Abstract
Magnitude frequency relation of earthquakes expected from the Go game model which was proposed in the second part of this paper is compared with actual data with special emphasis on the characteristic bent of magnitude frequency relations. Of particular interest is that the multiplicity of magnitude frequency curves as well as the characteristic bent can be expressed by assigning proper values of p to the Go game model.
It is also demonstrated that a formula recently presented by Utsu to express general downward bent of the magnitude frequency relations yields almost identical solution with Go game solution when parameters are properly chosen.
