I think it's not unreasonable or uncommon for big O to track separate variables without reducing them, just to highlight the (lack of) sensitivity of different parameters.