Unbalanced Schedules, Park Effects and Performance Deflators (Main Paper)

ABSTRACT:  both major leagues have adopted radically unbalanced schedules for the 2001 season, in which teams will play
up to 20 games against divisional opponents, but as few as 6 against most non-division league opponents.  This paper shows
that these unbalanced schedules are likely to create new headaches for analysts trying to adjust raw performance for context.
Because teams' road schedules overrepresent stadiums within one's own division, conventional measures of park effects are
likely to be misleading, especially if some divisions have better hitters parks than others.  Meanwhile, unbalanced schedules will
also create inequities in the quality of competition faced by players on different teams; Baltimore hitters, for instance, will have to
face the Yankee and Red Sox pitching staffs 19 times each in 2001, while getting to face Ranger pitching only 6 times.  The
main paper discusses the problem at an intuitive level and shows that the gaps between conventional estimates and reality may be
quite large in 2001 and beyond.  The technical appendix derives a new method for computing park factors and other performance
adjustments that is valid under unbalanced schedules.  The supplemental table presents results from applying my method to
unbalanced schedule seasons in both leagues since 1969.