The image at left shows BOG's view of the Case of the Purloined Light Ray.
Clicking on the nominal angles (45 pavillion and 44 crown) produces a rather dull stone, with only 33% ISO brightness and extremely poor star facet performance (see GemRay rendering and ISO values in the left panel).
Sliding over to the left into the better region (red area) to a pavillion angle of 37 degrees produces much better performance - the star facets light up brilliantly. Of course, it is cut well below the critical angle and does fish-eye in the center. Nevertheless, the ISO brightness, at 72% is more than twice as high as before.
Clicking around on other parts of the map does not produce a prettier solution. Some designs, Vargas' SRB among them, just don't do well with low index material.
This example shows how BOG can help pick the best cutting angles, based on the actual stone appearance as well as more objective criteria, such as ISO brightness. For a real stone, I would also include tilt performance (at least) in the evaluation.
The actual reason for this behavious was explained by Gregg "Sherlock" Glenn in a previous post.