Dice analyzer 1.5.1a, 1.5.1b and 1.5.1c

[tattooedude]

i will track my dice if you want. Something is wrong, i have gone four rounds without winning a throw, lost 15 vs 1 battles, and the list goes on. Very frustrating.

[lhcguy]

Hi.. i have not read all the 20 pages, but i think there should also be an measurement of the streakiness of dices, since most users complain about too long streaks.

Can you please add some more information in the script?

I.e.:
% that a dice is the same as the dice before
% of all 3 dices in an attack role are the same?
% of all 2 dices in a defensive role are the same?
% of the outcome of a battle is the same as in the battle before? (win, tie, loose)

If you add those in your script, i can get the ideal numbers for you with some math.
This might shed light on the streakyness of the dices (although i believe it will simply show those people, that they are superstitious)

Thx

[lancehoch]

Hi.. i have not read all the 20 pages, but i think there should also be an measurement of the streakiness of dices, since most users complain about too long streaks.

Can you please add some more information in the script?

I.e.:
% that a dice is the same as the dice before
% of all 3 dices in an attack role are the same?
% of all 2 dices in a defensive role are the same?
% of the outcome of a battle is the same as in the battle before? (win, tie, loose)

If you add those in your script, i can get the ideal numbers for you with some math.
This might shed light on the streakyness of the dices (although i believe it will simply show those people, that they are superstitious)

Thx

The user who created this script has not visited CC in a while. It does not seem likely that he will make any changes to this script. One possibility is if another user decides to adopt the script, but I do not know how likely that is.

[Kranberry]

I have a question about the odds. Where are people getting the odds for this and other scripts. I have created a spreadsheet that shows all the possible combinations of dice throws for 3v2. And according to dice analyzer the ideal win rate for 3v2 dice is ~37% of the time. I don't get this number. Mine is totally different...just wondering where people are getting their odds.

[scottp]

here's one statistical analysis:

http://www.plainsboro.com/~lemke/risk/

there are a bunch if u google "risk dice odds"

[Kranberry]

Thanks found a slight flaw in my logic and it has been fixed. My bad...

[Teflon Kris]

What a coincidence that dice analyser disaappears then the dice are rediculous for game ... after game... after game

Seriously though- Any chance of defending dice analyser? And opponents dice analyser?

Without that there is no way of telling if you are being ripped-off? I'm sure many players use dice analyser to check there isnt a dice conspiracy against them

Oh, and card analyser too

Another potential use of dice analyser could be to compare results and dice - to do this it may need to be tweaked to show game dice and recent dice (although again, for a full picture, defending dice and opponents dice would be needed).

Just some ideas- nice tool though

[istanbul39]

Is dice analyzer disabled? Tried to download and getting a script error.,

[DuGa]

Why does this appear in the top of my dice analyzer:

Game not found!! Cookie name: 'conquerclub'?


Thanks.

[Teflon Kris]

How many million dice need to be thrown before it is statistically impossible to have less than ideal dice?

[sbuck143]

i am discovering that the dice seemed to be skewed towards the defender an inordinate amount. I will install this script to verify through that.


Note I'm not just complaining as an attacker, but i've been the beneficiary of several of these events as well. But having been here for only 1 week, and already having been a party to at least a dozen 1 in a million events (literally 1 in a million, like 23 v 3 , 12 v 2), i am already extremely suspicious of the dice. Defenders win 3 v 2 battles way way more often than they should, and seem to roll double 6's like its going out of style.



What do we know about how the random dice are indeed generated randomly? There are whole college departments devoted to making computer randomizers truly random. What do we know about the seed used here?

[RADAGA]

i am discovering that the dice seemed to be skewed towards the defender an inordinate amount. I will install this script to verify through that.


Note I'm not just complaining as an attacker, but i've been the beneficiary of several of these events as well. But having been here for only 1 week, and already having been a party to at least a dozen 1 in a million events (literally 1 in a million, like 23 v 3 , 12 v 2), i am already extremely suspicious of the dice. Defenders win 3 v 2 battles way way more often than they should, and seem to roll double 6's like its going out of style.



What do we know about how the random dice are indeed generated randomly? There are whole college departments devoted to making computer randomizers truly random. What do we know about the seed used here?

You wont get anywhere. People here say it is random, and there is no way to prove it is not, even if you have a streak of three million attacker ONES versus two million defense ONES, it still can happen in a random streak. I just got a 10x4 turned to ruin because defense got 6,6 / 6,6 / 6,6 / 6,6 - this should happen once every 1.6 MILLION rolls.

To roll 1,6 mi. rolls, making 16 attacks a day in ten games (160/day) you would need 10.000 days, or about 30 YEARS playing...

Oh, and I have wrote Random.org about it once:

Don´t you think your numbers wield far too many streaks in a row to be considered random? I am a ConquerClub user, they use your numbers and claim them to be perfect. But the number of times we find strange things with the results is amazing.

(rolled 30 numbers at their generator and posted the results)

How can we trust it, if every time we get the die roller, it shows results that "are possible" but "should happen once every many rolls", not on "every other" roll. On those 30 dice, for instance, I´ve seen only one FOUR.. ....one four in 30 rolls ... where it should be something around five.

but I got 10 sixes .. 33,3% of the dice.

The reply>

Hello André,

So I'm finally finding some time to reply to your email. I have no idea who ConquerClub is and I have no commitment to them to reply to their users' emails.

Go figure....

[72o]

Radaga, have you installed the dice analyzer?

[RADAGA]

Radaga, have you installed the dice analyzer?

Yep, and with tens of thousands of rolls I have a 3x1 stats are (63.26% / 36.74%)

[72o]

So, if that was your percentage out of 10000 3v1 attacks,

65.97-63.26=2.71%*10000=271 more ones you should have killed out of ten thousand tries.

I don't think this variance is statistically significant, out of a relatively small sample size. Do you have big variances in other rolls, like 3v2?

[Teflon Kris]

So, if that was your percentage out of 10000 3v1 attacks,

65.97-63.26=2.71%*10000=271 more ones you should have killed out of ten thousand tries.

I don't think this variance is statistically significant, out of a relatively small sample size. Do you have big variances in other rolls, like 3v2?

Hi, I assume you are a qualified statistician?

If not, what you think doesn't count.

What's the standard variation for above stats?

[RADAGA]

Yes, to the opposite side (otherwise I would be a cook)

3x2: (42.63% / 32.71% / 24.66%) versus ideal (37.17% / 33.58% / 29.26%)

[Capitino]

I've been using this for a while and find it interesting. I have not read all 20 pages of comments, but is it true that you cannot see anymore what your statistics for a paricular game are?

[Capitino]

It looks like the dice analyser has not been updated in a while and that the one who created it, has not been around in a while. If possible I would like to see if I can develop it further. I'm not very experienced with greasemonkey scripts (as in not at all), but I should be able to learn it fast as long as I have something to start with. Is the source code for this script available somewhere?

[chipv]

It looks like the dice analyser has not been updated in a while and that the one who created it, has not been around in a while. If possible I would like to see if I can develop it further. I'm not very experienced with greasemonkey scripts (as in not at all), but I should be able to learn it fast as long as I have something to start with. Is the source code for this script available somewhere?

GM scripts are public source. When you install them, they get copied to your PC. You can see where by going to
Tools->Greasemonkey->Manage User Scripts then click on the script then the "Edit" button.

Or click on the userscripts.org link in the head post and then click on the green install button. You can see "Show Script Source" click on that.

Looks like hecter is hosting it right now, but you are right, no updates recently.

If you want to take this on, please PM both hecter and cyberdaniel and let them know, thanks.

[Capitino]

Thanx a lot! Never worked with JavaScript before, but I know C/C++ so it looks very familiar. First I'll go and have some fun with it myself. If anything useful comes out of it, I'll let all of you know.

[Capitino]

For anyone interested:
I did some looking at the dice analyser. It's major problem at the moment is the fact that after throwing a couple of thousands of dice, everything is probably pretty much 'averaged out'. By then also a few extremely lucky or unlucky rolls in a game, don't make much difference for the overall statistics, but it still made huge difference for that particular game! So basically, in my humble opinion, the dice analyzer is what it is and if you want to improve it, it would be analysis per game. That, unfortunately will be a bit difficult.
The problem is that there are people out there playing 300+ games at the same time. Now, although the information that needs to be stored is not that much, it still would not fit in one cookie. A single cookie for every game would not be possible since there is a limit on the number of cookies. A solution to this could be managing of the user which games (s)he wants to have tracked by the dice analyzer with a certain maximum number of games. Problem here is that this is not very user friendly and most of the time you only think afterwards that you wanted to see how bad/good the dice were and then it didn't track it.
So if people want dice statistics per game then, as far as i see it, the gods of CC need to keep track of this in the game. This would mean that for every player you need to count a certain amount of events:
6 counters for the dice thrown
14 counters for win/loss (3 counters for 3vs2, and 2vs2, and 2 counters for 3vs1, 2vs1, 1vs1, and 1vs2)
and maybe double it to separate attack and defence, which leaves 40 counters per player, but then you can see very very awesome dice statistics, like who was the luckiest player in the game and also, on a more general level, how often the luckier players win and thus we can shine a light on the matter what is the ratio between tactics and luck.
Downside is, of course, that the game could be going a bit slower, but i would not know how much that would be.

[Dako]

[url=Solution]http://diveintogreasemonkey.org/api/gm_setvalue.html[/url].

GreaseMonkey can save unlimited amount of data into inner variables and store them locally. You can store all the information for the past games you had (thou it will eat some space) and can retrieve them as easy. But who need old data? We can create an expiration mechanism that will delete previously finished games (7+ days old).

How is that?

[Capitino]

if greasemonkey allows to store in files, that could be very handy. I'll look into it.

[Dako]

It's not files, more like private script cookies. Only that script can access them and their size is unlimited.