Abstract
The rapid developments in the field of internet search engines underline the need for reliable mechanisms for performance evaluation. So far, the vast majority of researchers have relied on the "precision" and "recall" measures known from the field of Information Retrieval Unfortunately, both of them fail to assess how successfully they rank the returned documents according to their relevance. In this paper, we discuss this issue in some detail, and then propose a new mechanism for the evaluation of the quality of search-engine rankings.