Wednesday, May 27, 2009

More on Blown Saves

Eight days ago the FanDome featured a post concerning the dramatic rise in blown saves. Time was not taken to see if the blown saves were tied to an increase in appearances per game for pitchers. But without the math, a theory was espoused that the two must be tied together. Yesterday, the Fan commented on another blog about the blown saves and another reader commented that there were more blown saves because more pitchers were pitching per game. In other words, this other baseball fan had the same theory proposed here. But does the theory hold water? Well, the Fan had to stop being lazy and went to FanGraphs.com and got out Microsoft Excel to figure that out.

To refresh our memory, this is what the Fan wrote last week concerning the rise in blown saves:

Let's look at the past five years:

2005 - American League Save Percentage: 68%, National League: 68%. There were only two teams in the entire major leagues that had a save percentage less than 60%

2006 - American League Save Percentage: 68%, National League: 64%. There were five teams that had a save percentage less than 60%.

2007 - American League Save Percentage: 68%, National League: 67%. Again, there were five teams that were successful less than 60% of the time.

2008 - American League Save Percentage: 67%, National League: 62%. The number of teams that were successful less than 60% of the time jumped to eight.

2009 - American League Save Percentage: 63%, National League: 61%. An amazing 14 teams are currently under 60%. And four of those teams have been successful less than 50% of the time: Washington (37%!!), Houston (41%), Cleveland (47%) and Minnesota (46%).

To recap, if we look at that five year pattern, 2005 through 2007 were relatively stable. 2008 saw a plummet of success rate for the National League. 2009 shows a dramatic plummet in both leagues. So if our theory is correct, then 2008 and 2009 should show an increase in pitcher appearances for each year. The numbers do not support the theory. Here's what they look like:

  • 2009 - 3.94 pitcher appearances per game. 9.88 batters per pitcher appearance. The latter is the average per game including the starter and the relievers. So a starter could face 27 batters and the reliever 3, so for that game, the average batters per appearance would be 15 (30 batters, two appearances).
  • 2008 - 3.92 pitcher appearances per game. 9.87 batters per pitcher appearance. Not much difference.
  • 2007 - 3.97 pitcher appearances per game. 9.78 batters per pitcher appearance.
  • 2006 - 3.85 pitcher appearances per game. 10.06 batters per pitcher appearance.
  • 2005 - 3.71 pitcher appearances per game. 10.31 batters per pitcher appearance.
  • 1999 (for historical perspective) - 3.56 pitcher appearances per game. 10.98 batters per pitcher appearance.
  • 1989 - 2.88 pitcher appearances per game. 13.22 batters per pitcher appearance.
So what have we got here? Well, we can clearly see that the game has changed since 1989! But the game has been static in these two specific categories since 2007. There was a dramatic increase in appearances from 2005 to 2006 and then from 2006 to 2007. But since then, the staff usage remains the same for the past three years. So since only the last two years have seen a significant rise in blown saves, then the amount of appearances shouldn't be a factor or cause to the effect.

So what else explains it? We can pretty much rule out PEDs because the failure rate has only increased in the last two years and 2004 is the year you would look at for those kinds of number anomalies.

One possibility might be the number of walks, which has risen every year since 2005. Perhaps walks are reaching critical mass with more teams focusing on OBP and working deep into the count. Here are the walks per game since 2005:

2009 - 3.65
2008 - 3.36
2007 - 3.31
2006 - 3.26
2005 - 3.13

That 2009 number is quite a big leap. While it doesn't seem to be that much of a difference statistically, if the current pace continues, there will be 1400 more walks this year than in 2008. That's a lot of base runners a relief staff has to cope with.

The Fan doubts this is the final call on the blown save department, but we have at least trashed one original theory and have another in its place.