Modified Walk Moves

[n00613]

When using the Walk Move, I had troubles with nonnegative parameter space where the probabilities kinda look like a lognormal distribution. Walkers that are far from the modal area are having trouble getting closer. To go around with this used a modified walker from this article. where the walkaround the problem is having a scaling constant on the step size. However, is this modification still Affine Invariant?