willing restraint in Models