Saturday, December 03, 2011

Shutdowns and Meltdowns

Tigers closer Jose Valverde had a spectacular season closing games in 2011 accumulating 49 saves in 49 opportunities.  His ability to close out games was so impressive that he finished fifth in the American League Cy Young balloting.  Was he really that good though?

There are a few limitations of the Saves statistic in evaluating closers.  First, not all save situations are especially challenging.  For example, if a team has a three-run lead with nobody out and the bases empty in the ninth inning, there is about a 98% expectancy that the team will win the game with an average reliever.  So getting a save in that scenario is not necessarily a big deal, yet closers are routinely used in those situations and pile up rather easy saves.

An additional issue is that closers are often brought into games where there is no save opportunity.  They may come into a tie game and can have either a strong positive or negative impact on the outcome, but they get no credit for either a save or a blown save.  Valverde was particularly bad in non-save situations last year, but none of those games went into the Blown Saves column.  Moreover, many set-up men contribute substantially to their team's chances of winning games, but register no saves at all because are never called upon to do so.
      
What would be useful are some numbers which tell us how often a reliever added significantly to his team's chance of winning games and how often he significantly hurt his team's chances. Two relatively new statistics at FanGraphs - Shutdowns (SD) and Meltdowns (MD) do just that.    
 
The SD and MD statistics are based on the concept of Win Probability Added (WPA).  The WPA system gives relievers credit based on the effect each batter faced has has on the team's probability of winning.  These probabilities vary depending on the game state before and after each play.

Suppose Tigers reliever Joaquin Benoit comes into the game in the top of the eighth with a two run lead, no outs and a runner on first. There is an 83.7% expectancy that an average team will win a game given that situation. Now, assume that Benoit strikes out the first batter. There is now one out and the probability of winning increases to 88.4%. Subtracting the win expectation before the strikeout from the win expectation after the strikeout gives us the value of the play in terms of the probability of winning added by the strikeout, that is, 4.7% (88.4% - 83.7%).

Now, suppose the next batter, facing Benoit, doubles home a run. That gives the Tigers a one run lead with a runner on second and one out. The probability of winning goes down to 77.4% so Benoit loses 11%  (77.4 – 88.4) on that batter.  Summing all the gains and subtracting all the losses for all the batters Benoit faces during a game yields his WPA for that game

An SD is defined as any game where a pitcher had a WPA of 6% or more, that is, he greatly helped his team's chances of winning the game.  Similarly, an MD is any game where a pitcher has a WPA of -6% or less, that is, he significantly hampered his team's chances.  The 6% cutoff was chosen so that SD would be on a similar scale as saves.  For example, 40 SDs for a pitcher is about as rare as 40 saves is for a pitcher.  The SDs and MDs of all pitchers can be found on FanGraphs.

The SD and MD metrics are counting statistics which don't take numbers of opportunities into consideration.  So, the following rate statistics are also useful:

(1) The Shutdown Percentage (SD%) for a pitcher is the percentage of total games pitched which resulted in an SD.

(2) The Meltdown Percentage (MD%) is the percentage of games which resulted in a MD.

(3) Shutdown Meltdown Ratio (SD/MD) is Shutdowns divided by Meltdowns.

The AL Leaders for SD% are shown in Table 1 below. You can see that the best relievers had SD%'s above 50% and that Indians reliever Chris Perez and Yankees standout Mariano Rivera tied for the league lead at 57.8%.  Valverde was tenth in the league at 51.7%.  Benoit was 15th with 43.9%. 

Table 1: AL Shutdown Percentage Leaders, 2011


Player
Team
G
SD
SD%
Chris Perez
CLE
64
37
57.8
Mariano Rivera
NYY
64
37
57.8
Rafael Soriano
NYY
42
24
57.1
David Robertson
NYY
70
37
52.9
Andrew Bailey
OAK
42
22
52.4
Greg Holland
KCR
46
24
52.2
Joakim Soria
KCR
60
31
51.7
Grant Balfour
OAK
62
32
51.6
Jonathan Papelbon
BOS
63
32
50.8
Jose Valverde
DET
75
38
50.7
Jordan Walden
LAA
62
31
50.0
Daniel Bard
BOS
70
35
50.0
Chris Sale
CHW
58
28
48.3
Jim Johnson
BAL
69
31
44.9
Joaquin Benoit
DET
66
29
43.9

 Table 2 ranks AL pitchers according to fewest meltdowns.  Greg Holland of the Royals had just one meltdown all season giving him an SD% of 2.2%.  Valverde was fourth in that category at 6.7%.  Somewhat surprisingly, Tigers lefty Daniel Schlereth finished 11th with 10.2%.  That might have something to do with how he was used, but the fact that he had so few meltdowns is still a positive.      


Table 2: AL Meltdown Percentage Leaders, 2011


Player
Team
G
MD
MD%
Greg Holland
KCR
46
1
2.2
Frank Herrmann
CLE
40
2
5.0
Jonathan Papelbon
BOS
63
4
6.3
Jose Valverde
DET
75
5
6.7
Casey Janssen
TOR
55
4
7.3
Koji Uehara
AL
65
5
7.7
Jason Frasor
AL
64
5
7.8
Grant Balfour
OAK
62
6
9.7
Brandon Gomes
TBR
40
4
10.0
Cesar Ramos
TBR
59
6
10.2
Daniel Schlereth
DET
49
5
10.2
Dan Wheeler
BOS
47
5
10.6
Mike Gonzalez
AL
56
6
10.7
Mariano Rivera
NYY
64
7
10.9
Neftali Feliz
TEX
64
7
10.9

Finally, SDs and MDs are combined in Table 3.  Holland was the run away leader with an SD/MD Ratio of 24.0 (24 SD, 1 MD).  Valverde came in third at 7.6.  While his season may not have been quite as amazing as his perfect save percentage suggests, this final number confirms that Papa Grande was indeed one of the best at consistently shutting down opponents while keeping meltdowns to a minimum.  

Table 3: AL Shutdown Meltdown Ratio Leaders, 2011


Player
Team
G
SD
MD
SD/MD
Greg Holland
KCR
46
24
1
24.0
Jonathan Papelbon
BOS
63
32
4
8.0
Jose Valverde
DET
75
38
5
7.6
Grant Balfour
OAK
62
32
6
5.3
Mariano Rivera
NYY
64
37
7
5.3
David Robertson
NYY
70
37
8
4.6
Koji Uehara
AL
65
23
5
4.6
Casey Janssen
TOR
55
18
4
4.5
Andrew Bailey
OAK
42
22
5
4.4
Chris Sale
CHW
58
28
7
4.0
Joakim Soria
KCR
60
31
8
3.9
Jason Frasor
AL
64
19
5
3.8
Chris Perez
CLE
64
37
10
3.7
Vinnie Pestano
CLE
68
29
8
3.6
Neftali Feliz
TEX
64
25
7
3.6

No comments:

Post a Comment

Sabermetrics Book

Sabermetrics Book
One of Baseball America's top ten books of 2010

Blog Archive

Subscribe

501 Baseball Books

501 Baseball Books
Recommended by Tiger Tales

Stat Counter

Site Meter