Conquer Club

Dice Complaints and Various Suggestions to Fix Them

Suggestions that have been archived.

Moderator: Community Team

Random Number Generation

Postby Artius on Mon Apr 24, 2006 8:17 am

I know that this has been a hot topic in the forums in the past. I saw a link to an essay on Reddit that I thought might be of interest to Lack as well as the original posters.

http://www.qbrundage.com/michaelb/pubs/essays/random_number_generation
Private 1st Class Artius
 
Posts: 1
Joined: Thu Mar 02, 2006 12:36 pm

Postby lackattack on Mon Apr 24, 2006 8:56 am

Thanks for the article. I used the Mersenne Twister in the past (see come on now...just not frikkiin possible!!!) before I found out about random.org, which is the best because the numbers are generated from natural phenomena.
User avatar
Sergeant lackattack
 
Posts: 6096
Joined: Sun Jan 01, 2006 10:34 pm
Location: Montreal, QC

New Dice Generator

Postby AngryAnderson on Sun May 28, 2006 3:35 pm

I bet they have!

:x

Get's kind of ridiculous when defender rolls a 1 and you as attacker roll 3 1s.

Similarly the number of consecutive 6s rolled by defenders in many games defies probability.

Are there others that can be tried or is everyone happy with the current one?
You want to make alliances? Play doubles/triples games and stay out of my single player games lame brain!
User avatar
Corporal 1st Class AngryAnderson
 
Posts: 86
Joined: Sat Apr 15, 2006 10:09 am
Location: Norfolk

Postby Banana Stomper on Sun May 28, 2006 3:42 pm

i don't see your point. Ever play risk, rolling real dice. I played last night, and more than once the attacker rolled three sixes only to be foiled by a single six rolled by the defender. I don't have a problem with the dice on cc
User avatar
Corporal 1st Class Banana Stomper
 
Posts: 422
Joined: Thu Mar 16, 2006 4:39 pm
Location: Richmond, Virginia

Postby kingwaffles on Sun May 28, 2006 4:09 pm

No offense to anyone... But it really pisses me off when people complain about the dice. They are random! If you don't get the exact probable otucome then who cares??? It's random! There are thousands of dice rolls being ma de every day, of course there are going to be flukes like that, that's the lvely thing about chance!
Image
User avatar
Corporal 1st Class kingwaffles
 
Posts: 718
Joined: Mon Jan 23, 2006 9:05 am
Location: Pseudopolis Yard, Ankh Morpork, Discworld

Postby howie on Sun May 28, 2006 6:17 pm

Ya know Pearl Harbour happened because the Americans saw a low chance of the Japs attacking it, stop relieing on probablity it really is about chance
Private howie
 
Posts: 136
Joined: Fri Apr 28, 2006 3:01 pm

The dice

Postby Edward on Wed May 31, 2006 3:10 pm

I'm not saying there is a bug with the dice, but it seems to me that there are odd things going on when I play. Whenever I attack a contry that has 2 armies on it the contry defends very well, I usualy don't win unless i have more than 5 armies attacking it. In my understanding, If there is a battle where 3 armies where trying to attack a city or a contry with 2 armies defending it, the attacking force would have better chances. In another case, I had a force of 14 armys attacking a contry with 5 armies, he lost 2 of his and I lost all 14, now ok, ok, its unlikelly to happen but I admit its possible, and i ended up winning any way so its ok, but i just wanted to bring these things up because it does seem like sometimes, the dice aren't fair, especialy with those 2 armies, i don't know why but too me they defend like if they where 5 armies, please look over the programing for 2 army defenders to check for bugs.
User avatar
Sergeant Edward
 
Posts: 50
Joined: Mon May 08, 2006 4:10 pm
Location: Canada

Postby AndyDufresne on Wed May 31, 2006 3:51 pm

One thing people often neglect when they rant about the dice...is that most often than not you don't get to see how your opponents rolls against you. Logically speaking, if you the attacker has some horrible rolls, shouldn't anyone attacking you also? When looking only at the small sample pool of you attacking someone, the dice can seem way out of proportion. But once you take everything into account it balances out in the long run. Sometimes though there are streaks of bad luck, and streaks of good luck. Develop a routine perhaps before every dice roll--much like a preshot routine in golf. Maybe then it will feel like you can somewhat control the dice when they are 'misbehaving'. ;)

--Andy
User avatar
Corporal 1st Class AndyDufresne
 
Posts: 24919
Joined: Fri Mar 03, 2006 8:22 pm
Location: A Banana Palm in Zihuatanejo

Postby AngryAnderson on Wed May 31, 2006 4:35 pm

... but you always see what your opponent rolls in defence.

Hey if I hated it so much I'd move on and leave the site alone.

However I merely feel the random number generator isn't so random, and am just hoping that maybe we can try another at some point.

Then I can stop getting my backside kicked by rolling 3 1s in attack!
You want to make alliances? Play doubles/triples games and stay out of my single player games lame brain!
User avatar
Corporal 1st Class AngryAnderson
 
Posts: 86
Joined: Sat Apr 15, 2006 10:09 am
Location: Norfolk

Postby rocksolid on Wed May 31, 2006 4:58 pm

I am completely satisfied with the dice. 2 armies statistically defend 3 times as well as a single army, which has nothing to do with the dice generator. Attacking 2 armies from a country with 3 - i.e. only two attacking armies - has a very low probability of success, again without having anything to do with the dice generator. Streaks happen both ways, and I'd be more suspicious if they didn't.
User avatar
Lieutenant rocksolid
 
Posts: 625
Joined: Sat Mar 18, 2006 10:00 pm
Location: Mowwwnt Reeeal

Postby wacicha on Wed May 31, 2006 5:52 pm

i love the dice it is all in the wrist and as areyouincahoots has been quoted as saying just learn when to quit pushing the attack button
Image
User avatar
Major wacicha
 
Posts: 3988
Joined: Fri Jan 27, 2006 6:51 pm

Postby tonywalrus on Mon Jun 05, 2006 2:29 pm

The results are random and are generated from random.org. It is worth having a look at their site, if you want to know how the numbers are created. They have gone to a lot of trouble to generate true random numbers and I for one am pretty happy with them.

Even if the dice do piss me off!
User avatar
Sergeant tonywalrus
 
Posts: 51
Joined: Wed Apr 05, 2006 7:05 am
Location: Deep below the cold inky darkness of the Bering Sea.

Suggestion for random number generation

Postby Lizard_Pancake on Mon Jun 26, 2006 4:44 pm

I found this on MathWorld:

http://mathworld.wolfram.com/Rule30.html

Quoted from the site:

Rule 30 is of special interest because it is chaotic (Wolfram 2002, p. 871), with central column given by 1, 1, 0, 1, 1, 1, 0, 0, 1, 1, 0, 0, 0, 1, ... (Sloane's A051023). In fact, this rule is used as the random number generator used for large integers in Mathematica (Wolfram 2002, p. 317).


Perhaps the random numbers used on Conquer Club could be generated with this simple method. Here is what I'm thinking:

- Seed the initial conditions of rule 30 with random numbers taken off random.org
- Run the cellular automaton for as many iterations as needed in order to fill up a buffer on CC's server. The central column will serve as a bit-by-bit source of randomness.
- Read from the buffer when necessary, periodically re-generating the numbers.

If the site's programmers want, I'd be willing to write the generator myself in C (which could be used as a CGI script, right?).
Private Lizard_Pancake
 
Posts: 1
Joined: Wed Jun 21, 2006 2:08 pm
Location: Montreal, Qc.

Make dice rolls independent

Postby Ticma on Tue Aug 01, 2006 5:19 am

Q: How do the dice work?
A: The dice 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 dice are rolled, 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 dice rolls and is re-loaded when all the lines are used up. As of July 2006 we consume 68,500 lines of dice rolls per day.


By discarding lines from file dice rolls become dependent. Current roll also dependents how players rolled before that roll. While file becomes smaller the influence of that becomes bigger.

One extreme situation that describes it is this:
Two last lines could be attacker loose 2 and then that player,who rolls in that moment,is loosing that roll for sure(100% chance to loose, 0% to win) or if you lucky it could left to lines deffender loose 2 and then you win for sure. In both cases you dont have expected chances(3 vs. 2:
37-33-30).
User avatar
Sergeant Ticma
 
Posts: 13
Joined: Thu May 18, 2006 6:12 am

Postby mightyal on Tue Aug 01, 2006 5:54 am

I've read this repeatedly and don't understand. Any line could have attacker losing 2 and whoever gets that line whilst attacking will do poorly. You only have a 30% chance of any line resulting in the attacker losing 2 regardless of where you are in the file.
I don't understand your problem with discarding lines after using them. It has to be better than recycling them surely.
"I do not feel obliged to believe that the same God who has endowed us with sense, reason, and intellect has intended us to forgo their use."
- Galileo Galilei
User avatar
Captain mightyal
 
Posts: 280
Joined: Fri Jul 14, 2006 12:33 pm
Location: Banging the hag whilst Owl is busy banging hendy's mum

Postby Bonquida on Wed Oct 11, 2006 8:28 am

Just because the name of the website is random.org doesn't mean the dice are actually random. All I know is that if the dice in CC are random, then the ones in real life must not be; seeing as how dice in cc work nothing like dice in real life.
User avatar
Corporal 1st Class Bonquida
 
Posts: 6
Joined: Thu Aug 24, 2006 5:52 pm

Postby Bonquida on Wed Oct 11, 2006 8:44 am

43% of people are not satisfied with the dice. This is less than half. It is also a lot of people! People who are not satisfied with a basic, important element of the gameplay. That would be like 43% of people saying the brakes on their car don't work well. Would car manufacturers still produce the same brakes if 43% of their customers had a problem with it? Probably, but SHOULD THEY? nah, not really.

Anyway, it might just be human psychology that is ticking me off, so I just started to tally all of my wins vs. loss attacking. I should win more often attacking than the defender in 3 v. 2 situations and I'm going to see if that is happening.
User avatar
Corporal 1st Class Bonquida
 
Posts: 6
Joined: Thu Aug 24, 2006 5:52 pm

Postby hitandrun on Wed Oct 11, 2006 8:45 am

Bonquida wrote:Just because the name of the website is random.org doesn't mean the dice are actually random. All I know is that if the dice in CC are random, then the ones in real life must not be; seeing as how dice in cc work nothing like dice in real life.


If you actually read random.org you will see the dice are as random as we're going to get on a computer.
Maybe we should employ someone to sit with actual dice and roll for us, entering the thousands of results by hand!! :P
User avatar
Corporal hitandrun
 
Posts: 143
Joined: Mon Sep 04, 2006 6:00 pm

Postby hitandrun on Wed Oct 11, 2006 9:00 am

Bonquida wrote:I should win more often attacking than the defender in 3 v. 2 situations and I'm going to see if that is happening.


rocksolid wrote: Attacking 2 armies from a country with 3 - i.e. only two attacking armies - has a very low probability of success, again without having anything to do with the dice generator.


Take heed.
User avatar
Corporal hitandrun
 
Posts: 143
Joined: Mon Sep 04, 2006 6:00 pm

Postby sully800 on Wed Oct 11, 2006 12:56 pm

Bonquida wrote:43% of people are not satisfied with the dice. This is less than half. It is also a lot of people! People who are not satisfied with a basic, important element of the gameplay. That would be like 43% of people saying the brakes on their car don't work well. Would car manufacturers still produce the same brakes if 43% of their customers had a problem with it? Probably, but SHOULD THEY? nah, not really.

Anyway, it might just be human psychology that is ticking me off, so I just started to tally all of my wins vs. loss attacking. I should win more often attacking than the defender in 3 v. 2 situations and I'm going to see if that is happening.


Actually it would be like 43% of people saying "You know, I'm not sure if my brakes work properly or not. I'll complain about them a bunch but never back up my complaints with any data to show how they are not working".

The people who have tried to show whether the dice are random or not (most notably though the Dice Analyzer add on) have seen that as the number of rolls increases, the results reflect what probability says should happen. Check out AK_Iceman's data from when he attacked with some 15,000 men at once against a force of similar size. I believe after the thousands of dice rolls the results were a mere 30 or so off what "should" have happened. That has made me satisfied with the long run results of the dice.

Yes its a shame that they can determine the outcome of the game, but those are the risks you take.
User avatar
Major sully800
 
Posts: 4978
Joined: Wed Jun 14, 2006 5:45 pm
Location: Bethlehem, Pennsylvania

Postby Ishiro on Thu Oct 12, 2006 5:39 am

Real dice are not truly random. The reason you do better on your own Risk board at home is because you become used to the weight of the dice and shift your throwing pattern. A little more slide. A little more drop. Studies have shown that people will, over time, get a better than even result using the same set of dice because of forming new habits, even when they are unconscious of the "cheating".

Yeah, that's bullshit, but it sounds true... and people want it to be true. Have you ever seen the gyrations people go through with real dice trying to squeeze luck out of them? I have a friend who will only roll in the box top with a sidearm throw and the dice must bounce off two sides, not one, not three... he has to toss them toward a corner and have them bounce two sides and out. Otherwise he blames the number of collisions for his bad luck.
User avatar
Corporal 1st Class Ishiro
 
Posts: 324
Joined: Mon Jul 17, 2006 5:53 pm
Location: Under the Rainbow

Postby cyberdaniel on Thu Oct 12, 2006 9:46 am

Ok, again the the dice are random enough. Go here and you can see:
http://www.conquerclub.com/forum/viewtopic.php?p=101555#101555

And if we're going to do that table again with the current data we'll see that the dice are random.

What people are warried about might be because of the luck streaks. For a non-premium you can see that these streaks change every 2 games or so. Most times I observe that 2 games have "good" dice while after those I have 2 games that I get the worst dice. If you're a premmium its a bit better since in the 20 games you have, you'll see bad dice, ok dice and really good dice, even thow you still have those streaks.

I think (I don't know for sure, I might be totally wrong) that the algorithm used to get the set of data might have a bug or something (don't know). I think that this has been mentioned before in a post but I couldn't found it.

If indeed this is the problem then this might be corrected by using numbers from random.org for the dice chooser.(for the dice these numbers are ok)
"Every battle has been won before it's been fought." Sun Tzu
User avatar
Captain cyberdaniel
 
Posts: 281
Joined: Sat Apr 01, 2006 10:31 am

Postby joeyjordison on Sun Oct 15, 2006 11:24 am

lol. this has been goin on forever. they even hav a special section explainin how the dice aren't wrong people just think they r :)
User avatar
Major joeyjordison
 
Posts: 1170
Joined: Wed Apr 19, 2006 9:10 am

Postby Scorba on Sun Oct 15, 2006 12:12 pm

No, the averages may be correct but the streakiness is a huge problem.
Taking an enemy on the battlefield is like a hawk taking a bird. Though it enters into the midst of a thousand of them, it pays no attention to any bird other than the one it has first marked.
User avatar
Lieutenant Scorba
 
Posts: 396
Joined: Fri Mar 24, 2006 6:14 am
Location: Upon a pale horse

Postby iLLa on Sun Oct 15, 2006 12:45 pm

its all probability. i like it. i just wish the color around it was dif because its harder to see the yellow on top of the white. other than that i like it.
User avatar
New Recruit iLLa
 
Posts: 7
Joined: Thu Oct 12, 2006 3:14 pm
Location: Carrollton, TX

PreviousNext

Return to Archived Suggestions

Who is online

Users browsing this forum: No registered users