New "Intensity Cubes"

[Dako]

Also, if you want to discuss random numbers and they ways to get them I would like to offer you to create a thread in Bumpage and leave this Announcement thread alone. Because this has gone way off the main topic (which is to tell that site has improved).

[sherkaner]

Dako is partially right I think, though I read the argument slightly different. If you're gonna pick randomly from a list, just make sure that list has a good distribution.

And he has a point there, but if we have no knowledge about the list other than that it's random (which means in this case that every number has an equal chance of showing up at every spot), we have to assume it has a good distribution. And even if the distribution is off, since it ishard to misuse, very hard to find out and updated every hour, this is practically impossible to misuse.

And for the record, I have some certificate saying I'm a master at Applied Mathematics (and I vaguely remember 5 years at an university that were mainly about mathematics, though luckily most of it wasn't about statistics).

[carlpgoodrich]

File1. Get a list of numbers 1 to 6, each number repeated 10,000 times. The list will have 60,000 numbers in it. Randomize its order. This has uniform distribution, and is our pool of numbers. IE our DICE.
File2. Get a list of numbers 1 to 60,000, 600,000 long. Each number would be in there 10 times if it had uniform distribution, but it won't, because it's a random list. This list is how you choose which dice you get.

When a dice number is needed, read the next line from file2. It has a number, go read that spot from file1. That's your dice.

Do you know what your method produces in the end? He produces a list of random numbers. Which is exactly a mechanism CC is using right now. But we just don't need to generate that list of random numbers because they are generated for us by random.org site.

Exactly! If you create a third file that is the result of using file2 to get a number from file1 you will get precisely what we are getting from random.org. You will notice that every number (1-6) does not appear the same number of times, but it is still a sample of a uniform distribution. I also like Lack's explanation of thinking of this as a buffer.

Oh, and I don't appreciate the attitude of some people here. I'm trying to be polite and explain myself. I took statistics in college, and calculus for engineers etc. No need to get rude when discussing this. I actually enjoy it. Random numbers are fascinating!

I assume this was directed at me. I never meant to be rude, though I did get frustrated at times when it felt like you were ignoring everything I was saying. I too enjoy discussions like this, and I was honestly trying my best to explain what I believe to be true. If that came off as obnoxious at times, then I apologize.

[natty dread]

How about we just use this algorithm. Nobody can complain then.

[n00blet]

How about we just use this algorithm. Nobody can complain then.

This game has been won.

[bedub1]

FYI - I emailed the head of random.org for guidance. Here is his response and my original question.
EDIT: Removed Mads personal information.


Hi Bedub1,

If the list you are reading from originates from RANDOM.ORG, then the numbers in it are random and follow a uniform distribution. If the number that you generate locally is a pseudo-random number, then I would describe your system as a pseudo-random number generator (PRNG). If the number you generate locally is a true random number and you pick a new spot in the list for every new number you take from the list, then I would classify your system as a true random number generator (TRNG). However, in this case, you wouldn't really need the RANDOM.ORG-generated list in the first instance, since you already have a TRNG.

Did that answer your question?

Kind regards,
Mads
--
Dr. Mads Haahr <> | Lecturer in Computer Science
Department of Computer Science | Phone: +353 1
Trinity College Dublin | Web: http://www.cs.tcd.ie/


On Mon, 12 Jul 2010, Bedub1 wrote:

Hello,

I have a question about random numbers I can't seem to get my head around.
On random.org you state:

"When discussing single numbers, a random number is one that is drawn from a
set of possible values, each of which is equally probable, i.e., a uniform
distribution<http://en.wikipedia.org/wiki/Uniform_distribution_%28discrete%29>.

"

So if I have random.org provide me with a list of random numbers, and I use
a random number generator to choose a spot to read a number from in that
list, have I obtained a random number? Or is my number no longer random as
I am drawing from a list of numbers without equal probability and uniform
distribution as the list I am reading from is random, not uniform?

[Ace Rimmer]

(carryover from live chat)

You have proven nothing by emailing him. I'm sorry you can't understand what he is saying. He is NOT saying that we have to read the list randomly. Look at this:

"If the list you are reading from originates from RANDOM.ORG, then the numbers in it are random and follow a uniform distribution."

This means that even though we are getting numbers and storing them, these are still random numbers. End of conversation. Now by using the local RNG (which is a PRNG0 to determine where in the file we read from, we are introducing a PRNG into the equation. The number is still random, but I believe Dr Haahr would state that it would be best just to read the list of numbers we get from random.org sequentially, as these are still obtained from a TRNG.

I'm not sure how this can be explained more clearly.

[bedub1]

(carryover from live chat)

You have proven nothing by emailing him. I'm sorry you can't understand what he is saying. Look at this:

"If the list you are reading from originates from RANDOM.ORG, then the numbers in it are random and follow a uniform distribution."

This means that even though we are getting numbers and storing them, these are still random numbers.

Of course they are TRN's. I've agreed with this since the beginning.

Now by using the local RNG (which is a PRNG0 to determine where in the file we read from, we are introducing a PRNG into the equation. The number is still random, but I believe Dr Haahr would state that it would be best just to read the list of numbers we get from random.org sequentially, as these are still obtained from a TRNG.

I'm not sure how this can be explained more clearly.

Introducing a PRNG into the equation results in us obtaining a PRN instead of a TRN.

He is NOT saying that we have to read the list randomly.

it appears to me he is saying just that if we intend to achieve a TRN instead of a PRN:

you pick a new spot in the list for every new number you take from the list, then I would classify your system as a true random number generator (TRNG)

[carlpgoodrich]

If you have a list of TRN and your system reads consecutively from the list, then it is a TRNG. How would using a PRNG to read from the list of TRN make it not a TRNG? If you rearrange a list of TRN (e.g. using a PRNG to read from the list), you still have a list of TRN. The only reason Lack does this is to prevent a possible form of cheating that was discussed earlier in this thread, not to make the numbers more or less random.

[bedub1]

If you have a list of TRN and your system reads consecutively from the list, then it is a TRNG. How would using a PRNG to read from the list of TRN make it not a TRNG? If you rearrange a list of TRN (e.g. using a PRNG to read from the list), you still have a list of TRN. The only reason Lack does this is to prevent a possible form of cheating that was discussed earlier in this thread, not to make the numbers more or less random.

I don't know, but that seems to be what the guy that made random.org is saying. Using a PRNG to read from a list of TRN's results in a PRN, not in a TRN. But reading from a list of TRN's sequentially results in a TRN.

How about this. The attacker rolls 3 dice, the defender 2. Each dice is it's own TRNG. So have 3 lists of TRN's for the attacker, 1 for each dice, and 2 lists of TRN's for the defender, 1 for each dice, and read from them sequentially? A single list of TRN's would be a list of 1 dice being rolled over and over, not 5 dice. Is this right?

The lists could still be small to fit in ram and be fast, and you'd never know where from the list you are going to read, as some attacks are 2v1 and won't read from all lists...

[AndyDufresne]

TRNG, PRNG, Star Trek: TNG.

If mathematic, statistics, and logic were my forte, I'd comment more, but unfortunately they are not. Perhaps some this discussion would be better suited in a Off Topic forum topic?


--Andy

[Metsfanmax]

If you have a list of TRN and your system reads consecutively from the list, then it is a TRNG. How would using a PRNG to read from the list of TRN make it not a TRNG? If you rearrange a list of TRN (e.g. using a PRNG to read from the list), you still have a list of TRN. The only reason Lack does this is to prevent a possible form of cheating that was discussed earlier in this thread, not to make the numbers more or less random.

I don't know, but that seems to be what the guy that made random.org is saying. Using a PRNG to read from a list of TRN's results in a PRN, not in a TRN. But reading from a list of TRN's sequentially results in a TRN.

That is incorrect. You can use a PRNG to select a number of a list of TRN's, and the result is still truly random. Insofar as we accept that the list itself is composed of truly random numbers, and that no one actually has access to that list except for the dice selection mechanism itself, then it doesn't matter that the "selection" mechanism is not truly random. To convince yourself of this, consider the following thought experiment: say you used a PRNG to select a pseudo-random number corresponding to some time of the day. Using the results of that PRNG, you would have your program go to random.org and pick a random number at that time. Would you agree that the results of this experiment give us a random number? If you think about it, it must - because sure, one could hack the PRNG and figure out what time of the day you connected to random.org and got the random number, but the number you got from the site itself came from a TRNG, so it's irrelevant that you figured out what time of the day the site was called upon.

The current method is the same as the above method, except instead of connecting to random.org each time, the selection mechanism connects to a list of TRN's.

[sm8900]

gahhh. still trying to understand. ok. will reread this.

[danfrank]

Site is still rigged and unless proven otherwise that is what i will think. Take for instance tonite.. All my first rolls on everyturn has been double defensed . So i lose 2 armies to start every turn.. Is it bad luck ? Nope . Happens way too often for it to be luck.

[carlpgoodrich]

It has been proven otherwise, many many many times.

[bedub1]

Site is still rigged and unless proven otherwise that is what i will think. Take for instance tonite.. All my first rolls on everyturn has been double defensed . So i lose 2 armies to start every turn.. Is it bad luck ? Nope . Happens way too often for it to be luck.

lol...i was actually just saying the exact same thing. Last night it was about 90% i loose the first roll, defender wins. This morning it started the same, but quickly evened out and I won the first couple attacks. I stand by the statement of the author of random.org that our system is a Pseudo random number generator and is actually like rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time. But hey, we got to dump a server and reduced the load that much so who really cares?

[Metsfanmax]

Anyway, the author of that e-mail never said that this system was pseudo-random, and if you think about it for even a second, there's no way to get a pseudo-random number from any process that includes a random number generator at some step. You might be able to reproduce every other step in the process by learning the algorithms, but you could never reproduce the part that includes a TRNG. Therefore the whole process is random.

[carlpgoodrich]

... rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time.

What on earth is the difference? Forget the questions of PRNG/TRNG etc. for a second... Please, I want to hear this.

[jefjef]

Whats all the discussion about? The drops and dice are manipulated.

Now if CC would kindly move me to the positive manipulations then I will adamantly defend the randomness of CC.

[bedub1]

... rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time.

What on earth is the difference? Forget the questions of PRNG/TRNG etc. for a second... Please, I want to hear this.

The first is a single random number generator asked for a number 5 times, and then read sequentially. (you roll 1 dice 3 times, your buddy rolls 1 dice 2 times)
The second is five random number generators all asked for a single number once for the action desired, and read in parallel. (you roll 3 dice and your buddy rolls 2 at the same time)
Which is the way it works in real life? It is my belief we should attempt to mimic real life.

random.org states the way to read the lists of numbers obtained from them is to read them sequentially. We used to get a huge list and read it sequentially for the 5 dice, which was proper. It got changed, one of the reasons being in hindsight that it was a huge load on the servers. The new option is more efficient, but we use a PRNG that results in a PRN. Smaller lists seem better, but I don't care for the PRNG.

Proposal
Instead how about we get 5 lists and read them sequentially like we are supposed to, just one list for each dice.

Benefits
Now the streaks don't run through the 5 dice you see when you attack somebody, they run down your first attack dice, down your second attack dice etc.
Small lists for each dice should be a lower load
multiple lists makes cheating appear impossible
We get rid of the PRNG, and go back to using a TRNG in the correct fashion

I am hesitant to use a pseudo-random number generator because the numbers we use from random.org are truly random and that seems like a step backwards.

EDIT:
Version 1

The intensity levels are based on high quality random numbers from Random.org. The numbers are read from a large file containing columns of numbers from 1 to 6, in the format A1 A2 A3 D1 D2. When the intensity levels are generated, the game engine reads a line from the file and discards it. The appropriate numbers are used and the others are ignored. The file contains 500,000 lines of intensity levels and is re-loaded when all the lines are used up.

Version 2

This is how the intensity cubes now work:

* We have a series of 50,000 high quality random numbers from random.org
* Each time the game engine generates a random intensity cube, the next number is read in sequence from the series (e.g. in a 3v1 attack 4 numbers are read sequentially)
* When the last number in the series is read, we "rewind" and continue with the first number in the series

Version 2.1

I will replace our series of 50k random numbers with new numbers from random.org as often as their quota allows!

Version 3

This is how the intensity cubes now work:

* We have a series of 50,000 true random numbers from random.org
* Each time the game engine processes an assault or auto-assault, it select a random spot in the series to read from using a pseudo-random computer function
* Each time the game engine generates a random intensity cube, the next number is read in sequence from the series (e.g. in a 3v1 attack 4 numbers are read sequentially)
* The series of 50,000 true random numbers from random.org is replaced every hour

[Night Strike]

... rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time.

What on earth is the difference? Forget the questions of PRNG/TRNG etc. for a second... Please, I want to hear this.

The first is a single random number generator asked for a number 5 times, and then read sequentially. (you roll 1 dice 3 times, your buddy rolls 1 dice 2 times)
The second is five random number generators all asked for a single number once for the action desired, and read in parallel. (you roll 3 dice and your buddy rolls 2 at the same time)
Which is the way it works in real life? It is my belief we should attempt to mimic real life.

When you're referring to gathering data, they're the exact same thing as each number as the exact same probability of occurring, whether you throw them 1 at a time or all at once. Absolutely no difference.

random.org states the way to read the lists of numbers obtained from them is to read them sequentially. We used to get a huge list and read it sequentially for the 5 dice, which was proper. It got changed, one of the reasons being in hindsight that it was a huge load on the servers. The new option is more efficient, but we use a PRNG that results in a PRN. Smaller lists seem better, but I don't care for the PRNG.

How does the whole file result in a PRN when the only thing the PRNG does is tell you where to start reading? The list is still compiled from a TRNG and read sequentially as random.org states, so the data in now way magically becomes a PRNG.

[carlpgoodrich]

... rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time.

What on earth is the difference? Forget the questions of PRNG/TRNG etc. for a second... Please, I want to hear this.

The first is a single random number generator asked for a number 5 times, and then read sequentially. (you roll 1 dice 3 times, your buddy rolls 1 dice 2 times)
The second is five random number generators all asked for a single number once for the action desired, and read in parallel. (you roll 3 dice and your buddy rolls 2 at the same time)
Which is the way it works in real life? It is my belief we should attempt to mimic real life.

random.org states the way to read the lists of numbers obtained from them is to read them sequentially. We used to get a huge list and read it sequentially for the 5 dice, which was proper. It got changed, one of the reasons being in hindsight that it was a huge load on the servers. The new option is more efficient, but we use a PRNG that results in a PRN. Smaller lists seem better, but I don't care for the PRNG.

Proposal
Instead how about we get 5 lists and read them sequentially like we are supposed to, just one list for each dice.

Benefits
Now the streaks don't run through the 5 dice you see when you attack somebody, they run down your first attack dice, down your second attack dice etc.
Small lists for each dice should be a lower load
multiple lists makes cheating appear impossible
We get rid of the PRNG, and go back to using a TRNG in the correct fashion

I am hesitant to use a pseudo-random number generator because the numbers we use from random.org are truly random and that seems like a step backwards.

EDIT:
Version 1

The intensity levels are based on high quality random numbers from Random.org. The numbers are read from a large file containing columns of numbers from 1 to 6, in the format A1 A2 A3 D1 D2. When the intensity levels are generated, the game engine reads a line from the file and discards it. The appropriate numbers are used and the others are ignored. The file contains 500,000 lines of intensity levels and is re-loaded when all the lines are used up.

Version 2

This is how the intensity cubes now work:

* We have a series of 50,000 high quality random numbers from random.org
* Each time the game engine generates a random intensity cube, the next number is read in sequence from the series (e.g. in a 3v1 attack 4 numbers are read sequentially)
* When the last number in the series is read, we "rewind" and continue with the first number in the series

Version 2.1

I will replace our series of 50k random numbers with new numbers from random.org as often as their quota allows!

Version 3

This is how the intensity cubes now work:

* We have a series of 50,000 true random numbers from random.org
* Each time the game engine processes an assault or auto-assault, it select a random spot in the series to read from using a pseudo-random computer function
* Each time the game engine generates a random intensity cube, the next number is read in sequence from the series (e.g. in a 3v1 attack 4 numbers are read sequentially)
* The series of 50,000 true random numbers from random.org is replaced every hour

No, you didn't answer my question. I repeat: What is the difference between rolling an actual dice 5 times and rolling 5 dice one time each? One of the problems with this discussion is that you are not listening to what other people are saying, so it makes it very hard for you to understand.

[bedub1]

... rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time.

What on earth is the difference? Forget the questions of PRNG/TRNG etc. for a second... Please, I want to hear this.

The first is a single random number generator asked for a number 5 times, and then read sequentially. (you roll 1 dice 3 times, your buddy rolls 1 dice 2 times)
The second is five random number generators all asked for a single number once for the action desired, and read in parallel. (you roll 3 dice and your buddy rolls 2 at the same time)
Which is the way it works in real life? It is my belief we should attempt to mimic real life.

No, you didn't answer my question. I repeat: What is the difference between rolling an actual dice 5 times and rolling 5 dice one time each? One of the problems with this discussion is that you are not listening to what other people are saying, so it makes it very hard for you to understand.

Did I not explain it? I thought it was the first thing I did...

How does the whole file result in a PRN when the only thing the PRNG does is tell you where to start reading? The list is still compiled from a TRNG and read sequentially as random.org states, so the data in now way magically becomes a PRNG.

The guy that made random.org states it does:

If the number that you generate locally is a pseudo-random number, then I would describe your system as a pseudo-random number generator (PRNG).

Think of it this way. If the list had 6 numbers in it instead of having 50k numbers in it, and we use a PRNG to pick a spot to start reading, you are basically picking a random number. A PRN. I have to agree with lack that I'd prefer not to have a PRNG involved....

[carlpgoodrich]

... rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time.

What on earth is the difference? Forget the questions of PRNG/TRNG etc. for a second... Please, I want to hear this.

The first is a single random number generator asked for a number 5 times, and then read sequentially. (you roll 1 dice 3 times, your buddy rolls 1 dice 2 times)
The second is five random number generators all asked for a single number once for the action desired, and read in parallel. (you roll 3 dice and your buddy rolls 2 at the same time)
Which is the way it works in real life? It is my belief we should attempt to mimic real life.

No, you didn't answer my question. I repeat: What is the difference between rolling an actual dice 5 times and rolling 5 dice one time each? One of the problems with this discussion is that you are not listening to what other people are saying, so it makes it very hard for you to understand.

Did I not explain it? I thought it was the first thing I did...

No, you talked about generating random numbers, etc. I am asking what the difference is between two people playing physical games with actual dice that they put in their hands and throw on the ground. If you and your friend want to play risk but can only find one die, do you give up or do you roll that one die 5 times? You seem to think that rolling that one die 5 times would be different than rolling 5 dice one time each. I am asking you to explain this.

Think of it this way. If the list had 6 numbers in it instead of having 50k numbers in it, and we use a PRNG to pick a spot to start reading, you are basically picking a random number. A PRN. I have to agree with lack that I'd prefer not to have a PRNG involved....

I agree that using a PRNG pick a number between 1 and 6 gives PRNs, but that is not what we are doing, not at all. If you use a TRNG to generate a list of TRNs, and you want to read from this list, starting at the beginning is just as arbitrary as starting anywhere else. Hence using a PRNG to pick where we start does nothing to remove the "true" randomness of the list of TRNs.

[n00blet]

... rolling 1 dice 5 times to get the numbers, instead of 5 dice 1 time.

What on earth is the difference? Forget the questions of PRNG/TRNG etc. for a second... Please, I want to hear this.

The first is a single random number generator asked for a number 5 times, and then read sequentially. (you roll 1 dice 3 times, your buddy rolls 1 dice 2 times)
The second is five random number generators all asked for a single number once for the action desired, and read in parallel. (you roll 3 dice and your buddy rolls 2 at the same time)
Which is the way it works in real life? It is my belief we should attempt to mimic real life.

No, you didn't answer my question. I repeat: What is the difference between rolling an actual dice 5 times and rolling 5 dice one time each? One of the problems with this discussion is that you are not listening to what other people are saying, so it makes it very hard for you to understand.

Did I not explain it? I thought it was the first thing I did...

So you're saying that if we had 5 dice, and I rolled them all individually and wrote down those five numbers, and then rolled only 1 of them 5 times and wrote down those 5 numbers, that the second list is suddenly less random than the first list?