do the dice ever run out?

[moo_lol]

well??

[owenshooter]

sometimes they roll off the edge of my laptop and i have to have to have lack make an official call, as to whether i get a re-roll or just get what they landed on... oh, and my kid swallowed one, once... but he pooped out a 6, and i took that instead of a re-roll... true stories.-0

[Killer619]

sometimes they roll off the edge of my laptop and i have to have to have lack make an official call, as to whether i get a re-roll or just get what they landed on... oh, and my kid swallowed one, once... but he pooped out a 6, and i took that instead of a re-roll... true stories.-0

[Twill]

Owen, that was one of the few funny posts you've ever made - I actually smiled this time

and moo-lol, we run out about ever 17 hours or so.

Our dice file holds 500,000 unique rolls, right now we are averaging over 800,000 rolls a day I think.

when they run out, they cycle over again.

They are still random

[owenshooter]

Owen, that was one of the few funny posts you've ever made - I actually smiled this time

i find that hard to believe...-0

[-ShadySoul-]

sometimes they roll off the edge of my laptop and i have to have to have lack make an official call, as to whether i get a re-roll or just get what they landed on... oh, and my kid swallowed one, once... but he pooped out a 6, and i took that instead of a re-roll... true stories.-0

that was good

[Killer619]

sometimes they roll off the edge of my laptop and i have to have to have lack make an official call, as to whether i get a re-roll or just get what they landed on... oh, and my kid swallowed one, once... but he pooped out a 6, and i took that instead of a re-roll... true stories.-0

that was good

it was so funny......

[DiM]

They are still random

liar. we all know lack controls them. just a few moments ago he killed 16 of my troops and lost just 5

and the previous turn i attacked him and lost 13 while he lost just 6

[Bean_]

Owen, that was one of the few funny posts you've ever made - I actually smiled this time

and moo-lol, we run out about ever 17 hours or so.

Our dice file holds 500,000 unique rolls, right now we are averaging over 800,000 rolls a day I think.

when they run out, they cycle over again.

They are still random

This is bad.

If they cycle over, this is a definitive admission that the dice are obviously not random. This is actually somewhat disconcerting. There is a systemic bias in the dice (i.e., the outcomes will not match the exact probabilities that they should, and there are never "new" dice that are, a priori, random). Also, in theory someone who auto attacks 10,000 on 10,000 might be able to get a significant portion of the dice sequence and use it for later. (Yes -- it isn't easy to game the system that way, but that's how people win at blackjack, in essence.)

[Twill]

This is bad.

If they cycle over, this is a definitive admission that the dice are obviously not random. This is actually somewhat disconcerting. There is a systemic bias in the dice (i.e., the outcomes will not match the exact probabilities that they should, and there are never "new" dice that are, a priori, random). Also, in theory someone who auto attacks 10,000 on 10,000 might be able to get a significant portion of the dice sequence and use it for later. (Yes -- it isn't easy to game the system that way, but that's how people win at blackjack, in essence.)

Without saying too much about how the system works...it doesn't work that way

The dice file is random, thus meaning that if we cycle over it is still random...it's just the same random random numbers over again.

It is also impossible to predict/reserve which number you are getting, even if you had access to the file and a millisecond timer and trigger finger (there are what, 10 dice rolls ever second?).

The system doesn't just "cycle through" the numbers. there is a bit more randomness added in and the randomness which is added in is in fact based on randomness pulled from a random source. (that makes it several layers of random deep I think)

Aaaaaaaand, the dice file is replaced from time to time just to keep things interesting and add more randomness to the system.

Hope that puts your mind at ease there a bit

Have a good one
Twill

[demon7896]

i get what beand is saying. he/she thinks that we roll 5,000 dice, and when we run out, just use the same 5,000 numbers.

[Bean_]

Without saying too much about how the system works...it doesn't work that way

The dice file is random, thus meaning that if we cycle over it is still random...it's just the same random random numbers over again.

It is also impossible to predict/reserve which number you are getting, even if you had access to the file and a millisecond timer and trigger finger (there are what, 10 dice rolls ever second?).

The system doesn't just "cycle through" the numbers. there is a bit more randomness added in and the randomness which is added in is in fact based on randomness pulled from a random source. (that makes it several layers of random deep I think)

Aaaaaaaand, the dice file is replaced from time to time just to keep things interesting and add more randomness to the system.

Hope that puts your mind at ease there a bit

Have a good one
Twill

That's a little better, though still not random. Any time CC flushes the dice file and replaces it with a new set, *those* are random the first time they are used.

If CC reuses a set already used, it's not. For a simple example, let's say we take 12 dice rolls from random.org. It happens that the dice roll out as 1,1,1,2,2,2,3,5,5,5,6,6 (not necessarily in that order). Those dice were random when generated. But if that set is always reused, then the probabilities are not going to be 1/6 for each die, but are going to be 1/4 for 1s, 2s and 5s, 1/6 for 6s, 1/12 for 3s, and a 4 will never show up.

There are 7776 possible combinations of 5-dice. That's an average of 64.30 times each in the dice file. However, there are not going to be exactly 64.3 occurrences of each combination in the file; probably let's say between 55 and 73 for the most part (and there may be some outliers). If the same file is used, those deviations from randomness will always be there (until the file is flushed).

I don't know about the cycling algorithm -- that actually may make it less random if lines are being crossed off and not eligible to be reused until all 500k lines had been used (if you take a deck of cards, and draw cards out one by one, and in the first 39 draws you happen to take out all of the diamonds, then in the last 13 draws, there is *no* chance to draw a diamond).

I'm not saying that it's necessarily predictable or gameable, although it may be, but the dice really aren't random under these circumstances even if random.org generates true random numbers.

[hecter]

I think you're being a bit picky...

[Bean_]

I actually don't think so. It's very important that there be die roll independence from roll to roll. A 3v1 attack should always win approximately 65.97% of the time. That is not the same as having it be that way 98% of the time, but the 2% of the time that one is down to the last 10,000 lines, it could be anywhere from 62% to 68% depending on how the first 490,000 lines went.

In fact, if lines are really being crossed off and ineligible for a roll, that could account for some of the wildness that is being reported anecdotally. It's very easy for this to happen if someone makes rolls when the system is down to the last 1,000 rolls or so.

I don't know the details of the algorithm. The only real right way to do it is to take the die rolls once, and not reuse them. Recycling them linearly once or twice might be okay (though not perfectly random). CC shouldn't use an algorithm that by design generates black swan events, which really do ruin games if they happen more often than they should by true randomness.

[owenshooter]

I don't know the details of the algorithm.

measurements are 34-24-34, about 5'6", soft blue eyes, longish
dark hair... the algorithm is a taurus and enjoys long walks on the beach, cappucino with dessert, reading at night, naps, receiving foot rubs, discussing politics and the op ed's in the times... overall, the algorithm
is pretty amazing, and just wants to be held at night and told how special
it is...-0

[Twill]

I actually don't think so. It's very important that there be die roll independence from roll to roll. A 3v1 attack should always win approximately 65.97% of the time. That is not the same as having it be that way 98% of the time, but the 2% of the time that one is down to the last 10,000 lines, it could be anywhere from 62% to 68% depending on how the first 490,000 lines went.

In fact, if lines are really being crossed off and ineligible for a roll, that could account for some of the wildness that is being reported anecdotally. It's very easy for this to happen if someone makes rolls when the system is down to the last 1,000 rolls or so.

I don't know the details of the algorithm. The only real right way to do it is to take the die rolls once, and not reuse them. Recycling them linearly once or twice might be okay (though not perfectly random). CC shouldn't use an algorithm that by design generates black swan events, which really do ruin games if they happen more often than they should by true randomness.

I'm going to guess you know a lot more about probability/randomness than I do (I just sit here and [don't] look pretty)

Alls I know is that we don't run an algorithm ourselves, we let random.org do that (they do it much better than we ever could). And, that as the dice data set approaches large numbers the pool which they are pulling from (in a perfect world, an infinitely large pool) creates a system where the numbers are random enough...computer generated randomness never is exactly random, I'm sure you know that. I'd think that a half million rolls is a large enough pool (how many people have rolled that many dice in real life other than casino hosts?)

That being said, if you think you can design a more random dice set (or at least one that appears more random whilst still being as random as random.org) then please do get in touch with me privately so that we can discuss it.

Have a good one
Twill

[Bean_]

There are two levels at which numbers are being selected/generated. One is from random.org, the other is CC's use of the numbers that come from random.org. You can break randomness at either level.

So far I've been assuming that the numbers from random.org are truly random (they claim to be, and I've run some quick tests in the past from numbers provided directly from random.org). What I am concerned about is whether CC uses those numbers in such a manner so as to affect the randomness of the random.org-generated numbers.

One way this might happen is a reuse of the same random numbers over and over again. If you have 1000 random dice lines, they are random the first time you use them, but if they are reused (let's say in perpetuity, to take the easiest case), then they are no longer random, because the future population will always be tainted by the peculiarities in the sample. You wouldn't be picking based on perfect odds, but based on a finite sample that does not mirror those odds. The size of the sample matters insofar as it's more likely that the sample approaches perfect odds the larger the sample is.

Another way -- which could be a larger problem -- is if lines are being discarded as used. Again, the problem is that the remaining sample becomes more and more finite, with the odds becoming more and more distorted as the remaining sample becomes smaller. I don't know that this happens, but your previous responses could be interpreted that way.

Anyway, I'd like to know more about how the system works (i.e., CC's algorithm for using/selecting numbers, not random.org's generation processes) just in trying to improve it if possible (or if needed). If you are amenable and prefer the discussion take place privately, that is fine too.

Thanks.

[HappyCamper]

What the hell will you just ROLL PLEASE

[iamkoolerthanu]

Bean, if I understand correctly, you are saying that if I take a bunch of number, pick them randomly, and then re-use tose same numbers, they are no longer random. In reality, they are. If I take 6 numbers and put them on a dice, it is 1-6. Now if I roll that dice, it is random. If I re-roll that dice, it is just as random. If I roll it 5 more times, then roll it again, it is still completely random.

Now, what you were saying before about the numbers being 'corssed off', than that would no longer be random, you are correct on that. Because if I took that dice and rolled a 5, then I no longer could roll a 5 again untill the 1, 2, 3, 4, and 6 were rolled, than that was not random anymore.

I'm not sure which one you are trying to get across, because form post to post your opinion seemd to change.(to me at least. )
And I hope that the usage of smaller numbers would be easier to understand for some people, because it definately(spell) helped me out.

[Bean_]

Bean, if I understand correctly, you are saying that if I take a bunch of number, pick them randomly, and then re-use tose same numbers, they are no longer random. In reality, they are. If I take 6 numbers and put them on a dice, it is 1-6. Now if I roll that dice, it is random. If I re-roll that dice, it is just as random. If I roll it 5 more times, then roll it again, it is still completely random.

Now, what you were saying before about the numbers being 'corssed off', than that would no longer be random, you are correct on that. Because if I took that dice and rolled a 5, then I no longer could roll a 5 again untill the 1, 2, 3, 4, and 6 were rolled, than that was not random anymore.

I'm not sure which one you are trying to get across, because form post to post your opinion seemd to change.(to me at least. )
And I hope that the usage of smaller numbers would be easier to understand for some people, because it definately(spell) helped me out.

It's both, and my opinion hasn't been changing -- I just don't have all the facts so I'm guessing from what Twill has been saying. Both could be issues.

On the first issue, if you roll 5 lines of dice randomly (a line being 5 rolls, 3 attacker and 2 defender), and get

6,2,3 vs. 4,4
2,1,4 vs. 3,6
3,2,5 vs. 1,6
6,1,4 vs. 2,5
3,4,5 vs. 2,1

And then restrict yourself to using those same 5 lines over and over, then although they were random the first time you generated them, in the future you're limited just to using those 5 lines, each of which has a 20% chance of occurring.

[iamkoolerthanu]

Oh, I understand now. You are saying if you get a set of random numbers, then use those random numbers over and over again, then it wouldn't be random anymore. Then you are correct on what you are saying.

[fireedud]

There is more randomness than you think there is, not every time somebody attacks, 5 dice are used. Also, the dice may be broken up.



Lets say several people click the attack button within the same nanosecond( remember, it's almost impossible for something to be simultaneous). Then the first computer to connect to the server starts getting the dice it needs, but is interrupted by another computer, which in turn is interrupted by a third computer, etc.

[Kalten]

This is bad.

If they cycle over, this is a definitive admission that the dice are obviously not random. This is actually somewhat disconcerting. There is a systemic bias in the dice (i.e., the outcomes will not match the exact probabilities that they should, and there are never "new" dice that are, a priori, random). Also, in theory someone who auto attacks 10,000 on 10,000 might be able to get a significant portion of the dice sequence and use it for later. (Yes -- it isn't easy to game the system that way, but that's how people win at blackjack, in essence.)

Without saying too much about how the system works...it doesn't work that way

The dice file is random, thus meaning that if we cycle over it is still random...it's just the same random random numbers over again.

It is also impossible to predict/reserve which number you are getting, even if you had access to the file and a millisecond timer and trigger finger (there are what, 10 dice rolls ever second?).

The system doesn't just "cycle through" the numbers. there is a bit more randomness added in and the randomness which is added in is in fact based on randomness pulled from a random source. (that makes it several layers of random deep I think)

Aaaaaaaand, the dice file is replaced from time to time just to keep things interesting and add more randomness to the system.

Hope that puts your mind at ease there a bit

Have a good one
Twill

that makes no sense wht so ever

[iamkoolerthanu]

There is more randomness than you think there is, not every time somebody attacks, 5 dice are used. Also, the dice may be broken up.



Lets say several people click the attack button within the same nanosecond( remember, it's almost impossible for something to be simultaneous). Then the first computer to connect to the server starts getting the dice it needs, but is interrupted by another computer, which in turn is interrupted by a third computer, etc.

That is very true, and something I didn't even think of. Lol.

[Bean_]

There is more randomness than you think there is, not every time somebody attacks, 5 dice are used. Also, the dice may be broken up.



Lets say several people click the attack button within the same nanosecond( remember, it's almost impossible for something to be simultaneous). Then the first computer to connect to the server starts getting the dice it needs, but is interrupted by another computer, which in turn is interrupted by a third computer, etc.

That is very true, and something I didn't even think of. Lol.

Since they have always spoken in terms of "lines" of dice, my guess is that an entire line of 5 dice is taken each time, with the "unused" dice being discarded.

The argument that several different computers are connecting at (nearly) the same time addresses the linearity issue in part (it's equivalent to "hashing," I believe -- in any case, Twill has said that the lines are not taken linearly from the file), but would not affect the potential "finite pool" or "crossing out" issues.