Introducing Salary+

A few days ago, we found that that the relationship between wins and salary among MLB teams in 2014 was not that strong. Today we’re going to see if that was a fluke or the norm.

Salaries Over the Years

Graphs get bigger with one click!

Because the average team salary has increased by about $40 million over the past decade (and $80 million over the last two), we can’t compare wins to money spent straight up. To address this I have created Salary+, making the league-average salary for each season is 100. If you’re familiar with stats like OPS+ and ERA- you probably get it, skip the next paragraph. If not:

Like salary, most stats change over time. The average hitter in 2000 had a .345 OBP, but the average hitter in 1968 had a .299 OBP. In order to better compare players from those years, we take the player’s OBP, divide it by the league average for that season, and multiply it by 100. So if your OBP+ is 100 you were an average hitter for that year. If you’re above 100 you were above average, if you are below 100, below average. That goes for any stat you see a + after. Conversely we have ERA- because the lower the better, so an ERA- of 100 is average, but 60 is awesome.

Salary+ keeps things proportional, so the R-squared value is the same for each season whether you use the actual dollar value or the Salary+ value. In 2014, the R-squared for salary compared to wins was 0.087, which is not all that strong of a relationship. If we expand our pool back to 1988 (that’s all USA Today had in its database), that gives us 724 team seasons and an R-squared of 0.163. So this season was below average, but clearly there is more to a winning team than spending a lot of money. I left out the 1994 and 1995 as they were shortened due to the lockout.

Salaries vs Wins, R-sqared

I wanted to see how much the relationship fluctuated from season-to-season, so I found the R-squared of each season and plotted them out. In the early 1990s there was virtually no relationship, but it rose sharply toward the end of the decade.

Yearly Salary R-squared

Part of that I would credit to the increasing range of salaries, which doubled from 1992, when the R-squared was at its lowest, to 1999, when it was at its peak. This would seem to lend credence to the thinking that the Yankees (who won the AL from 1998- 2001) bought their way to the top. Since 1988 the Yanks’ R-squared for salary and wins has been stronger than the League-average at .295.

Over the past decade though, the relationship between salary and wins has again been on the decline. Teams with smaller budgets have been able to compete by minimizing their cost per win. Oakland has managed to average 88 wins on an average Salary+ of 66 since hiring Billy Beane. Still, in 2014, five playoff teams in 2014 came from the League’s top third, three from the middle, and two from the bottom. Spending money helps, but it does not guarantee anything: Four of the five highest salaries did not make the playoffs in 2014.

While we’re at it, here are a few more graphs using Salary+. Everyone knows the Yankees like to spend a lot of money, but I never realized it was this much.

Highest Salary+ Teams

And in the interest of symmetry, here are the lowest spenders.

Lowest Salary+ Teams