The site's dice are accurately pseudorandom.

Matty im saying A) our dice are unfair and not like real dice

Ok, then I was right in what I thought you tried to accomplish. If you have read the first post in this topic than you would have noticed that it references to alot of statistics (like if you kill 10 troops, you will approximately lose 10 (well, 9.8 or something like that) troops), and has two remarks about them, being: 1) If you roll our dice LOTS AND LOTS of times, they coincide to these statistics 2) If you roll our dice a couple of times, really weird stuff happends (like killing 0 losing 10), and this is what happends to real dice as well. 3) [size=4](he, a 3rd?)[/size] Point two is hard to understand if you do not know what true randomness means, however, if you do understand randomness, you would know that weird things happen with real randomness 4) [size=4](stop it, you said it only named 2 points)[/size] The reason that some other sites do not have these weird things is because they have fake dice (which is not a bad thing, we also plan on implementing an option for fake dice). Now the problem is that in real life its a bit hard to roll 5 dice 1.000.000 times, and with a program its easier, but if you really want to roll 5.000.000 times, you will get similar results as our dice do :)

Matty I know what randomness is, however its not randomness if in several games it happens more then once 3,4,5,6 times or more, and it happens a lot more then once.

Of course these things happen - just as you had more than 6 times that you had exceptionally good dice. You just did not really notice the good ones (as they don't hurt) but do notice the bad ones (as they do hurt). That's what being a human being is, I sometimes need to check the full log too to convince me that it's my fault im losing :)

Matty Theres the thing i notice the good ones, i had won several games due to extraordinary dices, once again im not whinning i may even have won more games that i lost due to dice system. I just dont thing that statistically speaking those dices are not accurate, statistically you can do 3vs1 attacking and lose 10 troops and then defending 3vs 1 win 10 troops and you will say its 10 losses to each attacking and defending its good. I say its not statiscally possible The chances of that to happen lose 10 troops is 0.00189% the chances of it to happnes twice is 0,0000035721. But with your count you will just say win 10 lose 10 is normal. i think it would be interesting to know what is standard deviation in this dice system. The standard deviation is the one who tells us if dice system is reliable... IMO

the sad part about this is i bet that if we gave the attacker more of an advantage then people would complain how there defence keeps getting bad rolls.

I think im not explaining well The objective in my opinion is not to give more advantage to attack, just to make the standard deviation lower closer to a normal distribution... In the end its the same system to everyone, i just would like it to be more strategy then dices... The attack as the same issues, there are times that no troops is lost attacking...

The problem with this reasoning is that dice rolls are not normally distributed. The only way I know of making dice rolls into a normal distribution is to approximate it by adding, say, 1.000.000 rolls. But then we are talking of quite large amounts numbers of rolls, and not a couple of examples you have had.

Matty This should be a normal (or Gaussian) distribution Just read the following links http://en.wikipedia.org/wiki/Dice statisticas part read the definition in https://en.wikipedia.org/wiki/Normal_distribution

The thing satsousa is failing to understand is that randomness is clumpy and he doesn't see that his clumps of bad dice are normal and expected. http://telescoper.wordpress.com/2009/04/04/points-and-poisson-davril/

Oh nice, I was correct: "[b]As the number of dice increases[/b], the distribution of the sum of all numbers tends to normal distribution by the central limit theorem"

Matty do you even know what is a normal distribution? "As the number of dice increases, the distribution of the sum of all numbers tends to normal distribution by the central limit theorem" So as a normal distribution you shouldnt have in same game 3 vs 1 and defender winning 6,7,8,9 troops 2,3,4,5 times in same game. but it happens in a lot of games... if it happens only times to times then it would be normal... so if losing 6 troops vs 1 is normal (0,15%) how about happen twice (0.000225%) and 3 times (0,0000003375%) Vexer i dont notice only bad dices, i notice good dices too, whats the part you dont understand when i say i won many games due to it? I was trying to make postive criticism but nobody wants to understand its fine by me, lets play a luck game instead of strategy game I wont say a single word anymore, only losing my time and bothering you all...

Yes, I know what a normal distribution is, do you?. Single dice rolls are not normally distributed. For dice rolls you can only approximate a normal distribution, and you will need a fair amount of them before the approximation is close enough to make sense. I am going to try to be simple again: Go to this site: http://riskodds.com/ . Fill in a 10 vs 10 attack, the chance to win is about 50%. Now lets assume you lose the first roll (about 50-50), now fill in the 8 vs 10 attack - wow, suddenly you only have 28.6% chance of winning. Wow, what happened here, how can one simple roll have so much influence?? Well, the variance (for as far as it makes sense to talk about variance in such low rolls) is quite big, and really weird things can happen. Quite often actually ;) Try the same for winning the first roll (enter 10 vs 8), again, the percentage is far off that pretty 50%. Apart from that, the probability of losing 1 vs 6 is 0.9%, not 0.15% Please, if you try to convince us, give us a nice mathematical calculation of what the variance should be according to you, and then calculate what the variance is in this site (using 1000+ rolls please).

My last post Losing 6 in 3vs 1 means you atack 9 vs 1 and lost 6 ok? do the maths... if not it would be 3 vs 1 3 times then 2vs 1 then 1vs 1 the probability of defender wins 3 vs 1 is 33,71% so the probability of winning 6 times is 0.3371^6= according to my calculator it is 0,1467413516623572783721% but it might be wrong too... the variation should be in a long run the one in normal distribution. http://en.wikipedia.org/wiki/Dice You even sayd Posted: Today, 3:36 AM | Post #41 Oh nice, I was correct: "As the number of dice increases, the distribution of the sum of all numbers tends to normal distribution by the central limit theorem" Iam not saying it cannot happens, im saying its not normal to happens so often... whatever... lets play casino games. Its a shame because this is by far the best site to play risk (boards, players, graphics) but the dice is not the best thats for sure. Do you play risk for real? does it happens teh same rolls that happens here? not to me...

Ooh nice, you actually show your calculations, good! However, you calculate the probability of 6 consecutive losses, not the probability of 6 losses in a game. A small note, there is a difference between a normal distribution and the standard normal distribution, what are you using? Imagine this: A player joins the site, and for three consecutive games, he loses every single dice roll. Is this possible according to you? (with unbiased dice of course)