Could you be using the real-space R-factors instead of the corr. coeffs?
This archive was generated by hypermail 2b29 : Tue Mar 20 2007 - 03:13:20 CST