Increasing model complexity reduces both bias and variance simultaneously.

As model complexity increases, bias generally decreases while variance increases, potentially leading to overfitting.

As model complexity increases, bias increases and variance decreases, leading to underfitting.

Bias and variance are independent of the model's complexity.