According to the COP documents, the caller decides the level and the judges the grade with differing criteria for each. Seems like extension, flexibility, body line are big parts of the grade, not the level. In the GPF Sasha ended up with the higher total element score even with a lower level rating; Shizuka 4.0, Sasha 4.3. Sasha did have a level 3 spiral in her Trophee Lalique short program. Her total element score was 6.2. Michelle could have a higher difficulty spiral but end up with a lower score than Sasha just as Shizuka did.