Clutch hitting revisited Pete Palmer and Dick Cramer

Why another study? Persistence of a controversy Bill James argues “Clutchness” is objectively non-existent vs. Human perceptions of pressure Bill James argues Clutchness might be obscured by the “fog” of random variation Might this “fog” be overcome by clever grouping of players? How do pressure situations affect batting generally?

Summary of Our Findings Game situation does not significantly affect average batting performance The “fog” of statistical variation is much thicker than almost anyone appreciates The variation in career “clutchness” among the 897 players with >3000 BFP’s between 1957 and 2007 seems random Ortiz and Mark Grace are tied for ~80th and ~100th rankings among the 897

All ML Hitters Under Pressure OPS when: “Pressure” Definition (who) Tense Other Tensest 10% of BFP (best 897) .779 .771 Elias “late close” 15% (all) .704 .715 Tense Situations are different: More intentional walks Better pitchers More pinch hitters

David Ortiz’s Clutch Performances by Season 2005 Win Value 2006 10 2004 2003 2007 2000 2002 2001 1998 Linear Weight Runs 100

Comparing the “Fog” to the Clutch “Results” Width => Fog Density: calculated several ways (probability theory, simulation). All agree. “Fog” Scott Fletcher “Results” Richard Hidalgo The other 895 players who: had 3000+ plate appearances between 1957 and 2007

“Fog” Density: Starting Points Many “pressure definitions” considered All, weighted by “pressure” 10% highest “pressure”, vs. other 90% Elias “late and close” (15%), vs. other 85% First 100 batting appearances of player Individual AB’s critical => noisy win average Example: Adam Dunn on 6/30/2006 tensest 2% BFP == easiest 35% BFP

Most and least “clutch” players Nellie Fox Don Lock Dave May Pat Meares Minnie Minoso Desi Relaford Jose Uribe Sandy Alomar Earl Williams Damian Miller Mike Lieberthal Michael Barrett Frank Thomas Dick Schofield Chris Sabo Manny Ramirez

Other “pressure” non-effects Clustering of “consecutive seasons”? (e.g., Ortiz 2005-2006). Overall, r2 for “clutchness” = .002 Overall, r2 for OPS = .43 The first 100 BFP’s of a career? Performance distributions by pressure situations compared to performance distributions by game date

Conclusions Yes, the fog that probability theory demands and empirical observation confirms is thick. But why believe in something whose existence can objectively be demonstrated to be unprovable? Especially since a “clutch” hitter must actually be someone who doesn’t always perform at his best!