Friday, October 21, 2011

One-Run Wins Boosted Tigers in 2011

Before the Tigers played the Rangers in the American League Championship series, many noticed a wide disparity in run differential between the two teams in 2011.  While the Tigers scored only 76 more runs than they allowed (787 versus 711), the Rangers outscored opponents by 178 runs (855 versus 677).  Even with that apparent disadvantage, the Tigers won just one less game than the Rangers during the season  In this post, I'll explore run differential and wins in more detail.

Pete Palmer (Hidden Game of Baseball, 1984) estimated that a differential of ten runs is worth approximately one win.  In other words, adding ten runs to a team's season total increases their expected wins (EW) by one win and subtracting ten runs decreases their expected wins by one.  Palmer developed the following formula for expected wins based on that principle:

EW= games/2 + (run differential)/10

According to their +76 run differential, the Tigers had 81 + 76/10 = 89 expected wins.  In actuality, the Tigers won 95 games. So, they outperformed their predicted wins by +6.  The Rangers, on the other hand, had 96 actual wins and 99 expected wins, a difference of -3.     The table below shows that the Tigers outdid their predicted wins by a bigger margin than any team in the American League.

Table 1: Wins Versus Expected Wins for American League Teams, 2011

TEAM
RS
RA
RS-RA
W
EW
W-EW
DET
787
711
76
95
89
6
CLE
704
760
-56
80
75
5
BAL
708
860
-152
69
66
3
CHW
654
706
-52
79
76
3
LAA
667
633
34
86
84
2
TOR
743
761
-18
81
79
2
TBR
707
614
93
91
90
1
MIN
619
804
-185
63
63
0
SEA
556
675
-119
67
69
-2
TEX
855
677
178
96
99
-3
OAK
645
679
-34
74
78
-4
BOS
875
737
138
90
95
-5
NYY
867
657
210
97
102
-5
KCR
730
762
-32
71
78
-7

Data from: Baseball-Reference.com

When a team performs far above its expected wins, as the Tigers did, it usually means that they either did very well in close games or they lost a disproportionate number of games by wide margins.  In the Tigers case, the gap between wins and expected wins can be explained by the league-best 29-17 record in one-run games. If they had instead played .500 ball in those games, they would have won six fewer games.

Interestingly, the Tigers were a league-worst 16-26 in one-run games in 2010.  What a difference year makes.  Why were they so much better in close games this year?  It could be due to better luck, managing, clutch hitting or relief pitching.  I'm guessing that Jose Valverde's perfect record in save situations was responsible for part of it.

The relief pitching theory might be supported by researchers Rany Jazayerli and Keith Woolner, who studied the relationship between bullpens and wins versus expected wins. They found that teams with the strongest bullpens tended to outperform their expected wins, while teams with weaker bullpens tended to under perform.  Their work was published at Baseball Prospectus in 1999. 

What does all this mean for 2012?  First, it's hard to perform so well in on-run games two years in a row, so we would expect the Tigers to win fewer games next year with the same run production and prevention.  Thus, it's important that they either improve internally, make some significant off-season upgrades or a combination of both.

The good news is that no team in the AL Central had as many as 80 expected wins in 2011.  Somewhat surprisingly, the Royals finished second in the division with 78 expected wins.  They under performed their expected wins by seven which was more than any team in the AL.  Part of the reason was a league-worst 34-46 record in games decided by two runs or fewer. A better record in close games combined with a developing core of young players could make them a divisional threat in the future.  For now though, I don't think their pitching is ready for contention.

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