Realistic Dice mod

[Berzerkerz]

Is there a mod for these stupid rolled dice who destroy the game ?

[revernet]

well of course you can do the math if you want

but when my dices were
10 skulls, 11 both down, 24 push, 6 stumble, 10 pow

and IA dices were
22 skulls, 18 both down, 47 push, 24 stumble, 15 pow

I don't need some confidence interval to understand that the IA don't use a D5 dice for his block dices ;)

[RodHull]

Originally posted by revernet:

well of course you can do the math if you want

but when my dices were
10 skulls, 11 both down, 24 push, 6 stumble, 10 pow

and IA dices were
22 skulls, 18 both down, 47 push, 24 stumble, 15 pow

I don't need some confidence interval to understand that the IA don't use a D5 dice for his block dices ;)

Wait the AI got that many block dice on you ??

[licker34]

Though, drawing any conclusion from the results of one game is pointless anyway.

Maybe the AI uses the D5 in every 4th game!

Maybe the AI uses the D5 in every 4th game on Wednesdays!

Maybe the AI uses the D5 in every 4th game when the players team has more than 8 letters in it!

I think you need to test all of those hypothesis to be sure...

[Kejiruze]

Originally posted by RodHull:

Originally posted by revernet:

well of course you can do the math if you want

but when my dices were
10 skulls, 11 both down, 24 push, 6 stumble, 10 pow

and IA dices were
22 skulls, 18 both down, 47 push, 24 stumble, 15 pow

I don't need some confidence interval to understand that the IA don't use a D5 dice for his block dices ;)

Wait the AI got that many block dice on you ??

Looks like a lot of base base base herp derp derp

[revernet]

Originally posted by licker34:

Though, drawing any conclusion from the results of one game is pointless anyway.

Maybe the AI uses the D5 in every 4th game!

Maybe the AI uses the D5 in every 4th game on Wednesdays!

Maybe the AI uses the D5 in every 4th game when the players team has more than 8 letters in it!

I think you need to test all of those hypothesis to be sure...

can also think a few seconds about ... why would programmers make a completly different RNG for IA when then have 1 for PvP, look like a huge waste of time, resources and money

also I just report 1 game, but think i have 20ish of them, all similar
If you REALLY think it can be otherwise, just run thoses tests yourself and show me I'm wrong ;)

Originally posted by Kejiruze:

Originally posted by RodHull:

Wait the AI got that many block dice on you ??

Looks like a lot of base base base herp derp derp

my skill (or lack of skill :D) as coach is not the question here :P
but yes I should probably dodge a little more when facing orks

[RodHull]

Originally posted by revernet:

can also think a few seconds about ... why would programmers make a completly different RNG for IA when then have 1 for PvP, look like a huge waste of time, resources and money

also I just report 1 game, but think i have 20ish of them, all similar
If you REALLY think it can be otherwise, just run thoses tests yourself and show me I'm wrong ;)

I think Licker was being sarcastic :)

[Kejiruze]

Originally posted by revernet:

Originally posted by Kejiruze:

Looks like a lot of base base base herp derp derp

my skill (or lack of skill :D) as coach is not the question here :P
but yes I should probably dodge a little more when facing orks

Any team vs any team, in the vast majority of cases you don't want to base up.

[Ugh]

That is why I would say that this experiment should be done by someone from the skeptical and someone from the non-skeptical crowd together so that each can confirm that no tampering of the sampling has taken place. They can then analyze the new unbiased sample together or give it to their trusted statistical expert for analysis or whatever.

I am in principle not interested whether this theory is true or not, but the use of the MT itself doesn't preclude its possibility, and I would trust any independent unbiased analysis more than either the word of the developers or the analysis on a possibly biased sample (it's more likely that those who suspect RNG fraud will look for games that support it while those who trust in the developers will ignore those as outliers, if they are even found).

[VoodooMike]

Originally posted by "Ugh":

I talk a lot but will never do any work at all

Called it.

[Ugh]

Originally posted by revernet:

can also think a few seconds about ... why would programmers make a completly different RNG for IA when then have 1 for PvP, look like a huge waste of time, resources and money

It's very simple, you just have to put a 5 instead of a 6 as the parameter to randint and map the 4, 5 differently for AI . This is maybe two lines of code for the two different versions, so it's negligible effort. As to the why, well .. who knows. The theory proposed is to make the AI more competitive, i.e. compensate for its lack of proficiency in the game.

[Ugh]

Originally posted by VoodooMike:

Originally posted by "Ugh":

I talk a lot but will never do any work at all

Called it.

I'm doing plenty of very gruelling work in my personal life at the moment, proving other people's theories is the least of my concerns. What people like you think, likewise.

[revernet]

unbiased samples ? right
here are ALL reports from my games with an underworld team in an "old world" league
can't be biased by any "selection" or "feeling" as there are ALL what I have with this team

skulls, both down, push, stumble, pow, opponent race
12, 18, 27, 7, 20 underworld
17, 3, 25, 9, 7 wood elfs
6, 8, 21, 9, 10 elf union
13, 12, 39, 17, 16 nurgle
9, 11, 14, 9, 8 wood elfs
11, 15, 24, 10, 13 nurgle
23, 19, 35, 20, 19 khemri
20, 9, 32, 12, 18 goblin
19, 13, 30, 18, 14 kislev
17, 11, 27, 14, 16 underworld
22, 18, 47, 24, 15orks
6, 7, 20, 7, 9 wood elfs
7, 14, 17, 12, 8 goblins

total :

182, 158, 358, 168, 173

ratio
0,175; 0,152; 0,344, 0,162; 0,167

expected ratio for a 6 face dice 2 faces giving "push"
0,167; 0.167; 0.333; 0.167; 0.167

expected ratio for a 5 face dice
0.2; 0.2; 0.2; 0.2; 0.2

There is no worse blind man than the one who doesn’t want to see.

[VoodooMike]

Originally posted by Ugh:

I'm doing plenty of very gruelling work in my personal life at the moment, proving other people's theories is the least of my concerns. What people like you think, likewise.

If this was the first thread you'd blathered on in that might be a valid excuse... but this is consistent across multiple years, ugh. You don't do any work at all regardless of the circumstances.

You seem to be under the rather pedestrian impression that you are an important "ideas man", and that other people should do the work to free you up for your important thinking. Thing is, that describes 80% of the population: people who think they're deep and brilliant, but who are too lazy and/or lack the competence to test those supposedly deep, brilliant ideas they churn out like farts. Rather than risk damaging their self-image through failure, they never lift a finger.

Originally posted by "revernet":

here are ALL reports from my games with an underworld team in an "old world" league
can't be biased by any "selection" or "feeling" as there are ALL what I have with this team

You can't simply eyeball data because that's the same method the people who hold the opposite belief used to come to their conclusion. It absolutely doesn't matter how obvious you feel it is there needs to be an objective method for testing the distributions... which there is: the chi-square goodness of fit test.

Here is a simple online calculator for chi-square goodness of fit.[mathcracker.com]

You need only enter the raw numbers in the first column, then the corresponding expected proportions (which you're calling "ratios") in the second, and set the significance level (the standard is 0.05, or 95% confidence) and voila you have an accepted standard test that doesn't involve any feelsies on the matter.

[revernet]

thanks


the "6 faces dice" theory has a "chi-squared summ" of 2.396 what is bellow the 9.488 limit
and so is not rejected

the "5 faces dice" theory has a "chi-squared sum" of 137.155, way above the 9,488 limit
and is rejected

[RodHull]

Originally posted by Ugh:

I am in principle not interested whether this theory is true or not

But thought you'd spend pages arguing about it anyway when even the OP has given up defending the position. So to summarise you don't actually care have no actual data to support one side or the other, but feel 'someone' should test all possible theories to see if the developers are telling fibs or not. That about sum it up?