Re: [Gretl-users] Multivariate infomation criteria