On Mazur's telescoping argument