A direct comparison between the observed WR/WRprogenitor number ratio within 2.5 kpc from the sun and the predicted value (using evolutionary computations of single stars of Maeder and Meynet, 1987, A.&A.182, 243) reveals a discrepancy of at least a factor of two. In a previous study (Vanbeveren, 1990, A.&A. in press) I proposed a solution based on the incompleteness of the observed OB type star sample within 2.5 kpc from the sun. In this summary, I propose a theoretical explanation for the discrepancy. The theoretically predicted WR/WRprogenitor number ratio critically depends on the adopted M formalism in evolutionary computations during the red supergiant phase (RSG) of a massive star, especially in the mass range 20-40 M⊙. Since any M formalism predicts the mass loss rate with an uncertainty of at least a factor of two, I have tried to look for solutions for the WR/WRprogenitor problem by using different values of M during the RSG (in the mass range 20-40 M⊙); the M values and formalism that were adopted were always choosen within the observational uncertainty (i.e. within a factor of two when compared to the formalism used by Maeder and Meynet, 1987).