Time to absorption in dis ounted reinfor ement models