Unrolling equations to transitions can be unreliable

[j-c-gibson]

I think the unrolling of equations into transitions could be reframed in terms of a linear algebra problem, hopefully giving more reliable outputs.