice some repetition otherwise but it's more difficult to see when a 12 or something repeats because it's less "important".

[Orion Invictus]

Originally posted by Erei:

As I said, I didn't notice a bias toward one side, but the roll ARE wonky. Yesterday I had 3 1 in a row, which would be very rare, but I had 3 20 the day before. I don't remember how to caculate that, but I used an online tool and apparently the odds to have the same dice roll from a 20d 3 time in a row is 0.0125% chance. While statiscally it can happen, it's not a regular occurence at all, and the odd of it happening more than once in such a small period of time is incredibly low.

Those repeating rolls both affected the enemy side, so clearly there is no bias (one time they made 3 crit miss, the other 3 crit hit). But clearly, the RNG is wonky. And those were memorable dice rolls due to their nature, I did notice some repetition otherwise but it's more difficult to see when a 12 or something repeats because it's less "important".

Time for a short explanation of this "X identical dice rolls in a row" thing.

The odds of getting 3 identical dice rolls in a row in exactly 3 attempts are small, yes. However, the odds of getting 3 identical dice rolls in a row in hundreds - possibly thousands - of attempts are actually quite high.

The problem is you're conflating the first with the second. You're looking at JUST those three results and completely ignoring the context in which those three results presented themselves. How many rolls do you usually get in an entire day? 500 rolls* are enough to "increase the probability"** over 450-fold.

Furthermore, people usually think that getting a specific dice roll means that the odds of getting it again is somehow reduced; that if you get a 1 on one roll, the odds of getting another one right away is smaller. This is not the case either. You have the same probability of rolling any given number in any given roll regardless of whatever number you rolled before.

*Assuming the calculator I found is accurate, because I didn't feel like doing the math myself, the odds of getting 3 identical rolls in a row over 500 attempts are around 5.745%.

**The probability never actually increases; instead, the whole problem has shifted from "odds of getting 3 identical dice rolls in a row in 3 attempts" to "odds of getting 3 identical dice rolls in a row in 500 attempts".

[Marlowe-33]

i know one thing my dwarven paladin seems to miss an awful lot , while he is getting beaten from all sides

[Erei]

Originally posted by Orion Invictus:

Time for a short explanation of this "X identical dice rolls in a row" thing.

The odds of getting 3 identical dice rolls in a row in exactly 3 attempts are small, yes. However, the odds of getting 3 identical dice rolls in a row in hundreds - possibly thousands - of attempts are actually quite high.

The problem is you're conflating the first with the second. You're looking at JUST those three results and completely ignoring the context in which those three results presented themselves. How many rolls do you usually get in an entire day? 500 rolls* are enough to "increase the probability"** over 450-fold.

Furthermore, people usually think that getting a specific dice roll means that the odds of getting it again is somehow reduced; that if you get a 1 on one roll, the odds of getting another one right away is smaller. This is not the case either. You have the same probability of rolling any given number in any given roll regardless of whatever number you rolled before.

*Assuming the calculator I found is accurate, because I didn't feel like doing the math myself, the odds of getting 3 identical rolls in a row over 500 attempts are around 5.745%.

**The probability never actually increases; instead, the whole problem has shifted from "odds of getting 3 identical dice rolls in a row in 3 attempts" to "odds of getting 3 identical dice rolls in a row in 500 attempts".

I just did another 3 same roll in a row. And let me tell you, I didn't roll thousands (or even hundreds) of die since the last one. And those I only the one I'm certain of (some I think I did but I'm not 100% sure). As I said, I may have rolled countless of 10, 11, 12 in a row and never noticed.
It is a very common thing in this game (and others, aswell as everywhere you get computer generated RNG numbers), and I know for a fact random generator that are not of high quality often produce the same number (or the same bracket) several time in a row.

I'm not sure about your calculator. If something have a 1% chance to happen, let's say, then 500 times of that chance should give 5 chance to happen (statistically, not guaranteed ofc). Good old 500*(1/100). In this case we have 0.0125% chance, that mean 500times that is equal to 0.0625 chance to happen. Which, while higher, is still incredibly low.
According to another calculator (which confirmed the first one), the odd to have that happen once is 1/8000, which mean, stastically, to reach a 1 chance to have such an event occuring only happen at 8000rolls and above. I'm not even sure I roll that many dices in an entire campaign :D

I'd want to add it's not a big complain. RNG generator on computer tend to do that because quality RNG are expansive. The algorithm behind them can be quite complex, and those are literally sold by company which make them. A game company do'nt have the resources and means to make it themselves, so they are reliant on pre-existing ones.

[Orion Invictus]

Originally posted by Erei:

I just did another 3 same roll in a row. And let me tell you, I didn't roll thousands (or even hundreds) of die since the last one. And those I only the one I'm certain of (some I think I did but I'm not 100% sure). As I said, I may have rolled countless of 10, 11, 12 in a row and never noticed.

But you HAVE rolled (or will roll) thousands of dice throughout your game. That's the point. You can't just look at a tiny subset of the sequence and say "This is unlikely". You have to look at the whole thing.

Originally posted by Erei:

It is a very common thing in this game (and others, aswell as everywhere you get computer generated RNG numbers), and I know for a fact random generator that are not of high quality often produce the same number (or the same bracket) several time in a row.

Bad generators don't necessarily produce the same number several times in a row. That's not something you look at when analyzing an RNG, because having the same number several times in a row IS a possible result and SHOULD occur.

To see if a generator is good you look at things like mean, variance, possibly a frequency diagram of all the values to make sure one isn't popping up more often than the others, and you use the generator MANY times before analyzing the data. I'm talking millions, possibly more times, depending on other factors.

Originally posted by Erei:

I'm not sure about your calculator. If something have a 1% chance to happen, let's say, then 500 times of that chance should give 5 chance to happen (statistically, not guaranteed ofc). Good old 500*(1/100). In this case we have 0.0125% chance, that mean 500times that is equal to 0.0625 chance to happen. Which, while higher, is still incredibly low.

That's not how statistics work. What you've just calculated is the expected number of times of getting the same number 3 times in 3 dice rolls if you did 3 dice rolls 500 times. And no, that's not the same as the probability of getting the same number 3 times in 3 dice rolls if you rolled the dice 1500 times.

Even IF that were the actual probability of getting 3 identical dice rolls in a row after 500 rolls (or 1500 rolls), a non-zero probability indicates that it CAN happen.

Originally posted by Erei:

According to another calculator (which confirmed the first one), the odd to have that happen once is 1/8000, which mean, stastically, to reach a 1 chance to have such an event occuring only happen at 8000rolls and above. I'm not even sure I roll that many dices in an entire campaign :D

I have no idea what kind of math you think you're doing, but I can guarantee you it's wrong.

I used a "losing streak" calculator used for bets. I put the odds at 5% (1/20), said I wanted to know the odds of getting a 3-losing streak over 500 rolls.

Originally posted by Erei:

I'd want to add it's not a big complain. RNG generator on computer tend to do that because quality RNG are expansive. The algorithm behind them can be quite complex, and those are literally sold by company which make them. A game company do'nt have the resources and means to make it themselves, so they are reliant on pre-existing ones.

That greatly depends on the kind of generator you want. You can make a "good enough" RNG in 5 minutes, for free.

EDIT: If you want an article that explains how to do this, see below. Keep in mind, it can get complicated. However, it also has Python code you can use, if you want to. It's technically for coin flips, but you can still use it if you set the odds at 0.95/0.05.

https://www.askamathematician.com/2010/07/q-whats-the-chance-of-getting-a-run-of-k-successes-in-n-bernoulli-trials-why-use-approximations-when-the-exact-answer-is-known/

[Aimee]

Originally posted by Grumpy Old Dude:

Here, is a recorded approx 40 hr play

Plugging your own youtube channel to get more views and exposure, is an entirely different thing than providing data on dice rolls.
If you ask something from other people, it is only normal others expect the same from you.
If you are so fed up about people discussing an experience in a game of all things, either you would have kept the data since you want to dispute anyones experience that does not fit your own. Or you would have said you dont have it anymore, because its not good evidence.
I honnestly expect the later, since most people dont even know how to do proper research and data gathering.
Regardless, you are not adding anything to this conversation other than insulting and flaming other people. Who simply share their experience and thoughts.

Originally posted by Tinball:

No - he is correct. Anecdotal evidence is weak when complaining about RNG. Show us the stats!

Anyone is free to share their experience and talk about it on the forums, thats what these forums are about. Not everything needs evidence or be proven right or wrong.

Rng is also never the same for everyone, some people just have more luck than others and rolls are never spread out evenly.

[Grumpy Old Dude]

I'm not monetized, and I didnt link it for views. I linked it, because it is a data set subject to review, by any number of independent persons. ie, the data can be independently collected and analyzed by multiple people, thus replicating/confirming the results/math.

Dont conflate hard data points, with ego. (Also, feel free to link anyone elses recorded data for use instead if you prefer. I really dont GAF)