Saturday, April 21, 2012

April means little when it comes to forecasting Phils' potential success

PBR - Fans often contend a team's success hinges on how it performs during the first month of the season. 

Though logical, this assertion often is inaccurate.

A recent study conducted by the Review of Phillies' monthly winning percentages (dating from 2000 to 2011) indicates a minor correlation between April success (or lack thereof)  and end-of-season winning percentage.

Instead, the study indicates performances in the months May and June are more telling than April.

To begin the study, I found Phillies' month-by-month records from 2000 to present on Baseball Reference and placed them in Excel. I then ran a multiple regression analysis with the Y variable equaling season winning percentage and the X variables equaling the various monthly winning percentages. I did not include the instances where games were played in March or October.

PHI monthly winning percentages, 2000 - 2011
Using the information above, the regression analysis resulted in the following coefficients (all statistically significant):
  • April: .127
  • May: .244
  • June: .219
  • July: .143
  • August: .175
  • September: .178 
The higher the number, the more closely that month is linked to the end-of-season winning percentage.
To further examine the importance (or lack thereof) April has in forecasting success, I looked at every club dating to 2000 that won the World Series. I made note of their April winning percentage and subtracted April's won-loss record to find overall season winning percentage. I did the same with the Phillies and charted the following (* means March games added in April, October games added in September):

Phillies WPct in April and End of Season
World Series Winners in April and End of Season.
The correlation for the Phillies was .652, meaning the winning percentages (April and non-April) were moderately linked, but not significantly. In contrast, the correlation of April winning percentages and overall winning percentages of World Series winners from 2000 to present was -.470, meaning April performances are lousy indicators of future success.

So, what does all of this mean?

Put simply, you can't read too much into April performances. Sure, winning early is a great opportunity to build a lead in a division, but the season is a marathon, not a sprint. 

Sustained success in the middle months of the season is much more meaningful than a solid winning percentage in April.

0 comments: