So...If the dice are Truly Random

[jusplay4fun]

We do not have objective memories. Our memories are impacted by our emotions of the moment. We remember painful losses; that is why they Teach US WELL. We do not forget those painful lessons.

The ones we win against such odds are forgotten much more readily.

SHOW me numbers that actually prove "Crooked Dice" other than a random and anecdotal citing of one such turn or battle. ALL dice complaints that I have read are all anecdotal citing of one such turn or battle.

Show me dice stats for SEVERAL TURNS and not jus your recollection. PROVE your allegation of "Crooked Dice."

Over several moves (and that may be 10 turns or battles or up to 30 or 50) dice odds balance, with the rare random streak one way up or down for 51 battles or turns. It is the nature of random dice. Did you ever the origins of the science and math of Statistics and Probability? LOOK into it. You might learn something.

Most such postings of bad dice are emotional rants of upset players, not rational thinking of objective persons.

[iAmCaffeine]

The dice system used is not a good one, but half the people complaining don't even understand why, and they don't understand what random means. If you want to read comments without your brain hurting, ignore people like jusplay4fun.

[Evil Semp]

I think one of the reasons people think the dice are so bad is that they don't see their defensive dice. I know my problem is not stopping when dice are bad. You attack 10v2 and lose because you roll just once more rinse and repeat.

My challenge to people is to find a better dice program for the site to use. I don't think it would be that easy.

If there is a better program out there why wouldn't admin use it?

[degaston]

I think one of the reasons people think the dice are so bad is that they don't see their defensive dice. I know my problem is not stopping when dice are bad. You attack 10v2 and lose because you roll just once more rinse and repeat.

My challenge to people is to find a better dice program for the site to use. I don't think it would be that easy.

If there is a better program out there why wouldn't admin use it?

Because the admin no longer cares about the site, and hasn't done anything to improve it in years?
bigWham's most recent post is Jul 20, 2019.
The one before that was Sep 25, 2016.

Back in 2013, it was shown that the dice were not at all random, and did not have an equal probability. The last I heard, they fixed the probability, but not the randomness. A simple pseudo-random number generator would certainly be better than what they're doing.

[jusplay4fun]

As I already said, get data to support your bad dice complaints. You offer NONE.

It seems you can't handle facts or the truth. Keep letting your ego and your head hurt. Better than alcohol or drugs. I think your brain hurting is due to you thinking too much; your brain apparently does little of that. And keep complaining; we all love to hear you complain. It is MUSIC to our ears. OK, enough of the ad hominem comments; now to return to our regular programming.

Read on and you will read that a better dice program is being requested and no one has yet suggested one. Do you have any suggestions there?

The dice system used is not a good one, but half the people complaining don't even understand why, and they don't understand what random means. If you want to read comments without your brain hurting, ignore people like jusplay4fun.

[DBandit70]

well not sure but I play risk and never see dice like CC's dice

[nietzsche]

i don't know the current state of the thing but.. the problem used to lay in:

a batch of random numbers were gotten from random.org, and when this was used up they got another one. for some reason, at some point they simply started to use the same batch over and over.. now this might not seem to be a problem if you consider you advance a pointer the batch after every use, say you use 5 numbers for an attack 3vs2, then use 4 for an attack 3vs1, etc. the pointer is then moving and given that it's highly unlikely certain player will get the same set of numbers even for an autoattack with some 1000 numbers used, because timing and whatever.. but i once asked and i was told that it wasn't the case, that the numbers were used 5 in 5, and if it was a 2vs1 attack it simply discarded the other 3 numbers.

so, it was more like in sets of 5, or lines of 5, however you want to picture it.

now, even if that was not the case, the random batch most likely has what for our case would be interpreted as streaks.. say that for a random reason, there's a big space in the batch that has numbers like 2,1,2,6,4 3,1,1,4,5 2,1,4,6,6... something like that.. for some 20 times.. imagine going 40vs2 and losing... that's not a problem ok.. it's part of the randomness.. statistically unlikely but it happens... but the thing is, it will happen again the next time someone does an auttoattack and the pointer is in that position of the batch


so, in a sense, using the same batch over and over is not truly random. however, i don't really see an issue with it, it is what it is..and it's big enough.. and unlikely enough. it would be better if the numbers were used one by one instead of 5 by 5 though, that would make it much more random, in fact, it would work as if the batch was 5 times bigger.


there seemed to be another problem, that for some reason they, at some point, decided to readjust the number of 1s.. i don't remember why was this.. but possibly because people were complaining they got too many 1s?

or, when it was shown that the batch had a larger amount of number 2s in it, or smaller, can't recall..


ideally, they should get a new batch every time.. but i'm not sure what the deal is here.. if they charge you a subscription or what.. maybe bigwham is mad at random.org because it used to be that it was a one time payment thing.. and suddenly they changed it to autorenewal and he didn't notice until he saw the credit card statement.

or if it's some technical problem with loading/unloading the batch in memory.. or simply they're too lazy to change the code no matter how easy it is.. or it's priority is low.. who knows, they never say anything.. you have to go by what you hear. even if they can't get another batch, applying some mathematical operation over the same batch, in smaller sets of numbers should be enough.. i know that wouldn't be super random, but given that the batch is already taken from the meassuring of particles in a 2dimes' fart, i think it's random enough.


there, i did more than caff in just saying most people don't know, i actually said what i heard.

[nietzsche]

well not sure but I play risk and never see dice like CC's dice

this is most likely a perception, because you would never in RL roll as many dice as you roll in here... you will be more likely to see things that have a small probability of happening.

here you just click autoattack many times a day and be done with it.. in RL you will roll the dice for 2 mins for each of those autoattacks.


but yeah, without analysing the batch of numbers it's impossible to see if it's not very streaky for all it's randomness. see, it's random.. if it contains large sets of

1,2,3,6,6
3,4,3,5,6
1,1,2,3,5
4,4,2,5,5
2,1,2,4,4
2,3,3,6,5

those numbers could be randomly generated, and yet for a risk attacker they suck.


so, again, to shut us up they should simply replace the batch of numbers once it's been used..

[degaston]

there seemed to be another problem, that for some reason they, at some point, decided to readjust the number of 1s.. i don't remember why was this.. but possibly because people were complaining they got too many 1s?

or, when it was shown that the batch had a larger amount of number 2s in it, or smaller, can't recall..

That all came out in this thread. Apparently, photobucket is down right now, so the graphs won't show, but yeah, they kept re-using one file and the data was slightly skewed to favor some numbers over others.

I either hadn't noticed, or forgot about them choosing 5 values for each attack no matter how many dice are actually needed. If the file-size is 50,000 rolls, then there are actually only 10000 possible sets of dice rolls that are available for the game. That's not a lot considering there are 7,776 possible sequences of 5 dice, so some roll combinations are going to be over-represented, and it's possible that some will never occur. Obviously, many combinations have exactly the same outcome, but they still have to be counted, because changing the number of times a combination can appear will affect the odds of that outcome happening.

Each time you attack, you should have a (2890/7776) = 37.165...% chance of winning 2, a (2275/7776) = 29.2567...% chance of losing 2, and a (2611/7776) = 33.5777...% chance of losing 1 each. But that's impossible with this system because you can't get more than two decimal places of precision out of 10,000 possible rolls. And I would bet that it's currently much worse than that, because variation is natural with truly random numbers, so the particular set they're using could be very far off from the win/loss percentages that people are counting on when they play.

But that's not even the worst of it. The real problem is when you think about sequences of winning and losing. For truly random numbers, as an attacker, you should expect to lose 8 battles in a row about once every 18,612 attacks. But with only 10,000 sets of rolls, you can never even come close to the expected rate for that result. If the dice data has just 1 such sequence, then the loss of 8 battles in a row would happen nearly twice as often as it should. Maybe such a sequence doesn't appear at all. Maybe it appears more than once. Who knows? Has anyone analyzed the data to see how often long sequences of winning or losing appear in the data? I doubt it. And with only 10,000 possible rolls, it would be impossible to balance all the sequences that should appear naturally, each with their own probability, in a manner anywhere close to what would be expected from truly random dice.

Things are somewhat better if you never use auto-attack (which I don't), so that you may skip over some rolls in the data set that are used by other players. It's usable, but as a programmer, I think it's an amazingly stupid system for generating dice.

[nietzsche]

following the link you posted, and the subsequent links in there.. some things are clarified


before mid 2010, this is how it worked:

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.

it was then modified "thanks to some prodding and advice from Dako, sherkaner and jakewilliams" to work like this:

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

we know that it was working the same around the beginning of 2014, and we have no reason to believe it changed. Then only thing in my mind left is that little change that i read somewhere was made that accounts for the 1's issue.


IMO, for randomness 500,000 sets of 5 is better than 50,000 individual numbers , but this is solely based on the idea of an offset at some point that leaves the pointer at a starting point different than one that's multiple of 5 and the rest of it moves in 3vs2 attacks.. i just simply can't come up with a way to calculate probability of the amount of different combinations, given that some attacks will use only 2 numbers, others 3 and others 4 (the vast majority will still be of 5 numbers though)

considering a lucky offset quickly, and subsequent 5 number attacks, we would be thinking of 10,000x5, which is a tenth as much possibilities than the previous 500,000 lines of 5 numbers (2.5million numbers)

it would be interesting to hear what was the "advice from Dako, sherkaner and jakewilliams".

perhaps it has to do that it isn't as predictable? some times you won't get a certain combination for a lot of passes on the batch. but 10,000 possible combinations are too low if we consider current escalating spoils numbers, even though the amount of games played isn't as much as in the past. I think a the batch is used up some 10 times a day at least.

considering that, it's understandable they don't get numbers from random.org every 2 hours or so, they have a quota limit, but it isn't undersatndable how they can use a large batch as a seed as degaston mentioned. i think i'd prefer the previous 500,000 lines from before, and every week or so this is used as a seed to get more numbers.. or pull more numbers from random.org steadily and change the batch when it's possible.

--

as a former programmer, it also boggles my mind why some things aren't fixed.. but then remember sometimes.. you just don't know what the issue is.. but meanwhile, since they don't say shit, we will just assume they're being lazy.

[Nymeria Stark]

The maths ignorance is strong in this thread.

[nietzsche]

i do not claim to know much about math or probability.. i know there're formulas that could be used i just don't remember the theory to apply them.


i do not claim it's not random.. ish. some aspects of it are random, others are not so much. it could be solved, if pulling a new set of numbers isn't an option, by using the same set of numbers as a seed to get new numbers (as degaston mentioned). it would still be dependant on the first set of numbers but it would be better.


if i made a mistake, point it out.

it's not that i'm super humble and mistakes don't affect me, it's that i'm awesome and part of being awesome is not caring about mistakes i make because i'm awesome anyway.

[NomadPatriot]

dice roll generator -->

https://academo.org/demos/dice-roll-statistics/

you can select # of dice then click autoroll..

average roll of (2) 6-sided dice is a 6-7 for 1,000 rolls
6-7 out of possible 12 is... 3.5 average

Nomad = A Better Tomorrow

[NorthmanII]

I have been on this site for 10 years. For the most part the dice have never been what I would classify as "normal" I have seen probably 15 different players post the same thing in chat during a game which is when they get to a certain rank it seems the bottom falls out and their dice go south for months on end. I know for a fact I have seen that happen probably 4 times and have currently seen it happen for the past 4 months to me. For me its every time I get into the 1900's. Unbelievable dice where attack stats for 25 rolls stays in the negative for a week on end. There is nothing that can be done when this happens. I am simply riding it out but at this point it has destroyed all that I worked to get up to and it does lend to having less fun when the dice are so one sided against you.

[Kaiser Szami]

I'd like to add some thoughts here, it's regarding the dice stats, in this case the last 5 battles, cause these are the stats I'm most used to. If dice are really random, why are stats always moving up and down like waves in a stock market chart? I assume other players also noticed that... Normally stats are going in a range of -20/-30 and +20/+30, sometimes also a little or way more down/up, but after a down phase there's always an up phase and vice versa. In my eyes this shows quite clear that dice are NOT random...

[berlin1945]

IF the dice are truly random, why am I the only one having bad luck out here? Why does everyone kill my big stacks with no losses, but I can't kill any of theirs? The dice aren't random for me, I know that for sure.

[hajduk1911]

COVID-19

[berlin1945]

I go long stretches of the dice being fairly neutral. Then they change, and i start losing 9 to 1. I have some games where i could sweep with neutral dice, but right now i may as well just not even bother. Does this happen to anyone else?

[Silly Knig-it]

I go long stretches of the dice being fairly neutral. Then they change, and i start losing 9 to 1. I have some games where i could sweep with neutral dice, but right now i may as well just not even bother. Does this happen to anyone else?

Wait until you get a month or two of dice like that every time there is an important roll. Of course, then there are the times when one cannot lose. (Those are fun.)

[berlin1945]

Two months! That would drive me mad. I'd walk away from the game. It's not worth it.

[Dukasaur]

Silly HAS walked away from the game.

But he came back...

[Silly Knig-it]

Silly HAS walked away from the game.

But he came back...

I did because I couldn't find anything better or a place with better people. Even the forums are fun to read after one foes the trolls.

[Dukasaur]

Silly HAS walked away from the game.

But he came back...

I did because I couldn't find anything better or a place with better people. Even the forums are fun to read after one foes the trolls.

Always glad to see you!

[Rotean9]

I had a recent run of luck that was hard to believe. I was promoted to lieutenant for the first time! Then I started having bad dice rolls. I mean, really bad. In every single game I play in. It hasn't stopped. My dice rolls are beyond pathetic. I lose to one army eighty percent of the time. I get - dice rolls almost every game. Like -60. I'm so tired of it because the games are not fun anymore. The dice also have patterns. They are fixed. The Dice, in my view, have been awful for about a year. They used to be much better. Now, you either win every roll or nothing. CC should be about skill and military tactics. NOT just the dice. Feeling frustrated and tired of bogus DICE....

[rockfist]

Does a bear shit in the woods?

There is not one person with more than 20-25 games played who hasn't had a poor run of dice, you just hope it comes in games of lesser importance.