1: \begin{abstract}

2: 

3: We review the assumptions on which the Monte Carlo renormalization 

4: technique is based, in particular the analyticity of the block 

5: spin transformations. On this basis, we select an optimized

6: Kadanoff blocking rule in combination with the simulation 

7: of a $d=3$ Ising model with reduced corrections to scaling.

8: This is achieved by including interactions with second and third 

9: neighbors. As a consequence of the improved analyticity properties,

10: this Monte Carlo renormalization method yields a fast convergence 

11: and a high accuracy. The results for the critical exponents are 

12: $y_H=2.481(1)$ and $y_T=1.585(3)$.

13: 

14: \end{abstract}

15: