As far as I know there is no computer random number generator. Computers just can not do it.
But if the software is good enough it will be virtually indistinguishable from actual dice rolling.
There is a "random" (not truly random but you will never notice) number generator built in to the hardware of a computer but even that requires a bit of software to make it even more random. And then a bit of software to convert it into simulated dice of your choosing.
The dice use a physics engine to move them around a virtual space in a manner that emulates the behaviour of real dice. The initial velocity (speed and direction) is determined by combining unique, local, not repeating values seeded from your system. For example; CPU percentage, number of apps running, device IDs, system clock etc. Some dice sets even have slightly different virtual weights and friction based on the materials their skins are portrayed as having.
All this combines to create virtual dice that function offline with a high degree of pseudo-randomness and a low degree of repeated initial state likely (the chance that the initial variables will be the same and thus produce the same random number). Even the number and type of dice you roll will increase the randomness.
1. You can't predict a result. If you can tell that the die will be a certain value next time or know it's more likely to be a given value, then you have a problem.
2. It doesn't favour specific values to meaningful degree. If it rolls 1s twice as often as other numbers, you have a problem. If the variance is only a couple of percent, then it doesn't matter.
Whether a die is truly random or only pseudorandom is purely academic. You can even a fixed order of numbers, so long as you don't know which number is coming next when making decisions and it's not going to be substantially more likely to be some numbers than others. Iirc, Beyond has an issue on small screens where the dice don't have enough space to roll properly which can bias results, but otherwise it's perfectly fine for D&D.
Rollback Post to RevisionRollBack
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
I still really wish they could add a feature to the app where you could shake your phone to send the dice tumbling.
(Not for actual game application but because it would fascinate me for longer than it should.)
I believe that's a feature on the android version of the app. It's not on the iOS version due to how iOS exposes the gyroscope to apps
No, it's not. At least not that I can see in the settings on the app and I can't seem to get it to work.
Rollback Post to RevisionRollBack
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
Shame. One of those useless but fun features that could swing me to using the app more.
Rollback Post to RevisionRollBack
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
They are psedo random and will never will be truly random.
Yes, which is saying nothing because that's true of every digital RNG. All that matters is the delta between the distribution of outputs of a "true random source" (which doesn't actually exist outside of mathematical constructs) and the pRNG. Which for the purpose of D&D is so small as to be completely irrelevant. It's like saying a ruler isn't exactly 30cm long compared to the international metric of length, but the difference is like 1/1000 of a mm.
As far as I know there is no computer random number generator. Computers just can not do it.
Of course there are. You can run simple random numbers using Excel, as well as simulate dice rolls.
A computer is a true random die roll, because a physical roll can be manipulate (by accident or on purpose) by how you throw the dice
Dice are truly random, if rolled properly. The fact that people can fix the roll, or do it poorly, doesn't change that. (It's mostly about the bounces, I believe.)
As for computers:
Any one who considers arithmetical methods of producing random digits is, of course, in a state of sin. -- John von Neumann
Computer-generated random numbers are not, in fact, truly random* -- if you know the current state of the random number generator, you can reliably predict the next results, because all they do is do a bunch of math on that state, and math is predictable.
However, for our purposes, we don't care. True randomness is a problem for people dealing with serious real-world cryptography and security.
All we care about is "good enough", and modern pseudo-random number generators, properly used, are easily good enough for our purposes. Your average D&D session has a relatively small number of random numbers generated, with the results compressed into a relatively tiny range. Even a pretty poor PRNG will work under those constraints.
* Because of the importance of true randomness for real-world security, there are all sorts of ways people get true randomness into the computer, which is usually then fed into a cryptographically strong PRNG to increase the number of secure bits they have available to work with. But that randomness must come from physical processes -- computers can't do it.
The dice use a physics engine to move them around a virtual space in a manner that emulates the behaviour of real dice. The initial velocity (speed and direction) is determined by combining unique, local, not repeating values seeded from your system. For example; CPU percentage, number of apps running, device IDs, system clock etc. Some dice sets even have slightly different virtual weights and friction based on the materials their skins are portrayed as having.
All this combines to create virtual dice that function offline with a high degree of pseudo-randomness and a low degree of repeated initial state likely (the chance that the initial variables will be the same and thus produce the same random number). Even the number and type of dice you roll will increase the randomness.
That is really cool! Very clever way to make random dice.
Actually their is a background line of code that produces random numbers. These random numbers are them converted by simple math into D&D dice numbers. IE1d4 1d6 1d8.......
What your see happen on the screen means nothing to those numbers. I used to program a little for fun. All computers have a random number generator inside them because its not exactly possible to program a truly random number.
They could also just list the random numbers in a little box.
The number is generated and then the graphics happen.
Random number generators are actually just long lists of numbers, each time you ask the computer for a random number it just takes the next group of numbers and then runs it though a program to make make one number and this sort of looks random. The original lists of numbers were short so short that they could find the repeating numbers and predict the next number. Now the list is very long hundreds of thousands long.
No the computer numbers are not as random as real life physical dice. But are those dice you throw actually random or could something make them roll one number more than others.
The computer numbers are now just as random as those dice you throw.
Probably more random, for the kinds of dice most people will have at home.
Rollback Post to RevisionRollBack
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
Probably more random, for the kinds of dice most people will have at home.
Nah. The dice may be biased, but they're random. A D20 that rolls 7 twice as often as it ought to be (which is an absurd amount of bias) still isn't going to do so predictably.
And the real bias is pretty tiny. It's in the realm of "center of gravity is a bit off", "this corner is rounder", etc. Things that skew the results slightly, but you're not going to be able to tell without doing significant statistical testing. Most of what we see in our dice is the human tendency to see patterns in randomness, because randomness is spikier than intuition expects.
Also, modern dice manufacturing is so much better than it was in the old days.
But just about nobody makes casino-quality dice for RPG play because it's way too expensive. Casinos care because a consistent bias might let somebody who knows what they're doing actually get a long-term statistical edge, and they're aghast at anybody but themselves having that. For D&D, even if you know your d20 has a slight skew it's not going to noticeably help. (And the distribution of numbers on the sides makes it much harder to get a consistent advantage from said skew.)
What your see happen on the screen means nothing to those numbers. I used to program a little for fun. All computers have a random number generator inside them because its not exactly possible to program a truly random number.
The screen stuff on DDB is actually used in the randomizing. They have a physics model of a die, they randomize the initial orientation and vector, and then let it bounce.
It's no more random than just using a decent PRNG, but it's not fake. Generating a random result then realistically spinning and bouncing the model to make the face come up is a much harder computation.
Random number generators are actually just long lists of numbers, each time you ask the computer for a random number it just takes the next group of numbers and then runs it though a program to make make one number and this sort of looks random. The original lists of numbers were short so short that they could find the repeating numbers and predict the next number. Now the list is very long hundreds of thousands long.
It doesn't work like that*. PRNGs have an internal state (one or more numbers, usually a lot more), generated from some initial data (called a seed). Each time you call it, it does a bunch of mathematical transformations on the internal state and extracts a bunch of random bits from it.
* OK, it does work like that. All PRNGs have a period, a point at which they start to repeat themselves. For a simple linear congruential generator, the period can be quite small (for values of "small" that are still quite large on the everyday person's scale). For something like the Mersenne Twister, it's ...not.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
Can someone explain how the virtual dice used by Beyond are as random as real world dice?
They are psedo random and will never will be truly random.
https://www.calculator.net/dice-roller.html#:~:text=Virtual dice, like the one,randomness than most physical dice.
As far as I know there is no computer random number generator. Computers just can not do it.
But if the software is good enough it will be virtually indistinguishable from actual dice rolling.
There is a "random" (not truly random but you will never notice) number generator built in to the hardware of a computer but even that requires a bit of software to make it even more random. And then a bit of software to convert it into simulated dice of your choosing.
The dice use a physics engine to move them around a virtual space in a manner that emulates the behaviour of real dice. The initial velocity (speed and direction) is determined by combining unique, local, not repeating values seeded from your system. For example; CPU percentage, number of apps running, device IDs, system clock etc. Some dice sets even have slightly different virtual weights and friction based on the materials their skins are portrayed as having.
All this combines to create virtual dice that function offline with a high degree of pseudo-randomness and a low degree of repeated initial state likely (the chance that the initial variables will be the same and thus produce the same random number). Even the number and type of dice you roll will increase the randomness.
Find my D&D Beyond articles here
To be honest, only two things matter:
1. You can't predict a result. If you can tell that the die will be a certain value next time or know it's more likely to be a given value, then you have a problem.
2. It doesn't favour specific values to meaningful degree. If it rolls 1s twice as often as other numbers, you have a problem. If the variance is only a couple of percent, then it doesn't matter.
Whether a die is truly random or only pseudorandom is purely academic. You can even a fixed order of numbers, so long as you don't know which number is coming next when making decisions and it's not going to be substantially more likely to be some numbers than others. Iirc, Beyond has an issue on small screens where the dice don't have enough space to roll properly which can bias results, but otherwise it's perfectly fine for D&D.
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
They just are its to basically make it fear both in app and on table
I still really wish they could add a feature to the app where you could shake your phone to send the dice tumbling.
(Not for actual game application but because it would fascinate me for longer than it should.)
I believe that's a feature on the android version of the app. It's not on the iOS version due to how iOS exposes the gyroscope to apps
Find my D&D Beyond articles here
No, it's not. At least not that I can see in the settings on the app and I can't seem to get it to work.
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
You're right, I think the feature was removed or maybe was on the previous app
Find my D&D Beyond articles here
Shame. One of those useless but fun features that could swing me to using the app more.
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
To know it once existed and is now lost is almost worse, hopefully they re-add it someday. :O
Yes, which is saying nothing because that's true of every digital RNG. All that matters is the delta between the distribution of outputs of a "true random source" (which doesn't actually exist outside of mathematical constructs) and the pRNG. Which for the purpose of D&D is so small as to be completely irrelevant. It's like saying a ruler isn't exactly 30cm long compared to the international metric of length, but the difference is like 1/1000 of a mm.
It's not worth splitting hairs over.
Find my D&D Beyond articles here
Of course there are. You can run simple random numbers using Excel, as well as simulate dice rolls.
A computer is a true random die roll, because a physical roll can be manipulate (by accident or on purpose) by how you throw the dice
Dice are truly random, if rolled properly. The fact that people can fix the roll, or do it poorly, doesn't change that. (It's mostly about the bounces, I believe.)
As for computers:
Computer-generated random numbers are not, in fact, truly random* -- if you know the current state of the random number generator, you can reliably predict the next results, because all they do is do a bunch of math on that state, and math is predictable.
However, for our purposes, we don't care. True randomness is a problem for people dealing with serious real-world cryptography and security.
All we care about is "good enough", and modern pseudo-random number generators, properly used, are easily good enough for our purposes. Your average D&D session has a relatively small number of random numbers generated, with the results compressed into a relatively tiny range. Even a pretty poor PRNG will work under those constraints.
* Because of the importance of true randomness for real-world security, there are all sorts of ways people get true randomness into the computer, which is usually then fed into a cryptographically strong PRNG to increase the number of secure bits they have available to work with. But that randomness must come from physical processes -- computers can't do it.
That is really cool! Very clever way to make random dice.
Actually their is a background line of code that produces random numbers. These random numbers are them converted by simple math into D&D dice numbers. IE1d4 1d6 1d8.......
What your see happen on the screen means nothing to those numbers. I used to program a little for fun. All computers have a random number generator inside them because its not exactly possible to program a truly random number.
They could also just list the random numbers in a little box.
The number is generated and then the graphics happen.
Random number generators are actually just long lists of numbers, each time you ask the computer for a random number it just takes the next group of numbers and then runs it though a program to make make one number and this sort of looks random. The original lists of numbers were short so short that they could find the repeating numbers and predict the next number. Now the list is very long hundreds of thousands long.
To answer the real question.
No the computer numbers are not as random as real life physical dice. But are those dice you throw actually random or could something make them roll one number more than others.
The computer numbers are now just as random as those dice you throw.
I trust them.
Probably more random, for the kinds of dice most people will have at home.
If you're not willing or able to to discuss in good faith, then don't be surprised if I don't respond, there are better things in life for me to do than humour you. This signature is that response.
Nah. The dice may be biased, but they're random. A D20 that rolls 7 twice as often as it ought to be (which is an absurd amount of bias) still isn't going to do so predictably.
And the real bias is pretty tiny. It's in the realm of "center of gravity is a bit off", "this corner is rounder", etc. Things that skew the results slightly, but you're not going to be able to tell without doing significant statistical testing. Most of what we see in our dice is the human tendency to see patterns in randomness, because randomness is spikier than intuition expects.
Also, modern dice manufacturing is so much better than it was in the old days.
But just about nobody makes casino-quality dice for RPG play because it's way too expensive. Casinos care because a consistent bias might let somebody who knows what they're doing actually get a long-term statistical edge, and they're aghast at anybody but themselves having that. For D&D, even if you know your d20 has a slight skew it's not going to noticeably help. (And the distribution of numbers on the sides makes it much harder to get a consistent advantage from said skew.)
The screen stuff on DDB is actually used in the randomizing. They have a physics model of a die, they randomize the initial orientation and vector, and then let it bounce.
It's no more random than just using a decent PRNG, but it's not fake. Generating a random result then realistically spinning and bouncing the model to make the face come up is a much harder computation.
It doesn't work like that*. PRNGs have an internal state (one or more numbers, usually a lot more), generated from some initial data (called a seed). Each time you call it, it does a bunch of mathematical transformations on the internal state and extracts a bunch of random bits from it.
* OK, it does work like that. All PRNGs have a period, a point at which they start to repeat themselves. For a simple linear congruential generator, the period can be quite small (for values of "small" that are still quite large on the everyday person's scale). For something like the Mersenne Twister, it's ...not.