This project has moved and is read-only. For the latest updates, please go here.

How to implement a Goodness of Fit test for a supposedly leaded dice?

Sep 7, 2014 at 11:58 AM
Edited Sep 7, 2014 at 11:58 AM
I seem to be unable to find out which classes and methods to use to check the resulting P value for a dice roll that gives a 19:1 probability to produce a six in stead of a one.
So when making 60 dice rolls, the expected frequency of any number coming up top is 10.

In my "leaded" dice the observed frequencies are:
1: 1
2: 10
3: 10
4: 10
5: 10
6: 19

ì entered the expected and the "observed" values on this website: http://www.quantpsy.org/chisq/chisq.htm and the resulting Chi Square value was 16.2 and the P value was 0.00629, indicating that the dice, indeed, has been tampered with.

I am unable however to produce this result using Metanumerics. I tried constructing an ContingencyTable and called the PearsonChiSquaredTest on it. But I get a P value of 0.07. I am unsure how to fill the table, it this is the rigt approach at all.

So, how to do t?