I've noticed frequently that elliptic looks like crap whenever the bottom is not the same as the top. My usual solution is to simply crop off the bottom half, but lately, I discovered that Apophysis doesn't have the issue. The problem is in the JWF elliptic. For whatever reason, we have:
// if ( pAffineTP.y > 0 )
if ( pContext.random() > 0.5 )
Notice that the correct implementation is commented out (but was originally in there, apparently). I understand that yes, if your fractal is always on one side of the x-axis, it's nice to have it reflected to both. But 90%, I don't want this. Just try rotating the x-form's pre-matrix by 45 degrees and you see a mess because of the imbalance.
I propose that the Apo method be used, but for the sake of offering both solutions, why not have a parameter that allows using the random # option if desired? This shouldn't be too much trouble to implement, and it would allow JWF to correctly load Apophysis fractals that use Elliptic.