Saturday, October 11, 2008

Tigers offense sabermetrically correct again

This year, many fans complained that the Tigers left too many runners on base and did not get the most out of their offense. They said the same thing about last year's team. In actuality, they scored almost exactly as many runs as they should have given their totals of hits, walks and various other positive batting events in 2007. Today, we'll take a look at 2008.

The Tigers finished fourth in the league with 821 runs scored in 2008.
Should the Tigers have scored more runs from their offensive output? One way to answer this question is to look at runs created (RC). RC is calculated from the number of walks, singles, doubles, triples, home runs and other things that a team does to score runs. Bill James showed in his Baseball Abstracts many years ago that RC is highly correlated with runs scored. This tells us that runs scored are essentially the sum of other team statistics and are not, in general, accounted for by unmeasured factors. There are many RC formulas but they all have the same basic structure:

RC= (A*B)/C

where A=number of runners on base, B=advancement of runners and C=the opportunities to do A and B.

One of the first runs created formulas was invented by Bill James in 1979. It was simply:

RC=((hits+walks)*total bases)/(ab+walks).

This version is still useful for some purposes but the formula has evolved over the years and more recent versions are more accurate in estimating runs scored. Sabermetrician Dan Fox gave a brief history of runs created on his blog. The formula I am using here is:

RC= (((2.4 *C + A) * (3 * C+B))/(9*C))-(.9*C)

where

*= multiplication
A = H + BB + HBP - CS - GIDP
B = 1.125*b1 + 1.69*b2 + 3.02*b3 + 3.73*HR + .29*(BB-IBB+HBP) + .492*(SB+SH+SF)-.04*so
C = AB+BB+HBP+SH+SF

This looks really complex but it does not give hugely different results from the earlier more simple formulas. So if you want to tune out this version of formula and just think of the (A*B)/C thing, you should be OK.

Table 1 lists the runs, RC, difference between runs and RC and % difference for all American League teams in 2008. Table 2 does the same for National League teams. The first thing you should notice is that all but 3 teams in the majors had RC estimates within 5% of their actual runs scored. This means that runs created gives a good estimation of runs scored in most cases.

A closer look at table 1 shows that the Tigers had 831 runs created in 2008. This is just 10 runs more than their actual runs scored total. So, they scored almost as many runs as they would have been expected (just 1.2% fewer) given their offensive output. In fact, there were five AL teams and seven NL teams that underperformed their runs created by more than the Tigers. That tells us that the Tigers were neither efficient nor inefficient with their offense.

The team which got the biggest bang out of their offensive output was the Minnesota Twins who scored 46 (or 6%) more runs than their runs created estimate. This is because of their abnormally high .305 batting average with runners in scoring position. The least efficient team was the Boston Red Sox who scored 41 (4.6%) runs fewer than their runs created.

In the National League, no team was especially efficient. The San Diego Padres (who were the most efficient team in the league last year) scored 41 (6.3%) runs fewer than they created than they created. The least efficient or unluckiest team in all of baseball was the cardinals who scored 61 (7.3%) fewer runs than their runs created estimate.


Table 1: Efficiency of American League Offenses in 2008

Team

Runs

Runs Created

difference

% difference

MIN

829

783

46

5.9

LAA

765

742

23

3.1

CHA

811

792

19

2.4

CLE

805

788

17

2.2

OAK

646

643

3

0.4

KC

691

689

2

0.2

BAL

782

785

-3

-0.3

TOR

714

722

-8

-1.1

DET

821

831

-10

-1.2

NYA

789

801

-12

-1.5

TEX

901

919

-18

-2.0

SEA

671

686

-15

-2.2

TB

774

803

-29

-3.6

BOS

845

886

-41

-4.6


Table 2: Efficiency of National League Offenses in 2008

Team

Runs

Runs Created

difference

% difference

PIT

735

722

13

1.9

FLA

770

767

3

0.4

HOU

712

714

-2

-0.3

PHI

799

803

-4

-0.5

CIN

704

709

-5

-0.8

CHN

855

864

-9

-1.1

NYN

799

809

-10

-1.2

ARI

720

729

-9

-1.2

WAS

641

652

-11

-1.6

MIL

750

770

-20

-2.6

LAN

700

722

-22

-3.1

SF

640

667

-27

-4.0

ATL

753

784

-31

-4.0

COL

747

778

-31

-4.0

SD

637

680

-43

-6.3

STL

779

840

-61

-7.3

2 comments:

  1. Does the National League typical skew towards performing under expected runs? (A URL would be enough- I don't mind doing my own homework.)

    If the formula used does not work well for particularly weak hitters I could see this because of the pitchers batting.

    ReplyDelete
  2. Jeff,

    It's been that way the past few years but I went further back and there doesn't seem to be a league trend. If you go back to periods where there was not much hitting (like the late 60s), teams in both leagues seem to skew towards over performing expected runs.

    Lee

    ReplyDelete

Twitter

Blog Archive

Subscribe

My Sabermetrics Book

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

Other Sabermetrics Books

Stat Counter