Double DER U

March 19, 2020
                                                        Double DER U

 

            Defensive Efficiency Record (DER) is an answer to this question:  When a ball is put into play against this team, how often does the defense turn it into an out?   I developed the DER back in the 1970s, the same era when I developed Runs Created and the Pythagorean Method and the Defensive Spectrum and stuff, but it did not reach an audience in the 1980s, did not come into common usage, because I had made the original formula so complicated that it was hard to use.  It was what I used to call a 40-pound garden hoe. 

            DER basically died off, but somebody came along later—I do not know who—and cleaned up the formula, which brought it back to life.  In  measuring DER I had tried to include Errors as plays not made—thus, indications of Defensive Inefficiency.   But since not all errors put runners on base, I then had to estimate how  many of the errors were not relevant, which complicated the process.  Also, my original formula tried to make two different estimates of Balls in Play, one derived from the pitching stats and one derived from the fielding stats, and then put them together, average them out.   The resulting formula was so complicated it was unusable. 

            The formula we use now is just this:

           

            BFP – BB – SO – HBP – HR  =  BIP (Balls in Play)

            DER= (BIP – (H – HR))/ BIP

 

            Over time, DER varies, but does not vary a lot.   In the last 120 years the normal DER has been as high as .729 in the 1950s and 1960s, and as low as .705, in the last decade (2010-2019).   The highest DER ever for a team was .763, by the 1906 Cubs (Tinker to Evers to Chance), and the lowest ever was .660, by the 1930 Phillies.  "Ever" meaning since 1900.    And, since I have come this far in that direction, these are the DER norms for each decade of our study:

 

From

To

Average

1900

1909

.723

1910

1919

.722

1920

1929

.710

1930

1939

.711

1940

1949

.728

1950

1959

.729

1960

1969

.729

1970

1979

.724

1980

1989

.720

1990

1999

.709

2000

2009

.706

2010

2019

.705

 

            And, since it is our habit to do this, these are the standard deviations:

 

From

To

Average

Standard Deviation

1900

1909

.723

.01803

1910

1919

.722

.01792

1920

1929

.710

.01216

1930

1939

.711

.01292

1940

1949

.728

.01171

1950

1959

.729

.01065

1960

1969

.729

.01177

1970

1979

.724

.01159

1980

1989

.720

.01124

1990

1999

.709

.01242

2000

2009

.706

.01050

2010

2019

.705

.01070

           

            There is a remarkable change in the standard deviation of team defense in 1920.  1920, of course, is the great line of demarcation in baseball history—the simultaneous emergence of Babe Ruth, the banning of the spit ball, the Black Sox Scandal, the Commissioner System, and the death of Ray Chapman.   The vortex created by the convergence of these events changed baseball dramatically overnight.  We see here another manifestation of that:  a standardization of defenses, with the standard deviation of DER dropping suddenly, probably due to the disappearance of common bunting. 

            DER is lower now because we are in the throw hard, swing hard era.  More contact is hard contact.   There are more strikeouts, but fewer balls in play result in outs. 

            We are interested, of course, not in DER directly, but in era-normalized DER, normalized by Standard Deviation.   Once we lead our numbers through their dance, the highest DER of all time belongs not to the 1906 Cubs, but to the 2016 Cubs—a World Championship team, as I recall. 

 

 

 

Year

City

Team

DER

2016

Chicago

Cubs

137

2001

Seattle

Mariners

132

1981

Detroit

Tigers

130

1939

New York

Yankees

129

1991

Chicago

White Sox

129

1990

Oakland

A's

128

2011

Tampa Bay

Rays

127

1975

Los Angeles

Dodgers

127

1972

Baltimore

Orioles

127

1968

Cleveland

Indians

126

 

            All ten of those teams had winning records, and four of them won 103 games or more.   In fact, in running down the list of teams with the highest DER ever, era-adjusted, you don’t find a team with a losing record until you get to the 50th spot.   The 1971 California Angels had a losing record (76-86) despite having a Defensive Efficiency Record that was two standard deviations above the norm for the era (120).  If you look at the roster, you can understand it; their lineup was littered with light-hitting defensive wonders like Jim Spencer, Ken Berry and Sandy Alomar Sr., but their 3-4 hitters were guys who should have been hitting 7th and 8th.  For the season, their leadoff hitters had an on base percentage of .295, and their #2 hitters and on base percentage of .285, while their third hitters had a slugging percentage for the season—SLUGGING percentage—of .319, and their cleanup hitters had a slugging percentage of .352.

            So that team had a historically outstanding defense but a historically awful offense, but the 49 teams who had higher normalized DER all had .500 or better records.    DER is more closely connected to winning than anything else we have studied:

   

Wins

Losses

W Pct

Raw DER

Highest DER

87

70

.555

.734

Second Highest

82

75

.521

.723

Average

77

79

.495

.717

Lower DER

75

82

.480

.710

Lowest DER

70

86

.449

.699

 

            I think the conclusion that the 2016 Cubs had a historically great team defense stands up to scrutiny.   Their first baseman, Anthony Rizzo, was named the major league’s best defender at any position—a conclusion that he himself mocked, saying that a first baseman should never win that award—but anyway Rizzo was an outstanding first baseman.  They had not one outstanding defensive shortstop on the team, but two, with Addison Russell and Javy Baez; Kris Bryant was really good at third base, and two outfielders were defensively outstanding, Jayson Heyward and Dexter Fowler.  Ben Zobrist moved here and there and held the defense together whenever and wherever there was an injury.   Their catchers had some defensive issues, but catchers don’t feed very much into DER; what catchers do is important, but it’s not DER.   They had some defensive issues in left field, but then, who doesn’t?

            I have made this point before about the 1906 Cubs—that that team, over a period of years, won more games than any other team ever; they went 116-36 that season, but over a period of ten years and more, they had better won-lost records than any other team, ever.  Tinker, Evers and Chance have to be given a lot of credit for that.   If you judge their pitchers by wins, losses and ERA, you will think that they had fantastic pitching—but if you look again, you’ll realize that (a) their pitcher’s strikeout and walk data is actually not all that good, (b) home runs are a minor element of the game at that time, and (c) almost none of those great pitchers was anything special when he wasn’t pitching for the Cubs.   Three Finger Brown was no doubt legit, but they picked up a lot of guys who were just very ordinary pitchers with other teams, and they suddenly became brilliant pitchers as soon as they put on a Cubs’ uniform.  The defense behind them made their pitchers look a lot better than they really were. 

            Look at the 2016 Cubs, you see something similar.  Jake Arrieta was 22-6, 1.77 ERA in 2015 and 18-8, 3.10 ERA in 2016.  The rest of his career, he’s just not that good.  Kyle Hendricks in 2016 was 16-8, 2.13 ERA.   He’s just not really that good.  The 2016 Cubs’ defense was making pitchers look better than they really were.   They haven’t been able to hold it together in the three years since then, keep it at that level, but I believe that that really was a historic defensive team. 

            DER confuses cause and effect, to a certain extent—cause, effect, and luck.  In other words, it may not be that when you have a great DER you win; it might be, rather, that when you win you have a great DER.  When things break your way, when you have a season when your own line drives go through and the other team’s line drives find somebody’s glove, then you have a high DER, and you win, and you look better in the statistics than you really are.   What I am saying is, you CAN over-value defensive efficiency, but you cannot deny that it is very closely tied to winning.   It’s a chicken-and-egg question; does a high DER cause winning, or does winning cause a high DER?   Some of each, probably.

            The 1930 Philadelphia Phillies had not only the lowest DER of all time, raw, but also the lowest compared to their era.  These are the bottom 10:

 

YEAR

City

Team

DER

1930

Philadelphia

Phillies

61

2007

Tampa Bay

Devil Rays

65

2012

Colorado

Rockies

72

1953

Detroit

Tigers

73

1987

Boston

Red Sox

73

1923

Philadelphia

Phillies

73

1999

Colorado

Rockies

73

1914

Baltimore

Terrapins

73

1997

Oakland

A's

74

1950

St. Louis

Browns

74

 

            The 1930 Phillies are a famously awful team that managed to lose 102 games despite having one outfielder who hit .386 with 40 homers, 170 RBI and another one who hit .383 with 22 homers, 97 RBI.  Obviously it was an atypical league, an atypical park, a kind of crazy combination of circumstances, and both the pitching and the defense were so bad that it is hard to say for sure which was worse. 

            All of those teams with the lowest DER lost 90 or more games except the 1987 Red Sox and the 1914 Baltimore Terrapins.  The 1987 Red Sox had a decent offense and the best pitcher in baseball, Roger Clemens, so they were able to go 78-84 despite their god-awful defense.  78-84 still is not good, but they avoided 90 losses.  The 1914 Terrapins finished 84-70, but that was in the Federal League, an upstart league; the statistical parameters are just a little bit misaligned with the era.   Basically, a low DER means that you lose. 

            The highest DER of all time, normalized for the era, was 3.7 standard deviations above the norm; the lowest was 3.9 standard deviations below the norm. 

            It has always been my belief that the #1 key to a high DER is speed in the outfield.  I can’t really PROVE that; it’s an informal observation, and sometimes it’s hit-and-miss.   I’m not saying that having range in the infield isn’t important; it is.   The 2001 Seattle Mariners had Mike Cameron in center field and Ichiro in right.   They had the second-highest DER ever, era-adjusted, and won 116 games.  The 1939 Yankees had DiMaggio in center field and four other outfielders who could have played center field if they had needed to.  There is no single key to winning, and having speed in the outfield is not the key to winning; you won’t win with fast outfielders unless they put some runs on the scoreboard as well.  But I still believe that having speed in the outfield is the #1 key to a high DER, and a high DER is one of the keys to a successful team. 

            That’s why I am relatively optimistic about the 2020 Red Sox, if we are ever able to get the season started.  With an outfield including Kevin Pillar, JBJ, Benitendi and Alex Verdugo, they’re going to catch everything that can be caught, and this will largely offset the effects of having maybe not the quickest shortstop in the league, or the second-quickest, or the third-quickest.   But we’ll see.  

 
 

COMMENTS (22 Comments, most recent shown first)

raincheck
Re: Baker Bowl. I always suspect that all of our adjustments break down a little in the extremes, such park adjustments for for extreme park are a good example of that.
1:40 PM Mar 24th
 
KaiserD2
OK, so here's Bill's list of the 10 most effective defensive teams in history, measured by normalized DER.

2016 Chicago Cubs 137

2001 Seattle Mariners 132

1981 Detroit Tigers 130

1939 New York Yankees 129

1991 Chicago White Sox 129

1990 Oakland A's 128

2011 Tampa Bay Rays 127

1975 Los Angeles Dodgers

127 1972 Baltimore Orioles 127

1968 Cleveland Indians 126

Now, using the dataset I developed from my book, based on Michael Humphrey's DRA, I'm going to give the team runs saved for each of those teams. (I have this number for every team since 1901.)

2016 Chicago Cubs 137 +89

2001 Seattle Mariners 132 +87

1981 Detroit Tigers 130 +57

1939 New York Yankees 129 +119

1991 Chicago White Sox 129 +93

1990 Oakland A's 128 +100

2011 Tampa Bay Rays 127 +65

1975 Los Angeles Dodgers 127 +70

1972 Baltimore Orioles 127 +40 (+115)

1968 Cleveland Indians 126 +33

My initial reaction is that the two methods are giving much closer results than I had expected. This is NOT a list of the 10 teams with the best DRAs--that would include the 1906 Cubs, the 1939 and 1940 Reds, probably the 1927 Yankees, and the 1948 Indians, I believe. But the top six teams on the list were indeed outstanding by both measurements. (You would have to correct for the strike-shortened season to get the proper comparable figure for the 1981 Tigers.) Regarding the 1972 Orioles, I'd like to ask Bill if that might have been a typo and if he meant 1973. The 1973 Orioles had a team DRA of +115 (all these figures for runs saved are park-adjusted, by the way), which was the highest figure in history.

There are two big differences in the way DER and DRA are computed that might account for the differences between them. Michael Humphreys did not include putouts for infielders, on the grounds that they are mostly from popups (except at first base of course) and thus do not reflect fielders' skill. That means that in my computation, popups are folded into pitchers' ability, not fielders, which I think is right. He also didn't compute pitcher fielding skill and thus pitcher's assists, which can vary a lot from one team to another, play no role in the team DRA calculation and are also counted by me as a measurement of pitching skill.

The 1906 Cubs, the 2001 Mariners, the 1927 Yankees and the 1939 Yankees had four of the highest winning percentages in history. All of them had extraordinary DRA figures in the neighborhood of 100 runs saved. Undoubtedly some luck was involved. Two of the fielding stars of the 1927 Yankees were Earle Coombs and Tony Lazzeri, neither of whom ever had comparable seasons in the field again. Apparently a lot of balls were hit just within their reach. So, this seems to be the only way to win as many games as those teams did.

DK
7:22 AM Mar 24th
 
willibphx
Went back and looked at the 1921 Phillies and this may highlight a potential flaw in your simplification of the DER calculation. By excluding errors as part of the analysis what gets left out is the fact they made 295 errors in 1921. This resulted in 248 unearned runs which I must admit I had to look at three times to believe. This was 119 more than any other team in the league. The average team (other than the Phillies) allowed 110 unearned runs. Interestingly the Dodgers and Cardinals made 232 and 219 errors respectively while allowing 121 and 129 unearned runs. So yes the 1921 Phillies did have a really poor defensive team but it may not be reflected in the simplified DER calculation.​
10:23 PM Mar 21st
 
willibphx
While I agree that the Baker Bowl was an extreme situation which makes analysis difficult are you saying that DER is not materially effected by park effects? I went back and looked at multiple years for the Phillies. The difference in their DER on the road as compared to home was:

1930 .025
1929 .039
1928 .030
1926 -.014
1925 .046
1922 .023
1921 .003

and their opponents DER was better when playing the Phillies away from Baker by:

1930 .054
1929 .058
1928 .011
1927 .027
1926 .029
1925 .035
1922 .030
1921 .019

So their opponents seem to have played even worse defensively in the Baker Bowl as well as measured by DER.
9:52 PM Mar 21st
 
bjames
Regarding the Baker Bowl, it is really difficult to normalize the context because the team was so phenomenally bad year after year, almost always losing 100 or more games in that era. The Phillies played in the Baker Bowl from 1895 to 1938. They had better than league DER in 1901, when they were 83-57, 1905 (83-69), 1906 (71-82), 1909 (74-79), 1911 (79-73), 1915 (90-62), 1916 (91-62), 1918 (55-68), 1934 (56-93). Their DER is low in most other years, but then, their won-lost records are really terrible, so it would be very surprising if they DIDN'T have a low DER.

If you gave them 20 points in DER, in many seasons or most seasons that would mean that their DER would be above average. In 1936, for example, their DER was .699, the league average was .709. If you gave them 20 points, they go from the worst DER in the league to maybe the best, or close to the best--but their won-lost record was 45-105. It seems REALLY improbable that they had one of the better defenses in the league. In 1921 they had a .695 DER against a league average of .706, but they finished 51-103. Do you think it is likely that their defense was significantly better than the league, rather than significantly worse?

I doubt that there was any DER disadvantage AT ALL to playing in the Baker Bowl. If there was, it 2 or 3 points.
2:57 AM Mar 21st
 
bjames
Regarding the '68 Indians, I think you shorted them one center fielder. Russ Snyder, who played 54 games/440 innings in the outfield that year, played mostly right field that year, but played over 3000 innings in center field in his career, and played far more in center than in left or right. He was also very fast, for that era; he was a career fourth outfielder, but a good one.

The '68 Indians were a good team, finished third. The 1-2 and 4 starters (Luis Tiant, Sam McDowell and Stan Williams) all had career low ERAs. Of course, a lot of pitchers had career lows in ERA in 1968, but a lot didn't, too.
2:41 AM Mar 21st
 
Fireball Wenz
The 1968 Indians team cropping up on this list is interesting. The infielders were CLEARLY nothing special - Tony Horton, Vern Fuller, Larry Brown and Max Alvis. Chico Salmon played a lot. Mediocre at best. The Of was Lee Maye (age 33), Jose Cardenal, and Vic Davalillo. Two of the three can play CF, Davalillo if RF because of the better arm. On the bench were Jimmie Hall (could play CF), Tommy Harper (could play CF but a speedy but not well regarded OF), and Lou Johnson. Cardenal, Harper and Davalillo probably among 6 fastest OF in league.
6:59 PM Mar 20th
 
willibphx
Good catch on the 30 Phils team, I mistakenly picked up Shibe instead of the Baker Bowl. Yes, per Diamond Mind park effects. 1b were 110% and 2B were 130% to 140% depending on the hitting side. Based on some quick math that could reduce the hits allowed for the Phils by 120 hits from 1B and 2B and improve their DER by over 20 points.
9:17 PM Mar 19th
 
TJNawrocki
What has not been mentioned yet, and knowing Bill, will also be added in is Park Effect, in particular singles. I don't think would explain the 30's Phillies, they were just awful.

It sure looks like extreme hitters parks might have an influence on low DER. The 1923 and 1930 Phils, the two Rockies teams, and the 1987 Red Sox all played in extreme hitters' parks. Don't know about the Baltimore Terrapins' home park.

Also, it's very interesting what happened to the Rays between 2007 and 2011.
4:27 PM Mar 19th
 
stevebogus
DER measures the rate that batted balls turn into outs. Flyballs which stay in the playing field become outs at a higher rate than groundballs. The HRs are subtracted from both sides of the equation, and flyball pitchers tend to produce more pop-ups, which become outs about 99% of the time. So, a flyball pitching staff will already tend to have a better DER. Now when you combine that with outfielders who can cover a lot of ground you can get the even higher DER than a groundball staff, so "fast outfields" dominate the top of the list.
3:18 PM Mar 19th
 
tangotiger
If it's not obvious, DER = 1 minus BABIP

BABIP is batting average on balls in play. And this makes up the core argument to DIPS, which is what rose Voros to prominence 20 years ago.

In other words, Bill James was on the cusp of DIPS 15 years before the rest of us.
3:07 PM Mar 19th
 
shthar
Does this mean that a 'great' defensive infielder isn't really doing that much more than an average infielder?

But a great fielding outfielder? whooo boy!


1:15 PM Mar 19th
 
shthar
How do double plays work into this?


1:12 PM Mar 19th
 
hotstatrat
Ooops, I meant no. 3
12:51 PM Mar 19th
 
hotstatrat
I was shocked to see the 1981 Tigers at no. 1.

I, too, have always felt having excellent outfield defense is of under-rated importance. But, the '81 Tigers still had Steve Kemp is left-field - well before they traded for one of Detroit's greatest defensive outfielders: Chet Lemon. Instead they had a mixture of Kirk Gibson, Rick Peters, and principally veteran right-fielder Al Cowens in center-field.

Morris, Petry, Wilcox, and Rozema must have produced mainly ground balls for Trammell, Whitaker, and Brookens to suck up.

It was also a shortened season - which produces more anomalies.
12:50 PM Mar 19th
 
stevebogus
We can be specific about when DER reached modern (current) values. The dividing line is the 1993 season.

1987 .715
1988 .722
1989 .715
1990 .717
1991 .719
1992 .719
1993 .710
1994 .704
1995 .706
1996 .702
1997 .703
1998 .704
1999 .702

While it is true that we are now in a throw hard swing hard period, that was less true in 1993 when the change occurred. 1993 is the transition year between two eras. The rate of HRs and doubles increased at the same time DER declined.
12:45 PM Mar 19th
 
evanecurb
I understand that the number of balls in play that are converted to outs involves some luck. At what sample size (as measured by number of balls in play) is luck no longer a significant variable?
12:37 PM Mar 19th
 
Gfletch
Speedy outfielders connected to high DER? Seems to me that if this is true you would see significantly fewer doubles and triples allowed.
11:57 AM Mar 19th
 
willibphx
What has not been mentioned yet, and knowing Bill, will also be added in is Park Effect, in particular singles. I don't think would explain the 30's Phillies, they were just awful. I would also note that this is the inverse of BABIP for hitters which at an individual level gets a lot of discussion.​
10:59 AM Mar 19th
 
bjames
'69 Mets DER was .748, best in the National League. They were 1.6 standard deviations above the norm, score of 116. 95th percentile of all teams.

9:56 AM Mar 19th
 
Robinsong
One way to try to separate luck from skill on DER might be to look at the correlation over time. The Chicago Cubs of 1905-1910 lilkely had good DER year after year. Sure, people get traded or get hurt or age, but it would help and could be compared to the consistency of, say, pitcher strikeouts or offensive stats. One could also try to split seasons to look at consistency first half to second halp.​
9:34 AM Mar 19th
 
evanecurb
How does the 1969 Mets’ DER compare with other teams in the study? Their raw DER (as calculated by BBRef- different method) was best in MLB that season.
8:04 AM Mar 19th
 
 
©2020 Be Jolly, Inc. All Rights Reserved.|Web site design and development by Americaneagle.com|Terms & Conditions|Privacy Policy