Re: An MDI regression