General periodicity of the complex cubic sin²-type algorithm
Prove that the deterministic sin²-type algorithm for complex cubic fields produces an eventually periodic digit sequence for every admissible trajectory, thereby establishing a general periodicity theorem for the non-totally-real case of Hermite’s problem.
References
The general periodicity problem remains open.
We state the phenomenon as a conjecture. What the data supports is invariance for the algorithm as fixed here --- polynomial p, lattice \mathbb{Z}[\alpha], sorting and tie-breaking conventions included:
Unit-exponent spread is the best single structural discriminant of reachable versus out-window admissible states (AUC 0.95), though no exact invariant characterizing reachability was found; we state its identification as an open problem in \S8.
OP8, sharpened: prove or refute that the terminal cycle depends only on the isomorphism class of (\mathbb{Z}[\alpha],\mathrm{conventions}).
Which pairs of unimodular moves are score-equal on which states, and why?