the inconsistency depending on fixpoint-selection

[Gottfried]

Again I didn't see some simplification...

Here it is:

Code:

´  using      g(z)= e z^2 + (1+(p-q)) z
   setting    e = (1+(p-q))

this is

Code:

´     g(z)= e z (z +  1)

second power of g(z) (not "iterate"!)

Code:

´   (g(z))^2 = e^2 z^2 (z+1)^2 = e^2 z^2 (c(2,0)z^2 + c(2,1)z + c(2,2) )

where c(a,b) is binomial(a,b) and this gives a binomial-matrix column-shifted accordingly to powers of g.

Can we describe the half-iterate from this?