Goldbach has too many trivial results is our main problem. Even $p+q=2n\implies n-1=k+j$ where $k,j$ are not of the form set out by the sieve of sundaram for values that $2l+1$ is composite.