Bias
Or the average of the error.
A random variable fluctuates around his expectations. We can therefore wish that the expectation of \(\hat{\theta}\) is equal to \(\theta\) , in other words, that in "average," the estimator is not mistaken.
The bias is defined as follows:
\(Biais\left(\hat{\theta}\right)\equiv b\left(\hat{\theta}\right)=E\left[\hat{\theta}\right]-\theta\) (1)