Answers are grounded in this paper and cite their sources.
Using the analogy between the statistics of the levels of quantum Hamiltonians and the eigenvalues of random matrices the authors use (an appropriate choice of) the latter in order to study the transition region near integrability. They show that the nearest-neighbour spacing distribution is linear for small spacings while the inverse of its slope is proportional to the amplitude of the (integrability-destroying) perturbation.
No takes yet. Share an insight, caveat, or question.
Caurier et al. (1990) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: