Conquer Club

As the dice are not random, could you add a button into the game: "Average dice", please. Using that you'd loose only the average amount troops that the theory says. Of course, for the small amount of defenders it's not relevant but for bigger that 6 or 7 defenders I'd always use it!

Great idea!

I would also like to see an "opponents never take more than 6 hours to take their turns" option.

And a script that changes the word dice into Rabbits in all the forums

Cachorro wrote:As the Rabbits are not random, could you add a button into the game: "Average Rabbits", please. Using that you'd loose only the average amount troops that the theory says. Of course, for the small amount of defenders it's not relevant but for bigger that 6 or 7 defenders I'd always use it!

What are average dice?

And how does one go about calculating them?


It probably is possible... make the rolls, and then eliminate any rolls that are statistically 10% or lower likelyhood to happen, and move on until something in the 90% range happens. Its still random, its just not fully random, and only 10% of rolls never make it to the board.

10% is rather arbitrary...it might be too high or too low.

its also irrelevant...I cant see it ever happening. however, you would never really complain that you didnt roll those unlikely events, so removing those hail mary passes, probably would not generate too many complaints...and removing the 5% chance rolls, will certainly even out the game, except it would place much more emphasis on drop and going first on many types.

It could be done too actually. Calculating the odds as you went is insanely impossible as the server would croak, but the rolls could be generated randomly ahead of time, and any statistically insane ones simply deleted, leaving only relatively random rolls, with no chance of 10 losing to a 1. You wouldnt ever miss those rolls either, because you simply wouldnt expect them to happen in the first place.

No one is going to complain that a 10 took out their 1. The question would be, at what percent do you limit the possibilities.

10% is way to high after considering this. I think maybe at 5%, its actually feasible. The reason we see so many crazy rolls in the first place, is because we throw so many dice, so over time, an individual game has the chance of crazy dice, that if you only sat down to one game ever, you would probably not ever see. Removing the worst of the worst would simply remove the rolls that wouldnt typically happen in any one game.

This would fundamentally change the game though. If you have a 95% chance of winning an attack, you would now know you have 100% chance of winning the attack, and in reality, there was a 5% chance of losing, which really would happen from time to time. That is at 5% which may still be way too high. But I certainly wouldnt complain if say the 2% or 1% rolls never happened. Theoretically, they would only affect 1-2% of the rolls anyways and would preclude the ones that defy logic, however faulty logic, that the roll was impossible.

Unfortunately, it would dramatically change the game as I said, and you would know when you were guaranteed to win a roll, which is not what the game is about.

AAFitz wrote:What are average dice?

And how does one go about calculating them?


It probably is possible... make the rolls, and then eliminate any rolls that are statistically 10% or lower likelyhood to happen, and move on until something in the 90% range happens. Its still random, its just not fully random, and only 10% of rolls never make it to the board.

Unfortunately, for any single roll, there is no outcome that is statistically unlikely. In a 3v2 situation, there are 3 outcomes: Attacker loses 2, Defender loses 2, or each lose 1. All three cases have a probability of close to 1/3. In a 3v1 roll, there are 2 outcomes: Defender loses 1 (odds: 66%) or Attacker loses 1 (odds: 34%). The odds that an attacker loses a 10v1 is 0.01%. However, the odds that an attacker wins 10v1 in the first two rolls on 10 different occasions is also only 30%. The odds of losing a single army when you attack 10v1 is 22%. Losing zero armies has a 66% chance. How many armies should the attacker lose in a diceless version? If you say zero, and this happens 10 times in a game, you've selected an outcome that should only happen 1.5% of the time. If you say 1, and this situation occurs 10 times in the game, you've assigned an outcome that should only have 0.000032% of the time.

You can't just eliminate single rolls. You could perhaps eliminate some results from an auto-attack. And perhaps that's what the OP meant by "average dice." It's easy enough to calculate the expectation value for the results of any given battle. Basically this suggestion is asking for a diceless option on a round-by-round basis. However, he is wrong is suggesting that

the dice are not random

In order for the dice to be truly random, they are unpredictable by definition. It is also a misconception to believe that random numbers are reasonably uniformly distributed. They aren't. There is an old experiment that I read about when I was teaching mathematics where the teacher assigned half the class to flip a coin 100 times and write out the results and the other half of the class to make up a set of random coin flip results and write them out. The teacher had a pretty easy time picking out which were the real coin flips and which were the imagined ones, because the made-up versions had very little streakiness. If you flip a real coin 100 times, it's to be expected for it to come up heads 10 (or more) times in a row every so often.

To be honest, I'm not sure I'd be particularly interested in a diceless version of the game. Part of the allure is taking chances and knowing that some battles will turn out poorly, but that sometimes you'll get lucky and will overcome the odds. If you take out the dice, you need another way to inject more variability into the game. Games like chess have a larger variety of pieces with different capabilities (imagine chess with only kings!). Other games, like diplomacy, rely on secret negotiations, lies, and alliances to yield a more random game.

Lets be honest about a diceless option. If you know the odds are against you, you would never choose to have the battle automatically give expected results since you hope for a bit of luck to give you a victory. Conversely, if the odds are in your favor, you'd be far more likely to pick the diceless option to prevent some random rolls from going against you. Ultimately this shifts the balance of power strongly in favor of the one that decides how the battle outcome is determined (presumably the attacker).
As an example, suppose that you have a set of battles. 50% of them have an average probability of only 25% chance of success. The other 50% have a 75% chance of success. Using random dice, you should expect to win about half of your battles. If you choose the diceless option for the 75% chance battles and the random option for the others, you increase your win rate to 63% of the battles. Basically you're shifting the odds strongly in favor of the attacker, which will make for an unbalanced game (I suspect it will strongly favor the person that goes first). Notice from the 10v1 example above that assigning a specific result may be reasonable for a single battle, but it produces an extremely unlikely result (that is very favorable to the attacker) over time.

The point of all this is that you can't really assign a "likely result" on single battles without producing very unlikely results for a large set of battles. If instead you want to get a large set of battles to fit to some sort of distribution, guess what? All you need is a random number generator, which is already in place! If you doubt the overall distribution of the numbers, install the dice analyzer script and you'll see that it's just fine.

Random
Function: noun

: a haphazard course
–at random : without definite aim, direction, rule, or method

Random
Function: adjective

: lacking a definite plan, purpose, or pattern

What about a 2nd button with the "order a pizza option", it could be handy when having multiple speed games ongoing.

Doc_Brown wrote:

AAFitz wrote:What are average dice?

And how does one go about calculating them?


It probably is possible... make the rolls, and then eliminate any rolls that are statistically 10% or lower likelyhood to happen, and move on until something in the 90% range happens. Its still random, its just not fully random, and only 10% of rolls never make it to the board.

Unfortunately, for any single roll, there is no outcome that is statistically unlikely. In a 3v2 situation, there are 3 outcomes: Attacker loses 2, Defender loses 2, or each lose 1. All three cases have a probability of close to 1/3. In a 3v1 roll, there are 2 outcomes: Defender loses 1 (odds: 66%) or Attacker loses 1 (odds: 34%). The odds that an attacker loses a 10v1 is 0.01%. However, the odds that an attacker wins 10v1 in the first two rolls on 10 different occasions is also only 30%. The odds of losing a single army when you attack 10v1 is 22%. Losing zero armies has a 66% chance. How many armies should the attacker lose in a diceless version? If you say zero, and this happens 10 times in a game, you've selected an outcome that should only happen 1.5% of the time. If you say 1, and this situation occurs 10 times in the game, you've assigned an outcome that should only have 0.000032% of the time.

You can't just eliminate single rolls. You could perhaps eliminate some results from an auto-attack. And perhaps that's what the OP meant by "average dice." It's easy enough to calculate the expectation value for the results of any given battle. Basically this suggestion is asking for a diceless option on a round-by-round basis. However, he is wrong is suggesting that

the dice are not random

In order for the dice to be truly random, they are unpredictable by definition. It is also a misconception to believe that random numbers are reasonably uniformly distributed. They aren't. There is an old experiment that I read about when I was teaching mathematics where the teacher assigned half the class to flip a coin 100 times and write out the results and the other half of the class to make up a set of random coin flip results and write them out. The teacher had a pretty easy time picking out which were the real coin flips and which were the imagined ones, because the made-up versions had very little streakiness. If you flip a real coin 100 times, it's to be expected for it to come up heads 10 (or more) times in a row every so often.

To be honest, I'm not sure I'd be particularly interested in a diceless version of the game. Part of the allure is taking chances and knowing that some battles will turn out poorly, but that sometimes you'll get lucky and will overcome the odds. If you take out the dice, you need another way to inject more variability into the game. Games like chess have a larger variety of pieces with different capabilities (imagine chess with only kings!). Other games, like diplomacy, rely on secret negotiations, lies, and alliances to yield a more random game.

Lets be honest about a diceless option. If you know the odds are against you, you would never choose to have the battle automatically give expected results since you hope for a bit of luck to give you a victory. Conversely, if the odds are in your favor, you'd be far more likely to pick the diceless option to prevent some random rolls from going against you. Ultimately this shifts the balance of power strongly in favor of the one that decides how the battle outcome is determined (presumably the attacker).
As an example, suppose that you have a set of battles. 50% of them have an average probability of only 25% chance of success. The other 50% have a 75% chance of success. Using random dice, you should expect to win about half of your battles. If you choose the diceless option for the 75% chance battles and the random option for the others, you increase your win rate to 63% of the battles. Basically you're shifting the odds strongly in favor of the attacker, which will make for an unbalanced game (I suspect it will strongly favor the person that goes first). Notice from the 10v1 example above that assigning a specific result may be reasonable for a single battle, but it produces an extremely unlikely result (that is very favorable to the attacker) over time.

The point of all this is that you can't really assign a "likely result" on single battles without producing very unlikely results for a large set of battles. If instead you want to get a large set of battles to fit to some sort of distribution, guess what? All you need is a random number generator, which is already in place! If you doubt the overall distribution of the numbers, install the dice analyzer script and you'll see that it's just fine.

The point isnt that they arent random. The point is, that because so many games are played, and dice thrown, often by one player...they typically see rolls that they would not expect to see playing one game. I understand that technically they might, but statistically speaking, if you attack a 1 with a 10, you will win, but if you play 8000 games, and throw millions of dice, you will absolutely see 10 lose to 1, even though its a very, very rare event, technically.

If the really low probability rolls were eliminated, the dice would not be random, however, they would still be fair, and arguably more so.

Because of the way the dice work, and because so many people are playing, while the dice are random, because so many are thrown, by so many people, and so often, we do see rolls that if you played just one game of risk, you would not expect to see.

I know that if I go into one game, my 6 will kill a 1. However, I also know that in my next few hundred games, my 6 will probably lose to a 1. Sure, it could be the first game, but it is highly unlikely. Just as possible, but still, unlikely.

So eliminating the extremely low probability rolls, the dice would be not random, but, would be possibly more fun, generate less complaints, and perhaps less pure...but upon reflection, would be an interesting thing to test out. The only problem is that on some rolls, there would be no uncertainty. If you eliminate rolls below a certain probability, you will know how a roll will turn out.

Its at least worthy of a discussion in any case. I actually just dont see any downside.

fumandomuerte wrote:What about a 2nd button with the "order a pizza option", it could be handy when having multiple speed games ongoing.

just copy this into your button bar on top http://www.dominospizza.com

AAFitz wrote:

fumandomuerte wrote:What about a 2nd button with the "order a pizza option", it could be handy when having multiple speed games ongoing.

just copy this into your button bar on top http://www.dominospizza.com

True, forgot the bookmarks thingy

If we are going to add an average dice button can we go ahead and add a "perfect dice" button. I would use that one a lot! Then, for the foolish, can we add a "completely terrible" button? This has potential!

AAFitz wrote:The point isnt that they arent random. The point is, that because so many games are played, and dice thrown, often by one player...they typically see rolls that they would not expect to see playing one game. I understand that technically they might, but statistically speaking, if you attack a 1 with a 10, you will win, but if you play 8000 games, and throw millions of dice, you will absolutely see 10 lose to 1, even though its a very, very rare event, technically.

If the really low probability rolls were eliminated, the dice would not be random, however, they would still be fair, and arguably more so.

Because of the way the dice work, and because so many people are playing, while the dice are random, because so many are thrown, by so many people, and so often, we do see rolls that if you played just one game of risk, you would not expect to see.

I know that if I go into one game, my 6 will kill a 1. However, I also know that in my next few hundred games, my 6 will probably lose to a 1. Sure, it could be the first game, but it is highly unlikely. Just as possible, but still, unlikely.

So eliminating the extremely low probability rolls, the dice would be not random, but, would be possibly more fun, generate less complaints, and perhaps less pure...but upon reflection, would be an interesting thing to test out. The only problem is that on some rolls, there would be no uncertainty. If you eliminate rolls below a certain probability, you will know how a roll will turn out.

Its at least worthy of a discussion in any case. I actually just dont see any downside.

Actually, I think you're missing my main point. There is no such thing as a "low probability roll." All possible outcomes from a single roll have equal probability. It is possible to talk about a low probability sequence of rolls. In that case, presumably the attacker would never lose a battle in which he has greater than, say 75% (or 90% or 98% or whatever) odds of success. But then you have to define what an "unlikely sequence" constitutes. A single 5v1 battle has over 97% of success. The odds of winning 10 of those (which could conceivably occur within a single turn on some of the larger maps) is less than 75%. Or set the bar a bit lower. A 7v3 has over 85% chance of success. But winning 5 of those has odds of 46%.

So simply by insisting that statistically unlikely sequences are not allowed to occur, you create a statistically unlikely result! So fixing rolls will not improve the randomness of the game. It will remove randomness from it. I think it's a bad idea to give the attacker the option to decide whether to roll or go diceless on an attack-by-attack basis. As I noted earlier, it essentially changes the odds in favor of the attacker. A diceless game would be fine. Or you could create a new distribution approach to the calculations for battle results. e.g. You could use a Gaussian distribution centered at the expectation value for the given battle (as evaluated the normal way) and with a fairly small variance (which would become smaller as the number of attacking armies increased). Either way, these options should be in place for the entirety of the game.

From looking at the suggestions forum, this subject seems to come up constantly. However you want to put it, I see fantastic results REPEATEDLY in most of my games and by fantastic I mean i consistantly raped by the dice. Its gotten to the point that I EXPECT to lose if i dont have a 3 to 1 ratio and even then it happens.

I know what people have said...so many dice..so many games blah blah blah fantastic things will happen. The point is, that my games are consistantly statistically impossible. Its like im winning the lottery of suck in every game. Whats funny is that the dice NEVER go the other way and Ive tried it. I said "okay" my 13 seem to consistantly get molested by 2 or 3 troops so let me attack with that many. Hasnt worked for me yet and Ive lost a few games because of my little experiment (Yes....yes I know, a stupid experiment but I tried it anyway)

Lets do the math....

"random dice" + repeated statistic improbabilities in my games = somethings not right.

I bet my little 'equation' fits the bill for a lot of people too.

Maybe its a faux pas to say this in this forum but I have a freebie account on a competing website and I do not have these problems.

Have you tried installing the dice analyzer plugin? http://www.conquerclub.com/forum/viewtopic.php?f=59&t=5655
That will keep track of all your dice rolls and can tell you exactly how lucky or unlucky you actually are. If you use that for a couple hundred rolls and your results are deviating significantly from expected results, then you really have a valid complaint.

@ Doc - Thanks...i will do so. It will be fun to track my rolls.

I think my 'complaint' echoes that of many others though. I've played risk all my life and have messed around other websites (this is by far the best one) and I hold fast to my statement that something is off. Although I agree with Doc Brown in that "average dice" is not a solution I think that Cachorro has hit the nail on the head as far as identifying the problem.

I think everybody is fixated on randomness when thats not really the problem. The problem, in fact, is that the conditions of the board game are not being replicated specifically with regards to the amount of dice rolls. CC has lets just say 100k die rolls a day while through a lifetime of board games you might have 10k dice rolls. If CC is cycling through die rolls or more accuratley probabilities at that much faster of a rate then it will NEVER be accurate because you will enocunter those 'impossible' probabilities.

I would like to make a point to seperate accuracy and statisitical probability. They are not the same which is why i can say that randomness is not the problem and that the website is not accurate to the board game. Basically I can play the BG all my life and never encounter the rolls that I have encountered here repeatedly.

The solution to this problem should be in matching the conditions of the board game to the site and not in worrying about randomness (which I believe has been achieved). Can it be done...I dont know. If this cant be done then the results will always be off. My field is nowhere near anything that would allow a technical exploration of my proposition so have fun shooting this down or exploring.


'nuff said (on my part at least)

Timur the Lame wrote:From looking at the suggestions forum, this subject seems to come up constantly. However you want to put it, I see fantastic results REPEATEDLY in most of my games and by fantastic I mean i consistantly raped by the dice. Its gotten to the point that I EXPECT to lose if i dont have a 3 to 1 ratio and even then it happens.

I know what people have said...so many dice..so many games blah blah blah fantastic things will happen. The point is, that my games are consistantly statistically impossible. Its like im winning the lottery of suck in every game. Whats funny is that the dice NEVER go the other way and Ive tried it. I said "okay" my 13 seem to consistantly get molested by 2 or 3 troops so let me attack with that many. Hasnt worked for me yet and Ive lost a few games because of my little experiment (Yes....yes I know, a stupid experiment but I tried it anyway)

Lets do the math....

"random dice" + repeated statistic improbabilities in my games = somethings not right.

I bet my little 'equation' fits the bill for a lot of people too.

Maybe its a faux pas to say this in this forum but I have a freebie account on a competing website and I do not have these problems.

I would just like to point out that every time you get "terrible dice" it means your opponent got "fantastic dice." By that logic alone, it all evens out. And I have used dice analyzer for a long time. I promise you, the attacker wins 2-0 approximately 38% of the time and loses 0-2 approximately 31% of the time. Just like it should be. Quit complaining, you have to have a back up plan for "what if the dice are bad" because there is a 31% chance they will be bad! Learn to stop rolling when they are going the wrong way and cut the bleeding. I get tired of reading threads that blame poor play on the dice....

Masli wrote:And a script that changes the word dice into Rabbits in all the forums

Cachorro wrote:As the Rabbits are not random, could you add a button into the game: "Average Rabbits", please. Using that you'd loose only the average amount troops that the theory says. Of course, for the small amount of defenders it's not relevant but for bigger that 6 or 7 defenders I'd always use it!

This is my favorite suggestion that's ever been mentioned on the forums.

Chuck, I would encourage you not to participate in anything you are tired of doing. The fact that you are tired of reading these indicates that the issue keeps popping back up which would perhaps suggest that people are more than just 'complaining.'

The real problem is that because you are 'tired' or reading these you do not take the time to actually read threads before commenting on them.

I would just like to point out that every time you get "terrible dice" it means your opponent got "fantastic dice." By that logic alone, it all evens out.

First off the purpose of my posts was not to complain. Second, your Cpt. Obvious logic here IS the point. Its not about me losing or me getting bad rolls its about 'fantastic rolls,' good OR bad, happening to an irregular degree (see my comment on accuracy and probability which I assume you just passed right over because it was 'long' and you were tired).

Also, the dice analyzer is flawed because the data collected is irrelevant to the point. Essentially, from what I've seen, it records in the form of a percentage amount of data and then displays the data in a limited format. The form it needs to take is more of a scatter plot which would highlight individual series or rolls. This way you could see how many 'statistical improbabilities' are occuring on both sides. Basically if 31% of the time I'm achieving lottery odds then maybe something is off. The point is that the things that Ive seen happen on this site and that other people write about constantly may only happen once in their life in a board game version of this. The question is, how accurate to the board game is this site trying to be? If its not then people just need to deal with the difference.

In the future Chuck please read the entire thread and understand what the posts are about before commenting otherwise you're just trolling.

Timur the Lame wrote:I think everybody is fixated on randomness when thats not really the problem. The problem, in fact, is that the conditions of the board game are not being replicated specifically with regards to the amount of dice rolls. CC has lets just say 100k die rolls a day while through a lifetime of board games you might have 10k dice rolls. If CC is cycling through die rolls or more accuratley probabilities at that much faster of a rate then it will NEVER be accurate because you will enocunter those 'impossible' probabilities.

I think this is a big part of the issue. If you're a fanatic Risk player (on the board game), I could see you playing maybe 2 games a week on average. So a sequence of rolls that has only 0.01% chance of happening might occur once a year for you. However, if you have 10 games going at a time here, plus you play a number of speed games, you could easily see something with a 0.01% chance occur 2-3 times in a single day!

Furthermore, I contend that there is a strong memory bias. People don't pay as much attention to their good rolls, but the bad rolls stand out. I've had plenty of bad luck as well. There was a period for a few days that I don't think I won a single battle regardless of odds. Luck changes though, and I've had plenty of good results mixed in. Keep in mind that a number of really good rolls is just as unlikely as a run of bad rolls. I just haven't seen anyone come in complaining about the dice being way too good!