[Gretl-users] Re: confidence intervals in NLS