How does using a dice file differ from rolling?

[Hoots]

Im interested in any opinions (or facts) about using a dice file like CC does. Does it mathmatically differ in any way from actually rolling dice? Are long and short term odds/probabilities effected in any way in comparison?

Does each player have their own file (dont think this is the case) or is it one massive file the entire site pulls from.

If we are all pulling from the same file wouldnt that drastically effect probabilities? (both long and short term)

If you dont want to be constructive please dont post. I would just like to better understand the differences (if there are any).

Thanks

[lord voldemort]

Dice are from random.org
There is a ff add-on that analyzes your dice. I personally havent done it but i know others have and it has been pretty close to 1/6 for each number

[trinicardinal]

Im interested in any opinions (or facts) about using a dice file like CC does. Does it mathmatically differ in any way from actually rolling dice? Are long and short term odds/probabilities effected in any way in comparison?

Does each player have their own file (dont think this is the case) or is it one massive file the entire site pulls from.

If we are all pulling from the same file wouldnt that drastically effect probabilities? (both long and short term)

If you dont want to be constructive please dont post. I would just like to better understand the differences (if there are any).

Thanks

I think it is that we are all pulling from one massive file and , yes that would affect the results (streaks?),... hence some of the strange things people see. my opinion only but hey, it would explain a lot

[Hoots]

I think it is that we are all pulling from one massive file and , yes that would affect the results (streaks?),... hence some of the strange things people see. my opinion only but hey, it would explain a lot

Well, this is exactly the kind of stuff that i want to discuss in here.

When talking about odds and probabilities we typically are talking about our OWN "virtual dice file"... So I if roll a 6-6-6 with real dice these three 6's come out of my own imaganary dice file which changes the probabilities of my next roll. Your rolls, the guy playing craps in vegas or the kids rolling dice on the corner have no effect on my dice probabilities.

With one massive file we all are changing each others probabilities and even odds. (even if it is so slight). Changing the odds is the interesting thing to me. On an equal dice 1 can roll a billion times and every single roll the odds of rolling a 6 is 1/6. With a dice file the actual odds of pulling numbers are being changed.

I have no idea how many numbers are in a dice file but lets just say 6 million as an example (1 million of each number 1-6).
So my 6-6-6 roll now leave only 999,997 6's for everyone else... (which is less than 1/6) Yes we are talking about very small difference but over time i can see how these small difference would add up.

How does this change this game? Thoughts?

again, im just looking for peoples thought and discussion points here. not OMG haxx and the site is fixed!

[trinicardinal]

I think it is that we are all pulling from one massive file and , yes that would affect the results (streaks?),... hence some of the strange things people see. my opinion only but hey, it would explain a lot

Well, this is exactly the kind of stuff that i want to discuss in here.

When talking about odds and probabilities we typically are talking about our OWN "virtual dice file"... So I if roll a 6-6-6 with real dice these three 6's come out of my own imaganary dice file which changes the probabilities of my next roll. Your rolls, the guy playing craps in vegas or the kids rolling dice on the corner have no effect on my dice probabilities.

With one massive file we all are changing each others probabilities and even odds. (even if it is so slight). Changing the odds is the interesting thing to me. On an equal dice 1 can roll a billion times and every single roll the odds of rolling a 6 is 1/6. With a dice file the actual odds of pulling numbers are being changed.

I have no idea how many numbers are in a dice file but lets just say 6 million as an example (1 million of each number 1-6).
So my 6-6-6 roll now leave only 999,997 6's for everyone else... (which is less than 1/6) Yes we are talking about very small difference but over time i can see how these small difference would add up.

How does this change this game? Thoughts?

again, im just looking for peoples thought and discussion points here. not OMG haxx and the site is fixed!

Actually I don't think it changes the odds (it's still 1 in 6) but it will change the amount of times that 1 in a million oddball factor (and especially the streaks) show(s) up. it's kinda like the real game.... you're all using the same dice on the board... if one player rolls all 6's does that mean you are less likely to roll a 6 on your turn? Not really... because future rolls don't depend on past rolls. its the same idea with the dice file (IMO).

[Hoots]

Actually I don't think it changes the odds (it's still 1 in 6) but it will change the amount of times that 1 in a million oddball factor (and especially the streaks) show(s) up. it's kinda like the real game.... you're all using the same dice on the board... if one player rolls all 6's does that mean you are less likely to roll a 6 on your turn? Not really... because future rolls don't depend on past rolls. its the same idea with the dice file (IMO).

I think this is one thing that is differnt. The odds HAVE TO BE CHANGING.

Im we assume a dice file starts with the same amount of every number (6,000,000 = 1millon individual 1,2,3,4,5,6's) then the first roll is a straight up 1/6 Odd to get a 6.

If i roll 6-6-6 vs 6-6 (bc i believe im actually pulling 5 numbers every roll even though they belong to the defense) there are now 5 less 6's in the pool. A dice file is a fixed amount unlike our "virtual dice file" that is endless.

So mathmatically in the above example i now hand the dice (file) to someone and he is pulling numbers from the remainingg 5,999,995 lines in the file. There is 1 million of everything except 6s so even if it is very small his odds of pulling a 6 is less than 1/6.

When we roll dice the only we change is our probabilities, never the odds. (for real dice that is).

Perhaps the dice file is large enough to round out the small odds differences.

It is interesting to me.

Other thoughts?

[karelpietertje]

Hoots, you make a good point about the odds changing a little bit when pulling say a 6 out of a pool.

However, the pool the dice are taken from could be considered as infinite. not because I think a million is close to infinite, but because there is actually not really a pool with 1/6 sixes, 1/6 fives etc, like you describe, but a huge sequence of 1s, ..., 6s.
no single pulled dice says anything about the other ones that have not been pulled yet.

[mibi]

people should start using the preface INAM (I'm not a mathematician, but...) in their posts.

[trinicardinal]

people should start using the preface INAM (I'm not a mathematician, but...) in their posts.

Good point so...

INAM but it seems to me that the dice file would have to be so large to run this site that pulling a few rolls out of that are going to have such a small effect to the percentages as to be negligible.... I mean how much difference is it to say that you have a 499,999 chance in 3,000,000 to roll a 6 instead of a 500,000 chance in 3,000,000 (just as an example that's what? about a 0.0000003 difference in probability? (feel free to correct my maths)

real random isn't likely to give you 500,000 6's in 3,000,000 rolls anyway

[khazalid]

The dice were tested from a pool of volunteers for any signs of 'streakiness' about a year ago. There were no statistically significant anomalies in the findings as far as I remember, although I'm certainly of the opinion that auto-attack does seem to jam up in either direction more frequently than single attacks.

[AAFitz]

The dice were tested from a pool of volunteers for any signs of 'streakiness' about a year ago. There were no statistically significant anomalies in the findings as far as I remember, although I'm certainly of the opinion that auto-attack does seem to jam up in either direction more frequently than single attacks.

Bruces war and I did a very small example of auto attack vs regular attack.

Personally, with big piles, I would always use auto attack and expect a different result.

In our game, we launched piles of 100 at each other.

The results were more inconclusive than you might have expected.

Ill grab them just to prove auto isnt always better...

Auto 100 v 100 = 2 left no kill
Auto 100 v 100 = 30 left after kill
Auto 100 v 100 = 21 left after kill
Auto 100 v 100 = 16 left after kill
Manual 100 v 100 = 11 left after kill
Manual 100 v 100 = 14 left after kill
Manual 100 v 100 = 35 left after kill
Manual 100 v 100 = 17 left after kill

[Hoots]

The dice were tested from a pool of volunteers for any signs of 'streakiness' about a year ago. There were no statistically significant anomalies in the findings as far as I remember, although I'm certainly of the opinion that auto-attack does seem to jam up in either direction more frequently than single attacks.

This isnt a witch hunt thread. Im just interested in what mathmatical differences there are.

I think we all can agree on this (assuming the dice file has equal amounts of every number)
1. A real world fair dice roll always has the same odds. 1/6. In a dice file with the above assumption, odds have to be changing by small amounts.


Can these small amounts compound? What happens when CC nears the end of the file? Does CC approach the end of the file and repeat? Have we got there yet? Do we get there every day?

Have any numbers ever been released about the file? How big is it, how many numbers, does it start with an equal amount of every number, etc?

[porkenbeans]

The dice were tested from a pool of volunteers for any signs of 'streakiness' about a year ago. There were no statistically significant anomalies in the findings as far as I remember, although I'm certainly of the opinion that auto-attack does seem to jam up in either direction more frequently than single attacks.

I have believed this to be true as well. So, I never use the auto-attack, unless I am in a situation that is very dyer, and only a miracle can save me.

[BigBallinStalin]

Maybe this is complete nonsense, but I'm very convinced that whenever I play Risk, the board game, I usually need about 1.5-2x (usually 2x) the amount my enemy has in order to kill him, and vice versa; however, with conquerclub, a slightly lesser amount (let's say ~1.3 or 1.5x) is usually sufficient to wipe them out.

Not sure why, but that's just how it's been over the the past year.

So, perhaps that means that the online dice are different, or are not truly random. As to why, I've got no idea, but I've always been suspicious of these internet dice, and have truly enjoyed the physical dice. Get those babies warmed up and BAM 665, BAM 654, BAM 652, BAM 664, and so on, and when ya need it. But the internet dice are lifeless, can't exert any subconscious or external energies from alternate planes into those suckers.

As for auto, don't trust it either, and it does seem to get stuck on those constant loss streaks more often, but once again, I really didn't take the time to gather any quantitative data, because honestly there are more important things to spend time on.

Perhaps, there's some automatic recorder for your dice as you roll them on CC. Then we can gather lots of info and maybe draw some conclusions, put these ideas to the test.

[e_i_pi]

IAAM, and I think as long as you have a decent random number generating function, I can't see the point of using a dice file. The mt_rand() function that ports with PHP passes the Diehard tests and has a period of 2^19937 − 1. Considering you can also seed the Mersenne Twister with the mt_srand() function, and use a seed that is a combination of non-replicable elements (such as microtime(), session live time, ip address, time between mouseclicks, last game number created, total number of games on xyz map, etc), then in my opinion, this is sufficient to generate randomness across small samples sizes, such as rolling a handful of dice in a game.

The main arguments against the randomness of the dice are along the lines of "omg I just lost 4v15 how on earth can that ever happen". Losses such as these are not at all impossible, as I have explained in many threads, and I would argue that it is far more believable that you could lose a 4v15 battle than it is that someone so goddam stupid as GWB could get elected President of the USA.

[trinicardinal]

IAAM, and I think as long as you have a decent random number generating function, I can't see the point of using a dice file. The mt_rand() function that ports with PHP passes the Diehard tests and has a period of 2^19937 − 1. Considering you can also seed the Mersenne Twister with the mt_srand() function, and use a seed that is a combination of non-replicable elements (such as microtime(), session live time, ip address, time between mouseclicks, last game number created, total number of games on xyz map, etc), then in my opinion, this is sufficient to generate randomness across small samples sizes, such as rolling a handful of dice in a game.

Ok so can we have that part again in English? aside fromthe fact that the complaints are generally just a waste of time and space (Other than in The Piece of Crap Dice Brigade of course )

[e_i_pi]

IAAM, and I think as long as you have a decent random number generating function, I can't see the point of using a dice file. The mt_rand() function that ports with PHP passes the Diehard tests and has a period of 2^19937 − 1. Considering you can also seed the Mersenne Twister with the mt_srand() function, and use a seed that is a combination of non-replicable elements (such as microtime(), session live time, ip address, time between mouseclicks, last game number created, total number of games on xyz map, etc), then in my opinion, this is sufficient to generate randomness across small samples sizes, such as rolling a handful of dice in a game.

Ok so can we have that part again in English? aside fromthe fact that the complaints are generally just a waste of time and space (Other than in The Piece of Crap Dice Brigade of course )

In plain English:

Multiply all these things together:

The number of nanoseconds since the universe began
The number of fundamental particles in the observable universe
The number of ways to arrange all the tiles in Scrabble
The number of legal chess positions
The number of combinations when dealing a deck of cards to 4 people

Once you get that number, multiply it by itself about 20 times.

That's about how often the Mersenne Twister repeats itself (ie - numbers generated before it is no longer random)

...and that's if you don't use a randomising seed.

[trinicardinal]

IAAM, and I think as long as you have a decent random number generating function, I can't see the point of using a dice file. The mt_rand() function that ports with PHP passes the Diehard tests and has a period of 2^19937 − 1. Considering you can also seed the Mersenne Twister with the mt_srand() function, and use a seed that is a combination of non-replicable elements (such as microtime(), session live time, ip address, time between mouseclicks, last game number created, total number of games on xyz map, etc), then in my opinion, this is sufficient to generate randomness across small samples sizes, such as rolling a handful of dice in a game.

Ok so can we have that part again in English? aside fromthe fact that the complaints are generally just a waste of time and space (Other than in The Piece of Crap Dice Brigade of course )

In plain English:

Multiply all these things together:

The number of nanoseconds since the universe began
The number of fundamental particles in the observable universe
The number of ways to arrange all the tiles in Scrabble
The number of legal chess positions
The number of combinations when dealing a deck of cards to 4 people

Once you get that number, multiply it by itself about 20 times.

That's about how often the Mersenne Twister repeats itself (ie - numbers generated before it is no longer random)

...and that's if you don't use a randomising seed.

Ok. I got it