Fair Dice

[drummer boy]

are the dice fair

[MeDeFe]

Where's the "they're random" option?
I will not vote until I get that one.

[wicked]

No, they're random. Random dice are like life, not always fair.

[misterman10]

No, they're random. Random dice are like life, not always fair.

of course they are fair, they are random

[sharrakor]

One could argue that you have a 50% chance of getting a good roll (4s and above), and a 50% chance of getting a bad roll (3s and below).

Sounds fair to me.

[Rocketry]

No, they're random. Random dice are like life, not always fair.

what????????

the dice are random, yes, and they are random for everyone AND THEREFORE FAIR!!!!!!!!!!!!

Rocketry

[homer_jay]

One could argue that you have a 50% chance of getting a good roll (4s and above), and a 50% chance of getting a bad roll (3s and below).

Sounds fair to me.

you're forgetting the situations where you get both high and low rolls...

the chances of all high rolls or all low rolls are equal, but not 50% each...

[bob72]

yeah but a draw means win for defender what are the odds in a 3-2 match?

Some people think that 3v2 means automatic win but even if you roll 2*6 you could lose 2 where as if defender rolls 2*6 he's won.

The odds are not really in favour of the attacker even with the extra dice are they?

[Pippin]

yeah but a draw means win for defender what are the odds in a 3-2 match?

Some people think that 3v2 means automatic win but even if you roll 2*6 you could lose 2 where as if defender rolls 2*6 he's won.

The odds are not really in favour of the attacker even with the extra dice are they?

Yes, they arn't because it's one of the realistic aspects of risk, the attacker is not as well prepared as defenders of have their own 'base' so as to speak

[snufkin]

I agree that random isn´t fair..

Go play chess if you don´t want fate or luck to be involved.

[gpellamjr]

The attacker does have the advantage in a 3-2 roll. So, a 10-10 attack still, I think, goes in favor of the attacker. But a 30-30 attack goes even more in favor of the attacker.

Greg

[Tashunca-uitco]

I think that the dice are equally fair or unfair to all.

However, I recently threw three ones in a row - and then 'threw' the same again the next roll - the chances of this are 46,656 to 1.

In a recent game, the defender, with one army, threw six 'sixes' in a row - a similar chance against this happening.

Perhaps we should ask how the dice rolls are calculated - please shoot me down in flames if I am talking b******s - but the last time I wrote programs for computers (many years ago when they were made of stone and powered by blood sacrifices) the one thing you can't ask a machine to do is to provide you with 'any number between 1 and 6' - or to 'take a card, any card'. You may as well ask it 'What is your favourite colour?'

The 'Basic' command RND(x) involved a complex method of producing what was to all intents and purposes a random number - but was in fact a the product of a long winded series of calculations.

There is no advantage for the site to operate a 'fixed' roll and, if the rolls work for and against all players equally, then there is no point in crying 'not fair' but, when players are assessing the odds accurately in order to decide tactics, then the dice rolls should be representative of the laws of probability.

Hey what do I know? I am 47 and cannot program my video and insist on 'tonight' in SMS texts instead of 'tonite'.

I should probably just shut up and take up bingo (lotto US), drafts (checkers US) or compile a dictionary of words different on either side of the Atlantic. (Do you know what 'Fanny' means in England? LOL)

[RiskTycoon]

go buy the board game, lose 8 in a row rolling 3 dice against a single......then come back to CC and do the same thing! it's like magic !

[Tashunca-uitco]

I didn't say I didn't appreciate that there is an element of chance (and I have owned the board game for over 30 years) - I only wished to point out that a random number produced by script is entirely different to one produced by tipping the die from your own hand.

This opinion is based on info that is nearly as old as the length of time I have owned the board game - and I invited correction - or instruction on how the die results are produced.

Savvy?

[Bean_]

I think that the dice are equally fair or unfair to all.

However, I recently threw three ones in a row - and then 'threw' the same again the next roll - the chances of this are 46,656 to 1.

In a recent game, the defender, with one army, threw six 'sixes' in a row - a similar chance against this happening.

Perhaps we should ask how the dice rolls are calculated - please shoot me down in flames if I am talking b******s - but the last time I wrote programs for computers (many years ago when they were made of stone and powered by blood sacrifices) the one thing you can't ask a machine to do is to provide you with 'any number between 1 and 6' - or to 'take a card, any card'. You may as well ask it 'What is your favourite colour?'

The 'Basic' command RND(x) involved a complex method of producing what was to all intents and purposes a random number - but was in fact a the product of a long winded series of calculations.

There is no advantage for the site to operate a 'fixed' roll and, if the rolls work for and against all players equally, then there is no point in crying 'not fair' but, when players are assessing the odds accurately in order to decide tactics, then the dice rolls should be representative of the laws of probability.

Hey what do I know? I am 47 and cannot program my video and insist on 'tonight' in SMS texts instead of 'tonite'.

I should probably just shut up and take up bingo (lotto US), drafts (checkers US) or compile a dictionary of words different on either side of the Atlantic. (Do you know what 'Fanny' means in England? LOL)

The dice are taken from random.org, which is supposed to be a true (not pseudo) random number generator taking events from physical phenomena. I have been trying to get to the bottom of this randomness issue; it appears that the overall incidences of 1s, 2s, 3s, 4s, 5s and 6s mirror the distributions that they should. No one has done a systematic test of streakiness, although DiM has a thread reporting his daily dice experiences, and so far those results do not indicate abnormal streakiness.

We have anecdotal reports like yours, but people are likely to post in these threads only if they have seen unusual events, so it is difficult to conclude much from them. A streak of 6 attacker 1s would be expected about 20 times per day on this site (we have been told that the site goes through about 1 million "lines" of dice per day, with each line being 3 attacker and 2 defender dice). Obviously, the chance that you individually see 6 1s and then shortly thereafter, 6 6s is fairly low.

[Tashunca-uitco]

Thank you for taking the time and trouble to explain.

Of course, one notes the unusual and one cannot rule out the effect of partisan-ship. English people constantly complain about referee's decisions 'robbing' us of important victories in the World and European cups - but conveniently forget our victory against Spain in 1996 when at least four crucial decisions went in our favour.

I also acknowledge that there is the same chance of producing any combination of numbers from two throws of three dice as there is of producing six 'ones'.

Once again, thank you for your time.

[lt_oddball]

these dices SUCK !!

Just because some "org" "CLAIMS" that their dices are "more" realistic (than what ?) does not PROVE it is !
You really believe anything sellers tell you ?
"our dices are the best randomized, please buy us now" . "Our cars are the cleanest and yet fastest engines on the world , please buy us now "
"We have the lowest mortgage interets rates, please take us !". " We guarantee you the biggest penis enlargement for the cheapest price, please buy from us!".


From my experience with the sets of 3 attack dices versus 2 defenders on this conquerclub it is CONISTENTLY worse than what STATISTICALLY SHOULD be the results !

I posted it before that any lengthy roll series (I find 20 rolls or more statistically quite representative) SHOULD give slightly more wins for the attacker than the defender.
And ALL TOO OFTEN this is NOT the case.

[lt_oddball]

quotes:

I put all the combinations of the attacker's 3 dices (56) against the defenders 2 dices (21).

and not to my surprise I find that of the 1176 (56x21) possibilities
420 give the attacker 2 wins,
378 give the attacker 1 win and
378 give the attacker 0 win.

So at any lenghty attack series the attacker SHOULD kill more than he 'd loose !!!



another one:

Example of 100 numbers between 1 and 6 from "random.org":
(read from left to right).
It just shows there are way too many freak series :

True Random Number ServiceRandom Integer Generator
Here are your random numbers:

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


56656256 ???
and 111212412 ???
and 4124545434 ???

Go try explain that with realistic dices.

[lt_oddball]

One might argue that at least the freak dice is/should be equally for all players and thus it is still "fair".

But actually it isn't;

The players that chose a attacking/agressive strategy -in the belief that their 3-2 attack dices SHOULD give them the edge- loose more than those that stay in a corner and build up.
Which means that overall, "attacking" is not rewarded as it should be (as in the REAL boardgame with REAL dices), and thus the "defensive" player has the advantage here.
And I always thought that the concept of Risk is to favour the attacker (as otherwise NO FUN happens in the boardgame).

So conquerclub (and more particularly Random.org) must invest some time /effort to get these dice results MUCH more in line with common statistics.

[Bean_]

quotes:

I put all the combinations of the attacker's 3 dices (56) against the defenders 2 dices (21).

and not to my surprise I find that of the 1176 (56x21) possibilities
420 give the attacker 2 wins,
378 give the attacker 1 win and
378 give the attacker 0 win.

So at any lenghty attack series the attacker SHOULD kill more than he 'd loose !!!



another one:

Example of 100 numbers between 1 and 6 from "random.org":
(read from left to right).
It just shows there are way too many freak series :

True Random Number ServiceRandom Integer Generator
Here are your random numbers:

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


56656256 ???
and 111212412 ???
and 4124545434 ???

Go try explain that with realistic dices.

This is very interesting. The experiment should be set up in advance with exactly what you are looking for, rather than searching for patterns after generating the rolls. The chance for 5 1s in 10 rolls is 1.3%. But we should look line by line (to prevent overlap problems that make the math very hard).

There is a group of 5 1s in the third line, but there is a 12.3% chance of this occurring (i.e., 1 or more groups of exactly 5 1s) in a simple of 10 lines of 10 rolls. Maybe I will run the experiment on more numbers from random.org and see how it turns up.

[Bean_]

One might argue that at least the freak dice is/should be equally for all players and thus it is still "fair".

But actually it isn't;

The players that chose a attacking/agressive strategy -in the belief that their 3-2 attack dices SHOULD give them the edge- loose more than those that stay in a corner and build up.
Which means that overall, "attacking" is not rewarded as it should be (as in the REAL boardgame with REAL dices), and thus the "defensive" player has the advantage here.
And I always thought that the concept of Risk is to favour the attacker (as otherwise NO FUN happens in the boardgame).

So conquerclub (and more particularly Random.org) must invest some time /effort to get these dice results MUCH more in line with common statistics.

The dice analyzer stats that have been reported pretty much show the battle outcomes in line with what is expected.

[misterman10]

Heres my take:

EVERYBODY STOP COMPLAINING ABOUT THE DICE. I'm sick of hearing, "well the dice should have", or "in the real boardgame risk", or "the chances of that were so and so", or "the dice are unfair"

Get over it. It's ridiculous how much you people complain about the dice, and THIS IS NOT RISK

Copyright © 2006 by Salamander Software.
RISK is a registered trademark of Hasbro Inc. Conquer Club is not associated with RISK or Hasbro in any way.

anyways, thats my take on the dice

[Bean_]

quotes:

I put all the combinations of the attacker's 3 dices (56) against the defenders 2 dices (21).

and not to my surprise I find that of the 1176 (56x21) possibilities
420 give the attacker 2 wins,
378 give the attacker 1 win and
378 give the attacker 0 win.

So at any lenghty attack series the attacker SHOULD kill more than he 'd loose !!!



another one:

Example of 100 numbers between 1 and 6 from "random.org":
(read from left to right).
It just shows there are way too many freak series :

True Random Number ServiceRandom Integer Generator
Here are your random numbers:

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


56656256 ???
and 111212412 ???
and 4124545434 ???

Go try explain that with realistic dices.

This is very interesting. The experiment should be set up in advance with exactly what you are looking for, rather than searching for patterns after generating the rolls. The chance for 5 1s in 10 rolls is 1.3%. But we should look line by line (to prevent overlap problems that make the math very hard).

There is a group of 5 1s in the third line, but there is a 12.3% chance of this occurring (i.e., 1 or more groups of exactly 5 1s) in a simple of 10 lines of 10 rolls. Maybe I will run the experiment on more numbers from random.org and see how it turns up.

I've just taken 20,000 dice rolls from random.org looking for exactly 5 of the same number in each row of 10. I would expect 26, and here is how many I got:

1s: 23
2s: 28
3s: 18
4s: 34
5s: 33
6s: 27

This does not look wildly off to me. The 33 and 34 are a little high, but still have a 5-6% chance of happening naturally. None of these are 50 or anything crazy like that.

UPDATE:

I've run off 50,000 dice from random.org and can find no unusual clumps. The following table shows the number of times a die roll appeared in a group of 10 (e.g., 271 times that 4 1s showed up in a group of 10). The numbers below appear normal.

Count: 1 2 3 4 5 6
Length:
4 271 261 258 268 267 284
5 68 66 49 78 75 60
6 13 13 16 11 12 19
7 1 4 2 0 0 1
8 0 0 0 0 0 1
9 0 0 0 0 0 0
10 0 0 0 0 0 0

You would expect:
271, 65, 11, 1.24, and 0.09 for lengths 4,5,6,7 and 8.

Looks pretty good to me.

[Minister Masket]

The attacker does have the advantage in a 3-2 roll

Greg

No he doesnt. A 3-2 roll means both sides use two dice, which the defender will win if the result is a draw. So actually, the defender has the advantage.

[Bob Janova]

Depends what a '3-2 roll' means. Rolling 3 dice versus 2 does give the attacker an advantage; attacking a 2 territory from a 3 doesn't, because the attacker only gets 2 dice attacking from a 3.

I put all the combinations of the attacker's 3 dices (56) against the defenders 2 dices (21)

There are 36 for 2 dice and 216 for 3. '1 6' and '6 1' both need to be included in the calculation (a '6-1' is twice as likely as a '6-6'), and likewise for the 3 dice case.