Example showing a run for which we randomly and uniformly initialised 32 walkers across the full allowed space 0 ≤ Π ≤ 100, 0.0 ≤ q_m ≤ 6.0, −10.0 ≤ q_s ≤ 10, and −5 ≤ γ ≤ 15. The walkers converged towards the same q_m ~ 2.9 as in the longer run, but the model evaluation times for some points far from the good solutions can be exceedingly long, making it computationally difficult to fully evaluate these chains. Because of this, we prefer to start walkers near our default solution.