In order for Scaffold to have access to enough data to properly build probability distributions, Proteome Software suggests users increase the peptide and protein FDR values to 1.0 (note the default is usually 0.01 or 1%). In Spectrum Mill, this can be done by validating your data (after extraction and searching) with peptide and protein FDR values of 100% (eg, 1.0).
Scaffold 4.8.1 introduced Prefiltered mode, which allows users to bypass Scaffold’s built-in probability model and filter your data using the search engine instead. If you would like to use this option you should not set the FDR value to 1.0 as described in the section above but set the value to an acceptable FDR cutoff. If you use Prefiltered mode you will not be able to filter your data in Scaffold using the Protein and Peptide dropdown menus.