Endpoint distribution for neighbor avoiding walks

In a recent paper, Ishinabe and Chikahisa enumerated neighbor avoiding walks, and showed that analysis of this data by a standard method led to an exponent value δ=5, in disagreement with the value δ5=4 predicted by scaling. The purpose of this note is to extend the walk length, and to make some observations on the difficulties of the analysis of this problem. Even with greatly extended walk length, it is difficult to determine the value of δ for this problem by fitting the endpoint distribution.