Is GGV Convex Out? Might need new method

[kierancoz]

    I have a concern on the G-G diagrams i've been seeing this producing. The current form seems to imply that there's a concave section of the G-G at the negative long accel cases, but we just convex hull over it. There's a huge region of unattainable states, per our MMM, that are being considered for the lapsim as a result. One of these is true:

1. Our MMM generates an incorrect set of states and a bug exists somewhere else
2. GG slices of the GGV are not inherently convex, and using a convexhull is misleading when computing laptimes

Originally posted by @Astronauty in #67 (comment)