Are the dice fair on this website? Dice log and analysis. vF

[army of nobunaga]

I guess everyone knows by now that I have endless time to screw around. With that, coupled with the fact that Im more of a doer than a complainer, I have decided its time to rest this dice argument. I kind of get tired of endless complainers that cry about a sore foot, but never check their shoes.

I am going to record defensive dice until its statistically relevant. I believe the statistical term is that I have to have a high enough "power" to make the results meaningful.

For the record:
-I have no idea going into this if the dice are statistically random or not.
-I will not be surprised with results of the die showing biased or showing complete randomness.
-I will keep this as scientific as possible considering time constraints (I have a few) involved with my life.
-I will not lie in any way. I will try to record all data by video and upload the results on a file share site so my data are reproducible.



The end graph of my tally should look like this 2 die graph below. I understande there will be varience and I hope to calculate it to make sure the graph is statistically sound. Anacova or something... been a while since ive been at school.










Ill try to update this every day. And I might even show a neat bar graph tomorrow. But until I get the right power, Im more concerned with scientific data collection.

Vers1



vers2



vers3


vers4



Final version

[army of nobunaga]

ok at end of turns today, power at 79ish... nice little bellgraph forming. Sort of like seeing a baby in the womb. I think I see its nose mama!

One turn today- out of 9 doubles def dice rolls, the guy rolled "7" 7 times... Creepy.

[AAFitz]

Define fair.

[army of nobunaga]

I think the burden of the word definition lies upon the 100's of people that complain about dice being "unfair" and making threads and posts on the subject. For me fair would be for dice to operate in a fashion that they would in real life. Expecting a better chance for 7's to hit, a little less of a better chance of 6's and 8's to hit and so on. Unfair dice would be hot and cold dice that are not only unrandom but given to quarks throughout the day. Noone would call dice fair where in one turn a person rolls through 5 territorys losing 4 armies to an opponents 10. That would be unrealistic and thus what I would call "Unfair".

[army of nobunaga]

vers 3 out. percentage of '7's rolled is out of whack. Lets see if it evens out as the number of data entered increases.

[AAFitz]

I think the burden of the word definition lies upon the 100's of people that complain about dice being "unfair" and making threads and posts on the subject. For me fair would be for dice to operate in a fashion that they would in real life. Expecting a better chance for 7's to hit, a little less of a better chance of 6's and 8's to hit and so on. Unfair dice would be hot and cold dice that are not only unrandom but given to quarks throughout the day. Noone would call dice fair where in one turn a person rolls through 5 territorys losing 4 armies to an opponents 10. That would be unrealistic and thus what I would call "Unfair".

But if you play a thousand games, it would be unfair for this not to happen. It certainly would in real life.

[army of nobunaga]

I have played 1000 games. And did it in a span of about 3 days. The dice may "seem" hot and cold. But in reality they should be obeying basic laws of statistics. Just as so far in this little analysis Im doing, they seem to be.

[AndyDufresne]

I think what is needed is a little flashing icon in the top right corner, of flaming intensity cubes. Thus, whenever the icon pops up, that means the cubes are hot. When it is gone---advised on your own risk to take your turns and roll.


--Andy

[Falkomagno]

The problem here is call STREAKNESS.

That's why people recommend to not use Auto attack (which is utter silly answer, imho)

show me a model which proves, or refutes that statement " The dice is streaky". Then, I will believe you.

[natty dread]

Blame the atmosphere.

[army of nobunaga]

The problem here is call STREAKNESS.

That's why people recommend to not use Auto attack (which is utter silly answer, imho)

show me a model which proves, or refutes that statement " The dice is streaky". Then, I will believe you.

I proving RIGHT now with this that there is no "Streak" there are bad and good rolls that I have been getting.. but Im about to post vers 4


I am starting to believe the "Streaks" and the "Hot and cold" is actually a myth. Its just a damn myth we make up in our "Why is it ALLLWAYS me" programed heads.


I think instead sir.. that you people that talk about streaks... should actually PROVE the streaks.. HOW about some screens shots.. I want to see it! But you guys never try to prove it. you jsut log on to the forums and complain about how unlucky or unfair the dice have treated you.


well cheers. Here comes vers 4

[army of nobunaga]

vrs 4.. the 7's dropped a little. But still outside what I would expect. At 500 dice inputs, if It still looks out of whack Im going to start using SAS to try and get a number to put to this. I would also say at 500 rolls (just a guess) that the power is strong enough to make something this far off the curve point to bad dice.

But I am optimistic that it will even out. Ill make no more turns tonight so tomorrow we will see vers 5.

[Timminz]

Didn't all this bullshit used to get lumped together into one big "teh dicez sux0rs" thread?

[army of nobunaga]

dont know.. I rather create bullshit than read it.

[RADAGA]

I think instead sir.. that you people that talk about streaks... should actually PROVE the streaks.. HOW about some screens shots.. I want to see it! But you guys never try to prove it. you jsut log on to the forums and complain about how unlucky or unfair the dice have treated you.

The problem is that the site works against you when talking about screenshots. HOW can I screenshot a streak of seven 1x1 losses when the dice windo refreshes every roll I make, and there is NO WAY to see that round´s history?

Even if I autoattack and experience something like that, the results window show only two and a half rolls, we have to scroll it to see the results.

So, screenshots are impossible. Unless you are suggesting we FILM our desktops, and post them, wich, then, I´d say it is WAY too much trouble just to prove something we know happens to people we know don´t care about it, as long as the money keeps flowing in.

[army of nobunaga]

yeah I meant film it.. but good point.

[Trephining]

It would be better to actually log the dog results themselves, not the sum of the two dice. Rolling (3,4) is not the same as rolling (6,1), as (6,1) guarantees one off the attacker.

[army of nobunaga]

agreed. But that is too labor intensive. And I am assuming in this project that if the totals are normal, then the individual die are normal. because the bellgraph works off of a ratio chance of each throw of 2 dice. THAT ratio constitutes the various permutations of two different die together.

[army of nobunaga]

at 255 sample, there should be 42 7's. there are like 48. Other than that I see no major problem with the dice, and I think that being 6 die off of the stat is fine.

This project is too labor intensive and too under-appreciated. I proved the dice are fine for myself, and that is really all I was after. In the last couple of days, I have concluded:

-Hot and cold streaks are part of your(our) psychy
-The Dice actually perform fine.
-Even on the rolls I got CLOBBERED on, when I broke it down, it wasnt that crazy, and It all averaged out fine.

[AndyDufresne]

I'm glad you've reinforced your own initial thoughts---though we all know that unfair/random/streaky/etc intensity cubes will be a persistent issue until we do with them completely.

Maybe a new feature, that after the first round, the game ends, and assigns victory to whomever it feels like. But even this...can't escape randomness neither! I don't think we'll ever be able to escape it...


--Andy

[retrocrush]

Just attacked an army of 21 v 5 and end up with 1 v 4 ... give me a break. Wonder the odds of that are? Have to be ridiculous.

[AAFitz]

Just attacked an army of 21 v 5 and end up with 1 v 4 ... give me a break. Wonder the odds of that are? Have to be ridiculous.

the odds are ridiculous, but youve played nearly 2000 games and have probably thrown hundreds of thousands of dice. Now, if you lose 21 to 5 every time, thats ridiculous.

but remember, 21 to 5 losing is a difference of 15 armies, but its just 10 bad rolls, with the defender winning ties. Ten bad rolls is not that ridiculous. You could easily flip a coin ten times and get the same result eventually too.

If you sit down to one game, odds are you will see some pretty average rolls. If you sit down to 2000 of them, odds are you will see some pretty ridiculous rolls.

The problem is, when you sit down to your 2000th game, and see something insane, you forget about the 1999 games that were more or less normal. You forget all about the countless perfectly seemingly random dice, and then see one roll which you think means all those other rolls were not random.

Its really basic psychology.

[army of nobunaga]

I think we finally have to agree Fitz... its leaving a bad taste in my mouth too =D

[Fruitcake]

.....Fitz... its leaving a bad taste in my mouth too =D

apparantly that's what all the girls say to him.

[RADAGA]

You could easily flip a coin ten times and get the same result eventually too.

Easily? 1/2^10 chance = one streak of ten in every thousand and twenty four....

so, is it easy to flip ten thousand times a coin? 2 seconds for each flip, sit nearly six HOURS doing nothing but flipping a damn coin. If thats easy, I dread to see what you consider hard.