For many examples of couples $(μ,ν)$ of probability measures on the real line in the convex order, we observe numerically that the Hobson and Neuberger martingale coupling, which maximizes for $ρ=1$ the integral of $|y-x|^ρ$ with respect to any martingale coupling between $μ$ and $ν$, is still a maximizer for $ρ\in(0,2)$ and a minimizer for $ρ>2$. We investigate the theoretical validity of this numerical observation and give rather restrictive sufficient conditions for the property to hold. We also exhibit couples $(μ,ν)$ such that it does not hold. The support of the Hobson and Neuberger coupling is known to satisfy some monotonicity property which we call non-decreasing. We check that the non-decreasing property is preserved for maximizers when $ρ\in(0,1]$. In general, there exist distinct non-decreasing martingale couplings, and we find some decomposition of $ν$ which is in one-to-one correspondence with martingale couplings non-decreasing in a generalized sense.