Increasing model complexity reduces both bias and variance.

Bias and variance are independent of the sample size.

An estimator with low bias often has high variance, and vice-versa, impacting the total Mean Squared Error.

Bias is always preferable to variance for large sample sizes.