In MSE, the as the bias goes up, the variance goes down, and vice-versa. A good MSE will try to minimize both parameters.