SABR 35 - September Falloff Revisited (Research Presentation)
|
|
| By The SABR Office |
September Falloff Revisited
Andy Andres
King II
Previous research by Harold Brooks described the September performance declines for batters. He states batting average (-.002), on base average (-.002), and slugging average (-.006) decline after the end of August. Brooks' research was based on data from the eight seasons, 1984-91. He also determined that playing time had an impact on decline – increased playing time in September (those players who rested four or fewer games) led to greater performance drop offs. This performance decline was also greatest in throwing infielders (2B, SS, 3B).
This is an updated look at Brooks' analysis. Data from the seasons 2000 through 2003 are used. For each season, the data representing April though August of each of these seasons has been gathered, approximately the first 22 weeks of the 26-week MLB season.
The last four weeks of the season (called the September data here) has been determined by the difference between the last April through August data and the final season statistics. The data from September has been compared to the April-August data. Since regularly playing batters were wanted for this analysis, only players who had a minimum of 50 AB in September and 275 AB (approximately 12.5 ABs per week) in April through August were used.
The results are interesting. The combined 2000-03 data supports the results of Brooks from 1984-91. There does seem to be some impact of September falloff compared to the earlier months of the season. In addition, it looks as if the year 2000 is somehow different than the others, an interesting observation that proposes more questions about the nature of that season. However, in contrast to Brooks' results, we do not see the same positional differences in September Falloff. Here, the positions most negatively impacted by September seem to be 1B/DH, SS, C, and RF.
Many questions may be asked about this: How can the data from September 2000 be explained, was that season somehow different? Are the differences seen in September, as shown by Brooks and here, significant? What is the impact on run production that OBP falls 2 points in September? What is the true impact of position on the September falloff? How can we better predict the September falloff? Is it true, as Brooks suggests, that rest will improve the falloff?
ANDY ANDRES earned his PhD in 1999 and currently teaches at Tufts University, offering the course “The Analysis of Baseball: Statistics and Sabermetrics.”
|
|
| | Created On: 2005-07-12 |
|
|
|
|
|
Page Link: http://www.sabr.org/sabr.cfm?a=cms,c,1345,34,0
|
|
|
 |
|
