Again, Nice graph.

So the 0.1 density is the poorest of all and yet is the basis of traditional Zone System tests.

This is where I come to the conclusion for myself... I base my EI speed on 0.1 density of the curve that meets ISO conditions closest...

Then I take Stephen's example in the Delta-X Criterion paper as proof of "approximate invariance of speed" (my quotes), to justify not changing my speed regardless of the developing time I choose.

I do the tests for any particular film/developer combination - to find the closest ISO curve and take its speed. The rest of the curve family gives me Time/CI which I use to fit development time to shooting conditions.