An estimator summarizes a [[parameter]] of a distribution given some observed data. Estimators differ from estimates in that estimates take on a single value; estimators are in fact [[random variable|random variables]]. Estimators are often denoted with a "hat" or line over the top. For example, $\hat{\theta}$ is an estimator of the generic parameter $\theta$. An unbiased estimator is one where the [[expectation|expected value]] of the estimator equals the parameter of interest. However, note that a biased estimator with smaller variance may beat an unbiased estimator depending on the task at hand.