In Chapter 2, we talked about the query optimization process and how it uses the indexes, constraints, and other objects within your database in order to figure out how best to satisfy the query. All those objects in combination are how the optimizer arrives at its estimated values. However, one of the single biggest driving factors in what leads the optimizer to choose one method of query behavior over another is row counts. Since SQL Server can’t realistically count all the rows ahead of time, it maintains information about the distribution of data within a column. That information is called statistics. Statistics are used to establish estimates on row counts for the optimizer so that it can make better choices.