[phenixbb] discrepancies between completeness stats

[Kendall Nettles]

Dear PHENIX developers, 

We are getting a group of structures ready for publication and noticed that the table 1 description of the completeness is showing higher completeness in the last shell than overall. See Table 1, attached, for one example, with completeness at 81.73 (96.75). This table was calculated from the last major release of PHENIX, and was the same on the latest nightly build. The data was highly anisotropic so we truncated it with the UCLA diffraction anisotropy server. Validation program suggests completeness is 13% in highest res shell, which makes more sense with anisotropic data.  I also noticed that the refinement log lists the overall completeness at 85% while the validation and table 1 programs give 82%. 
Thanks for you advice!

Kendall