I think I have a pretty good understanding about building and tuning queries, and the use and creating of indexes.  What I am not sure about is statistics.  In working on slow performance of our PeopleSoft Finance system after upgrading to SQL Server 2005, I ran an hour long trace and used that with DTA.  It recommended we build 35 indexes and a lot more multi-column statistics.  We have implemented those recommendations and have seen a significant improvement in performance.  I can understand how the indexes help, but how do the statistics?  If you are doing this manually, how do you decide if a statistic is more appropriate than index?  Without going into a 3 to 30 page explaination, how are statistics used to improve query performance?

Or should I invest in the book I mentioned earlier by Karen Delaney?

Also, I work for a school district, training funds are very sparse.