It seems as though every year around voting time for the Hall of Fame there is a debate on various baseball mailing lists about the HOF credentials of long-time Pirates' 2B Bill Mazeroski. His proponents like to point out that Mazeroski was considered the best defensive 2B *ever*. His detractors point out that Maz was a terrible hitter (lifetime BA/OBP/SLG of .260/.302/.367) and his superior defensive statistics don't overcome that.
I've done an analysis of Mazeroski's defensive value, estimating the number of runs Mazeroski saved for the Pirates while he played for them. An early version of this analysis - posted here previously with several mathematical errors - estimated Mazeroski's offensive value at 57 runs saved per 162 games. Several commentators (notably Clifford Blau, who posted an analysis on his Web site, here) took issue with that conclusion. I decided to revisit the issue, adjusting Mazeroski's defensive numbers using a method first published in the 1998 Baseball Prospectus to estimate the extent to which Mazeroski's defensive statistics are skewed by the number - and direction - of balls in play (as well as fixing my other mistakes). I wrote a paper summarizing the results of my adjustments, also posted here. Clifford weighed in with some further suggestions for refinements, and I revised the paper in response. I greatly appreciate the amount of time and effort that Clifford put into this review.
The latest (and hopefully final!!) version of the paper is now available in MS Word 7.0 and Adobe Acrobat PDF format (comments are welcome; please send them to mwe@nomos.com).
I conclude that the value of Mazeroski's defense was approximately 22-25 runs per 162 games, rather than the 57 from the original analysis. When combined with the value of his offense, this modified estimate takes Mazeroski out of the realm of slam-dunk HOFers. This is about as complete an analysis as I can make from traditional defensive statistics, absent detailed play-by-play descriptions (maybe someday we'll get them in Retrosheet).