Dice analyzer 1.5.1a, 1.5.1b and 1.5.1c

[gavin_sidhu]

So whens the next update coming Cyberdaniel? I really want to see the individual stats for the dice now that World 2.0 has come out.

[cyberdaniel]

It's coming, I just have to finish my finals first. I do want to get in the second semester of university.

[zip_disk]

25k attack rolls, 12.5k defender

[sully800]

Nice zip! Very even!

[cyberdaniel]

Ok, I've been working on the new update and I've stumbled into a potential problem. The thing is: Do you guys want the individual game statistics for the games you're playing (they still remain for 8 days after the game is over) or do you want them for future reference.

In the first case there is no problem, however in the second case there will be a need for another storing system (cookies will not be enough).

[cyberdaniel]

OK, the next update (1.4) is ready. Go to the first post to see the updates and to get it.

[wicked]

I'm not suprised the dice AVERAGE out ok over time. That's what an average of a large data set is... the median of all the spikes and dips. However, in the case of CC, the spikes and dips are what people are complaining about, not the average. People want to see a more even distribution in a much much smaller sample size, i.e. they want "fair" dice over their 3-4 rolls per turn, not over their 20 games over 3-4 weeks. So yeah, keep analyzing... but I think you're analyzing the wrong thing.

[AndyDufresne]

Hehe...the whole idea of fair dice over 3-4 rolls per turn is absurd...due to probability. Anything can, and most likely will happen, in that short span.


--Andy

[wicked]

Never said it was possible, just said that's what people want.

[Evil Semp]

I don't question the percentage of the dice I question the true randomness.

I have been keeping track of 6's rolled for me on offence and for the defender. I rolled 497 6's and the defender rolled 250. The defender killed 110 of the offinsive 6's or 22.13%. The defender rolled 250 6's and killed 110 of the offinsive 6's or 44%. that percentage seems a little high to me. There were a total of 3035 offensive dice and 1163 defensive dice.

I think enough people have been complaining about the dice, but here is some proof. Maybe someone can look a little more into this to see if if there is a problem with the dice.

[silvanthalas]

If I could make a small request.

Dice Analyzer currently loads into a new window.

Any chance it could be set to load in another tab of the current window?

Also, it appears I found a problem. If I'm on the Dice Roll Simulator tab, and then click to go back to General Statistics, it doesn't do anything. I have to reload the window to get back to General Stats.

[previsualconsent]

for some reason my analyzer doesn't show specific game info.
there is no tab for it. are you in the process of changing that?

let me know

[TFA-LAPDANCE]

The dice analizer is basically pointless. It doesn't matter what you roll, only when you rolled it. I have been involved so far in a very limited amount of games and in almost all of them I haven't went with any type of odds. More than once I have attacked one army with 7 or more and have lost 6 out of the 7. A game I just recently started, I lost 13 armies while I only killed 1.

I think before anything can be done, the game needs to start showing the rolls in the box where it displays the actions and placement of armies. I seriously doubt the true ramdomness of the rolls. Good know how many games of real risk I have played and I have never seen screw jobs like I have seen here.

[sully800]

Hehe...the whole idea of fair dice over 3-4 rolls per turn is absurd...due to probability. Anything can, and most likely will happen, in that short span.


--Andy

It's not that every short span should be identical and follow what the odds predict- the true test of randomness is whether everything in the short term follows the distribution you would expect to see based on how often it should occur.

If you lose twice in a row and win twice in a row much more often than would be expected there is certainly a problem. But even with 4 rolls it would reflect the expected distribution, which creates a problem for finding and analyzing any trends that seem out of the ordinary.

The dice analyzer has effectively proven what most people suspected all along. If someone wants to actually analyze the randomness of the dice a program will need to be created that keeps track of streaks. How many times do you win 2 in a row? 3 in a row? etc. This would need to be done for each rolling situation, though I think 3v2 and 3v1 are the only ones that really matter. Then once a ton of data is collected those occurrences could be plotted and we would see if it matches the expected distribution. Lots of fun work huh?

And you could do all that and the trend may match the distribution exactly....at least it would settle the debate once and for all though!

[KingCoolDaddy]

I think it's probably just YOU that the dice don't like.

Statistically, out of a 1000 rolls, in a 3-on-2 situation, ON AVERAGE,
Attacker wins both about 37% on the time, Both lose 1 army 34% of the time, and the Defender wins both 29% of the time - roughly.

Keep in mind that this is an AVERAGE. Some days you end up above the averages, and some days below. And as many others have stated - You remember the bad rolls way more than the good.

Happy rolling!

[subjekt]

Damn for some reason mine wont run??? all the other scripts run? any ideas?

[yeti_c]

I don't question the percentage of the dice I question the true randomness.

I have been keeping track of 6's rolled for me on offence and for the defender. I rolled 497 6's and the defender rolled 250. The defender killed 110 of the offinsive 6's or 22.13%. The defender rolled 250 6's and killed 110 of the offinsive 6's or 44%. that percentage seems a little high to me. There were a total of 3035 offensive dice and 1163 defensive dice.

I think enough people have been complaining about the dice, but here is some proof. Maybe someone can look a little more into this to see if if there is a problem with the dice.

You might be onto something here...

Way too many times are my 6's beaten by a 6 and rarely do I win with a 6!! I'm more likely to win off of a 4 than anything else - which breaks the stats...

Perhaps we should be analyzing these Stats - this would need to count the number of combinations received and display those.

C.

[Fireside Poet]

I am working on a couple of theories. One is a "morning dice" vs "evening dice". I find that if I play in the mornings, I have a proneness for getting more 1's, versus the evenings where I seem to have better ... luck? Maybe the creators of random.org were hard at work in the hours of 8-5 and then after that, the night crew slacked and was boozing it up. I don't know, just sharing my equidelirium with you.

My other is the "Every Game Is Winnable" theory. I plan on collecting a big prize for the completion of that work. Estimated time of completion: TBD.

Yes, I'm having a slow day today and I'm behind in my shipment of Guiness to the boys at random.org for this evening.

[cyberdaniel]

Hmm, how could that work though since the 100000 dice throws that get updated daily are updated at once not over a period of time and random.org generates these numbers on the spot, it doesn't collect them in a database.

[cyberdaniel]

Damn for some reason mine wont run??? all the other scripts run? any ideas?

Do you have the "Dice Analyzer" link on the left menu (Personal Menu)?

[Birdy_ch]

Does this script also run on Mac OSX?
I have downloaded it, but somehow I'm not able to start it...

[cyberdaniel]

Yep, you need firefox and the greasemonkey extension for firefox. After that just go to the script and an install message box should appear.

[max is gr8]

will there be the individual game stats again any time soon?

[max is gr8]

bump

[cyberdaniel]

My current implementation uses cookies but, unless you have only 4 games going, you're gonna end up with a lot of cookies and at a certain point the most important one will get deleted by the web browser. If anyone knows a nice way of saving them on the users hard drive let me know. I was also thinking of making a web server that will host these stats. That wouldn't be hard but there will be security issues since anyone can change these scripts.
However this idea would make possible to view everyones luck in a certain game.