New "Intensity Cubes"

[carlpgoodrich]

This makes it more random and makes it much less likely to duplicate a roll. Jake

No, this does not make it "more" random (random is a binary state, you either are or you aren't). In a list of random numbers, there is the same correlation (i.e. none) between the first number and the second as there is between the first and the third. Throwing away some numbers does not affect the randomness of the dice. The only advantage I can see is that the new way is slightly more efficient, but if that was really a problem that could easily be solved with a combination of a sudo-random number generator and the "true" random numbers from Random.org.

However, I do have some concern with the new changes (maybe I am not understanding Lack). It sounds to me like the same list of 50,000 numbers is used repeatedly. Is this within the same game or over all games? And how often does it get changed? My concern is that when you start repeating numbers you open yourself up to abuse. I doubt anyone would go through the trouble of mapping the list just to win at CC, but it is certainly doable.

[edwinissweet]

*head explodes*

yea, my mind is pretty blown. Im not bright enough to understand how it works.

[Knight2254]

This makes it more random and makes it much less likely to duplicate a roll. Jake

No, this does not make it "more" random (random is a binary state, you either are or you aren't). In a list of random numbers, there is the same correlation (i.e. none) between the first number and the second as there is between the first and the third. Throwing away some numbers does not affect the randomness of the dice. The only advantage I can see is that the new way is slightly more efficient, but if that was really a problem that could easily be solved with a combination of a sudo-random number generator and the "true" random numbers from Random.org.

However, I do have some concern with the new changes (maybe I am not understanding Lack). It sounds to me like the same list of 50,000 numbers is used repeatedly. Is this within the same game or over all games? And how often does it get changed? My concern is that when you start repeating numbers you open yourself up to abuse. I doubt anyone would go through the trouble of mapping the list just to win at CC, but it is certainly doable.

You could never do this as there are hundreds of attacks happening all at once.

[danfrank]

I really do not understand this concept either. BUT.. Lets just do some simple math...


If the attacker has 3 dice and the defender 2 then at most there are 7776 different combinations... I come to this conclusion by mutliplying 6 x 6 x 6 x 6 x 6 .. So why is there a list of 50,000 thousand numbers which would logically suggest duplication ?


Also what is intensity level, and what determines this factor.. My belief is, this is a script that is written by said master and fused with random. org number generator. If this is the case then it is fair to say that certain members can have a more favorable equation in said script then others .

[stahrgazer]

I really do not understand this concept either. BUT.. Lets just do some simple math...
If the attacker has 3 dice and the defender 2 then at most there are 7776 different combinations... I come to this conclusion by mutliplying 6 x 6 x 6 x 6 x 6 .. So why is there a list of 50,000 thousand numbers which would logically suggest duplication ?

Agreed. The most random would be each roll generates one of 7776 combinations; the next number would pick from 7776 combinations.

However, the way they're doing it does help identify why people get series of really good rolls and series of really bad rolls; and for some, the 'randomness' doesn't necessarily equal out. Basically, the accusations that someone else got all the good dice can fairly true, if all the "good numbers" got used up before they reload the entire 50,000 combinations.

[carlpgoodrich]

This makes it more random and makes it much less likely to duplicate a roll. Jake

No, this does not make it "more" random (random is a binary state, you either are or you aren't). In a list of random numbers, there is the same correlation (i.e. none) between the first number and the second as there is between the first and the third. Throwing away some numbers does not affect the randomness of the dice. The only advantage I can see is that the new way is slightly more efficient, but if that was really a problem that could easily be solved with a combination of a sudo-random number generator and the "true" random numbers from Random.org.

However, I do have some concern with the new changes (maybe I am not understanding Lack). It sounds to me like the same list of 50,000 numbers is used repeatedly. Is this within the same game or over all games? And how often does it get changed? My concern is that when you start repeating numbers you open yourself up to abuse. I doubt anyone would go through the trouble of mapping the list just to win at CC, but it is certainly doable.

You could never do this as there are hundreds of attacks happening all at once.

Give me on the order of 50,000 rolls (they don't have to be consecutive) and a day or two to write the code and I'll do it (note to the mods, I'm not going to:) ).Like I said, I don't know how often the list of 50,000 numbers is replaced, so if its often this won't be relevant. But if it is not replaced regularly, it doesn't matter that there are hundreds of attacks happening all at once. All you need is a large number of samples from that set of 50,000. For example, if there are a larger than expected number of repeated numbers in the set than defending is slightly more favorable than it should be. There are tons of things like this that you can write a program to analyze, and if you have enough statistics it can tell you things like "be aggressive" or "don't be aggressive".

[hulmey]

Wow, that was rather quick. Too bad I don't have a stacking game to test out 10k attack.

true randomness cannot be judged with such a small amount. test a couple of million

[jleonnn]

big words!!!!! ahhhhh!!!!!!!!!!1

[n00blet]

Awww mannnnn

Don't get me wrong, this sounds nice and all, but when I read the headline's title I was for a moment convinced that Lack had taken the "bunnies holding up placards" idea seriously...Needless to say, I was ecstatic.

[MadagascarAdam]

Yeah, seriously, I'm wondering why they have to choose numbers consecutively in a sequence? Is it really that hard to get it to choose one random set from the sequence and then another random set from the sequence? Why does it have to choose the number in order?

Anyway, I'm definitely confused now. All along I thought my dice really were random, i.e. no one would ever know what was coming next. But now I'm thinking it is a bit rigged.

Then again, I don't know what kind of feats of processing strength are required to make this thing random, and I guess a list of numbers that were generated randomly is at least better than nothing. Can the starting point in the list not just be randomly selected for each individual set of rolls though? Why does it have to cycle through sequentially for the sum total of all ConquerClub rolls?

[sherkaner]

Yeah, seriously, I'm wondering why they have to choose numbers consecutively in a sequence? Is it really that hard to get it to choose one random set from the sequence and then another random set from the sequence? Why does it have to choose the number in order?

Anyway, I'm definitely confused now. All along I thought my dice really were random, i.e. no one would ever know what was coming next. But now I'm thinking it is a bit rigged.

Then again, I don't know what kind of feats of processing strength are required to make this thing random, and I guess a list of numbers that were generated randomly is at least better than nothing. Can the starting point in the list not just be randomly selected for each individual set of rolls though? Why does it have to cycle through sequentially for the sum total of all ConquerClub rolls?

Yes, it's hard to get truly random numbers, and choosing a random set from the sequence would involve random numbers again, and thus have about the same disadvantages.

The methods computers can generate are pseudo-random, meaning there is some correlation between the numbers. If translated to the dice, it would typically be behaviour like
'if the first die is a 6, then the 4th die will have a 30% chance to be a 1'.

That being said, I'm not truly happy with the smaller sample of dice.

[Dako]

Let me try to explain some moments.

First, it is way (I mean very very) cheaper to keep numbers in memory then to generate them over and over at each roll request.

So now CC has some sort of pregenerated random numbers (let's say 1mil numbers). They look like a string 122546352635215 ... If CC need to roll - it just picks up new dices from the memory and serves them to user. When the list gets to an end - it rewinds the list to the start and reiterates. I doubt CC changes those numbers often.

The dice are still random and each processor thread (let's say 10 threads for CC) has it's own cope of the dice string. So each dice request uses this random dice and there are more than 10 mil attacks each day (means more than 50mil dices are thrown each day). And you never know on which position you are right now - and this position always changes and to predict whether the dice would be good or bad is impossible or extremely hard.


Did I make it any clearer for you? Any more questions?

[pearljamrox2]

this is how i thought it should have been . good change lack.

[carlpgoodrich]

Dako, there is a big difference between 1 million and 50,000. If it is just 50,000, then an auto-assult with around 10,000 vs 10,000 will give you the entire string of numbers in order. Like you said, you will still have no clue where on the string you are at any given time, but you can still calculate statistics about that string to give you a slight advantage.

Can I make a practical suggestion? If you are storing these numbers and just reusing them, why not use a pseudo-random number generator? sherkaner, I don't know what generator you are talking about, but I use pseudo-random number generators all the time for work and there are very very good ones. The correlations you mentioned simply do not exist at the level we are talking about (they run into some trouble if you need to pick from over a billion integers... if the first number is 372,936,028, the chance that the next one is 372,936,029 is about twice what it should be. Please believe this correlation will make no difference in CC where you only need to pick from 6 integers). Anyways, pseudo-random number generators are very inexpensive computationally (they amount to a few if statements), and you can get a new seed from Random.org every couple minuets if you want to make sure people don't complain about it not being "true" random.

[Dako]

I do not have the exact knowledge of how many numbers CC uses. And also, even knowing the whole string... 1 roll shift will change the entire effect on the rest of the rolls. So there is no way to calculate if you have an advantage or no.

I guess CC numbers are pseudo-generated from random.org, but it is faster to access them from memory rather than to generate them for each time (even with few if/else statements). It is all about speed here. 1mil attacks per day = 11 attack per second, 40-50 rolls per second. I guess it will be quite hard to call pseudo-generator for each roll on the fly. Even if it is just 5 if statements - it will be 250+ operations per seconds just from rolls. And we distribute here all the dice evenly but everyone knows that American daytime is a peak of rolling.

So.... I think keeping them in memory is better, and the real count of CC numbers is not that relevant if it is bigger than 10k.