The systematic analysis of baseball statistics, often called "sabermetrics," has evolved in recent years to resemble something of a science, attracting fans from diverse professional and educational backgrounds, all fascinated by the analysis itself and its insights into the game. But one problem has defied solution: estimating runs saved by fielders throughout history. Traditional statistics include errors and plays made, but not hits that could or should have been prevented. The latter can now be estimated using records of the location of every batted ball, but the underlying data exists only for recent seasons and has generally been withheld from the public. Now, in Wizardry, comes the long-awaited breakthrough. Drawing solely on freely available baseball statistics, Michael A. Humphreys shows how to apply classic statistical methods to estimate runs saved by fielders going back to 1893. Humphreys tests his results against other fielding measures, including published ratings based on proprietary batted ball location data, and explains their respective strengths and limitations. He also introduces a method for adjusting historical player ratings for increased competition due to population growth, integration, and international recruitment. Position by position, Humphreys identifies and profiles the greatest fielders of all time with anecdote-rich essays. Sabermetrics changed baseball and introduced a generation to the art of statistical inference. Wizardry makes the case for the most significant changes in historical player valuation in decades, while opening up new approaches for further exploration.