Suppose that you have two biased estimators of the same population parameter. Estimator A has a bias equal to 1/n (that is, the mean of the estimator is 1/n unit away from the parameter it estimates), where n is the sample size used.
Estimator B has a bias equal to 0.01 (the mean of the estimator is 0.01 unit away from the parameter of interest). Under what conditions is estimator A better than B?