I am looking for an intuitive explanation for the bias-variance tradeoff, both in general and specifically in the context of linear regression. Thanks.