<div dir="ltr">On Fri, Oct 3, 2014 at 11:58 AM, Sneha Rangarajan <span dir="ltr"><<a href="mailto:rsneha@umd.edu" target="_blank">rsneha@umd.edu</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="white" lang="EN-US" link="blue" vlink="purple">
<div>
<p class="MsoNormal"><span style="color:rgb(31,73,125);font-family:Calibri,sans-serif;font-size:11pt">I did another round of refinement with default settings (XYZ,realsp, IndB and occ) with and without weight optimization.</span><span style="color:rgb(31,73,125);font-family:Calibri,sans-serif;font-size:11pt"> </span></p>
<p class="MsoNormal"><span style="font-size:11.0pt;font-family:"Calibri","sans-serif";color:#1f497d">Without weight opt, the Rfactors are 23/36 with RMSbonds-0.0108 and RMSangles-1.750</span></p></div></div></blockquote><div><br></div><div>One idea would be to run AutoBuild again. I've seen cases before where it didn't converge using the default settings, and feeding a previous result back into the program for a second run produced significantly better models. It might help get rid of the overfitting.</div><div><br></div><div>-Nat</div></div></div></div>