Effect of ink load on gamut volume?

Having a grand old time comparing profiles on iccview.de, which gives a very nice 3d graphic comparison any two, type 2 profiles. Also gives information about gamut volume and lots of other stuff. All of which probably can be completely wrongly interpreted by persons such as myself.

I noticed for instance that the Crystalline profile I generated for the 8300 is about 9% bigger than the BC canned profile, but both profiles' "gamut balloons" are almost exactly the same shape and generally overlay each other when graphed LAB mapped to YXZ.

I was wondering what might account for the increased gamut volume in my profile, beyond different instrumentation and software.

Mainly I am wondering if I could further puff up the gamut balloon by running profiles with different inkloads. Which is a lot of work per iteration! Can anybody give me some insight as to what I could expect to see from ink load variations, and what direction might be optimal for a first approximation. I suspect I might be near optimal, and that BC's lower volume might represent some useful compromise of which I am unaware.

******************************************

FWIW my iPF8300 Crystalline profile is the wireframe, BC's iPF8300 Crystalline profile is the colored envelope. The 90 view is very similar in character.