Moments and binomials

 

My model of three semi independent binomials, one for spin, charge and magnetism (x,y,z).

If I manage three binomials I can equalize them such that there is no noticable kurtosis, or better said, the kurtosis is in the round off error. So I construct a composite binomial of the tree matched binomials. Is will have a p of one third and will have no noticible kurtois because Markov 3-tuples bury the kurtosis in round off error.  

If I add a relative prime and make Markov 4-tuple, then the next moment up is removed from round off error.

There is a binomial relationship which match pr and N such that Markov n tuples are created with properly aligned binomials. That is what the Markov is really doing. 

The value n*(x*2 + y^2 + z^2 +... ) should always give the weighted mean of a binomal that aggregates the other N-1 binomials.  The x are the pr, the share of bandwidth allocated. np is the mean.

So npp is the weighted mean when p=x represents the probability of picking x as heads. So using this tricks and collecting the proper terms on 3xyz, it should be possible to reconstruct x,y,z as shares of coins toss with x+y+z = 1.  Then calculate n, then work backwards to recover  the Markov 3-tuple.

I think this works, there is a lot of cancellation going on because probability of selecting a binomial is related to number coin. I have not verified all of this, I am to careless. But the x,y,z should be shared according to entropy.