Does anyone know if there's a stronger and more general bound than math.dartmouth.edu/~carlp/PDF/paper88.pdf for Miller-Rabin for integers chosen uniformly at random? Or is that still the best one known?