I will have a go at answering my question, although I am not certain I have it right. I think the expected RMS displacement between two points undergoing independent random walks in
dimensions after n steps is given by


with a standard deviation of

This seems to agree with the outcome of simulations, but I do not have a good source to back it up.