thoughts on the Mariners, MLB draft, and more homelinksdraftabout me

Payroll Kind Of Matters I Guess

I keep trying to come up with fresh ideas for blog posts. The obvious one for free agency is a grand off-season plan. That would have been fun, and I have my preferences. However, I don't expect the Mariners to do much in the free agent market, and trades are nearly impossible to sniff out. So what's the point?

Instead, I have an overwhelming (but interactive!) data extravaganza. Free agency boils down to money, and in particular the Mariners payroll seems to be a topic of concern most off-seasons. The payroll has sunk about $20 million while revenues across baseball have risen, and it is fair to wonder if the shrinking funds have something to do with the team's struggles. Below is a visualization tool so you can explore just about anything you want relating team payrolls to team victory totals from 1997 through 2012. I would say more, but I still am yet to play with all the options:

Bottom line: payroll does not seem to matter as much as we might intuitively think. There is a connection between payroll and success, but not much of one in baseball.

I crunched R^2 values for every season from 1997 to 2012 looking at payroll versus victories. The reason I prefer R^2 is because, by definition, it represents the percent of variation of a dependent variable explained by the independent. In this case, I treated payroll as the independent variable, and win totals as the dependent. This differs greatly from correlation because correlation only measures the strength of linear relationships. Correlation does not equal causation.* R^2, on the other hand, does imply causation.

*Cannot stress this enough; it is a common misconception.

Here are the R^2 values, year-by-year, expressed as percentages:
  • 2012: 4%
  • 2011: 17%
  • 2010: 12%
  • 2009: 14%
  • 2008: 11%
  • 2007: 24%
  • 2006: 29%
  • 2005: 23%
  • 2004: 29%
  • 2003: 18%
  • 2002: 20%
  • 2001: 10%
  • 2000: 10%
  • 1999: 50%
  • 1998: 47%
  • 1997: 22%
There is no special reason I stopped at 1997. Salary data is available back to 1998, courtesy USA Today, and I plan on adding all that data in when I get time.

I am not still not entirely sure what to make of all the data, other than that it seems to strongly support that things beyond team payrolls impact success. I plan to say more as I digest all this information, and get all the data from 1988-1996 entered. I can provide a list of my own wonderings though:
  1. I wonder if the sinking R^2 values are actually the product of a non-linear relationship between payrolls and victories. It's only in the last decade that payroll gaps have become so massive in baseball, and intuitively my understanding of supply and demand would suggest that premium free agents cost exponentially more to sign.
  2. I wonder if expansion impacts how much money matters. The two highest R^2 values by far in the data I've crunched are 1998 and 1999 - the first and second years of the Diamondback and Rays. Going farther back will at least give a glimpse at when the Marlins and Rockies entered the league.
  3. I wonder how much of the variation is due to "luck." The Orioles should have had a losing record this year with their run differential, yet ended up one game short of the ALCS.
  4. I wonder what else matters, potentially more than team payroll, when it comes to winning baseball games. I would imagine a team's farm system plays a major role, but it would be hard to statistically measure that. It definitely would not be as simple as payroll.
This is just some food for thought (and data for your own exploration) as free agency begins, particularly with the Mariners poised to potentially spend little. I still hope they add some offensive firepower, but I won't despair if they doesn't change much.