In article <js2l98$2pln$1@matchbox.inf.ed.ac.uk>, I wrote:
>Call the number of ways that even N can be expressed as the sum of two >(unordered) primes g(N). g(N) tends to be much higher when N is a >multiple of 3, because only a third of the "potential pairs" a+b=N are >ruled out by a or b being a multiple of 3, while for other numbers two >thirds are ruled out.
