Increasing complexity reduces bias but typically increases variance.

Increasing complexity reduces variance but increases bias.

Total mean squared error is the sum of bias squared and variance.

Bias and variance are independent of each other.