Catchers: Better as Veterans

by Tom Hanrahan SABR, The Best of By The Numbers


A study which shows that as a catcher moves from rookie status to spending years with the same team, the pitching staff ERA improves dramatically.

INTRODUCTION

A catcher’s main job, everyone knows, is to call a game and handle the pitching staff. Everyone may know this, but in a game that has statistics for virtually everything, there seems to be precious little time and energy devoted to measuring how well catchers do their main job. Rather, we see catchers’ defense measured by how many base stealers they throw out, and all of their other defensive skills (“framing” the pitch, setting up the hitter, bringing along the pitcher) are defined anecdotally by TV and radio announcers. This study is an attempt to determine if catchers’ defensive abilities as a whole improve as they mature and adjust to a pitching staff, and to quantify this as much as is possible.

            We start by asking the question “what general factors might affect a catcher’s ability to handle a pitching staff?”. I suggest that his ability might vary with:

1)                  His age and experience

2)                  His familiarity with the pitchers he is catching

3)                  His familiarity with the batters his pitcher is facing

There may be other specific factors like special tutoring under a specific coach, but these kinds of things will not help us answer the question in general. Tests could be set up to measure any one of these. My focus in this study was to see if catchers’ defense improved in the whole, as they became familiar with a pitching staff. At the end of this paper, I will look at how the findings here might be dissected into the individual factors.

 

THE STUDY


            How do we best measure a catcher’s defensive abilities? I propose the only reasonable answer is the ERA of the team for which he is catching. How could we best isolate the catchers’ defensive ability from all of the other factors that cause a team ERA to rise or fall? I attempted to do this by using all of the teams that had the same primary catcher in consecutive seasons. I defined “primary” as having caught at least 85 games during a season. I used the years 1946-1987; beginning after the players returned from WWII, and going through the last year in my Baseball Encyclopedia. No adjustments were made for strike years or change in length of the season. I did not use years when a team changed cities (Brooklyn to L.A.).

There were 104 catchers used in the study. The total number of catcher years was 539, representing about 60,000 total games caught. This gave me a large set of matched pairs of teams in consecutive years using the same catcher. I found the team ERA for each year, and compared it to the league average. I also recorded the number of games the team’s catcher had caught in his career prior to the start of the season. Obviously there was always some movement of pitchers between years, some hurlers improving or declining, changes in the team defense at other positions, and changes in ballpark dimensions. But if I could get a large enough sample that all of these other factors got washed out in the noise, I would be able to see if the number of games caught by the catcher was an important contributor to the team ERA.

            As an example, I will use Bob Boone’s career. He caught at least 85 games every year except the strike year of 1981.


Table 1. Bob Boone's Career

Year

Team

Previous Gm Caught

Tm. ERA

Lg. ERA

Difference from League

Difference from Prev Yr

Difference from 3 Years Ago

1973

PHI

14

4.00

3.67

+0.33

 

 

1974

PHI

149

3.92

3.62

+0.30

-.03

 

1975

PHI

295

3.82

3.63

+0.19

-.11

 

1976

PHI

387

3.10

3.50

-0.40

-.59

-.73

1977

PHI

495

3.71

3.91

-0.20

.20

-.50

1978

PHI

626

3.33

3.58

-0.25

not used

-.44

1979

PHI

755

4.16

3.73

+0.43

.68

.83

1980

PHI

872

3.43

3.60

-0.17

-.60

.03

1982

CAL

1085

3.82

4.07

-0.25

not used

not used

1983

CAL

1228

4.31

4.06

+0.25

not used

not used

1984

CAL

1370

3.96

3.99

+0.03

-.22

not used

1985

CAL

1507

3.91

4.15

-0.24

not used

.01

1986

CAL

1654

3.84

4.18

+0.34

.58

.09

1987

CAL

1798

4.38

4.46

+0.08

-.26

.05

 

 

            In Boone’s rookie year, having only caught 14 games prior to 1973, the team ERA was .33 runs per game higher than the league average. In his second season, this improved slightly to .30 higher than the league average; just .03 runs per game better. Continued improvement was shown in the next two years, after which there was a meandering slow drop off until he retired. We have 7 pairs of years while he was with the Phillies (73-73, 74-75…. 79-80), and 5 pairs of years with the Angels. Even if he had caught full time in 1981, the pair of years 80-81 would not be used in the study because he switched teams (and pitching staffs).

           

ORGANIZING THE DATA

            I built two groups of data. In the first, I grouped the year-pairs in bins of hundreds of career games caught: 0-99, 100-199, etc. I only used those pairs of consecutive years where the catcher’s games crossed from one grouping to the next. Thus, we can use Boone’s 73-74, but not 77-78, because he crossed right through the 500’s. This grouping was used to focus on changes from one year to the next, so I could build a function over time. Controlling the number of games caught (by 100’s) allowed me to use that as the variable that could link one group to the next. There were 306 matched year-pairs using this method. I think this method gives a good deal of organization to the data (it’s easy to use and see the trends), but I did lose a few of the samples.

Secondly, I compared rookies to veterans directly by comparing years that were somewhat further apart. In the next grouping, I again organized the data into bins of hundreds, but this time I compared them not to the previous year, but to their record 3 years ago, having caught for the same team for 4 consecutive years. I did not control how many career games the catcher had 3 seasons prior. Going back to the Boone example, between 1973 and 1976 the Phillies’ ERA improved relative to the league by .73 runs per game (from .33 to -.40). So, I recorded one data point for a catcher with career games caught in the 300s, a team ERA of -.73 compared to 3 years ago. This second grouping contained fewer points, because not as many catchers started at least 85 games for the same team for this length of time. I chose 3 years as the comparison point because

a)      the more years apart, the less data there are, so using a longer time span would be difficult, and

b)      the results of the first data grouping suggested that a 3-year span would show noticeable differences.

After trying this 3-year comparison, I wound up focusing exclusively on comparing raw rookies to veterans, since this is where the most obvious differences appeared.

 

COMPARING CONSECUTIVE SEASONS (GROUPING ONE)

            I found 49 consecutive year-pairs where the catcher’s career games caught went from between 0-99 to between 100-199. The average team had an ERA of .07 runs per game lower when the catchers had the extra year (= 100 games) of experience. Table 2 shows the data from every bin. As the amount of data became small for catchers with over 1000 games, I combined the last groups to ensure my sample sizes were at least 15.

 

Table 2

 

Career Games Caught  Between the 2 Years

Number of samples (catcher year-pairs)

Average ERA difference (lower is better)

Standard deviation (SDEV) of ERA differences

Cumulative ERA difference

000s – 100s

49

-.07

.40

-.07

100s – 200s

44

-.02

.43

-.09

200s – 300s

39

-.12

.40

-.21

300s – 400s

29

-.04

.37

-.25

400s – 500s

25

-.02

.36

-.27

500s – 600s

28

-.02

.36

-.29

600s – 700s

22

.00

.34

-.29

700s – 800s

20

.06

.34

-.23

800s – 900s

15

.04

.25

-.19

900s – 1000s and

1000s – 1100s

19

.08 per year

.37

-.11         -.03

(1000s)   (1100s)

1100s – 1200s thru

1600s – 1700s

16

.11 per year

(very limited sample)

.27

Sample is too small to make projections for 6 years

            The data in Table 2 strongly suggest that the defensive ability of the catchers improves steadily until they have caught somewhere between 400 and 800 games with the same club. The team ERA drops about three tenths of a run per game from the time they have their first full season until they reach this level of maturity. After this there is a slow rise in the team ERA until the catcher retires.

STATISTICIAN’S CORNER (ignore this if you’re not into the real technical math stuff): The variation (or “noise”) in measuring ERAs from year to year is measured by the SDEV (column 4). This typically was about .35 to .40 runs per game. We can use this to measure how certain we are that the average ERA difference is not just a random chance happening, assuming the data is normally distributed, which seems to be a reasonable assumption in this instance. In the first row, the average ERA difference is -.07, and is based on 49 samples. The SDEV of the average ERA difference (that is, the SDEV of column 3) is found by dividing the sample SDEV by the square root of the number of samples; in this case,


So, the average ERA difference is -.07, plus or minus .06. We can create a confidence level that states that the average ERA difference in row 1 is between -.19 and .05 (plus or minus 2 standard deviations) with 95% certainty. However, it is difficult to determine the cumulative SDEV between data in non-adjacent rows because of correlation in the data. The most conservative estimate, assuming zero correlation, would be obtained by taking the root sum square (RSS) of each row’s SDEV of the average. For example, in column 5 of row 5, we see the difference in team ERA between the catchers with less than 100 games caught and those with 500-599 games caught is -.27 earned runs per game. The SDEV of this number is .15, computed by RSSing each of the data in column 4 divided by square root of column 2 data. In reality, the true variation is likely much lower than this, since random noise causing a high value in row 2, for example, also causes a low value in row 3, since some of the same data are used to compute each number.

The right most column of Table 2, cumulative difference, is plotted as chart 1. The bars show the worst case confidence limits of the cumulated difference in ERA from year 1 through the year plotted.

Chart 1. Cumulative ERA Differences by Catcher Experience



COMPARISONS OVER 3 YEARS (Grouping Two)

Table 3 shows the catcher year-pairs organized by bins of hundreds in a different manner. The 279-300s row shows that there were 14 catchers we could use to compare the team ERA between the year when they had between 279-399 games caught under their belts, to the team ERA 3 years prior to that. The average number of games caught in each career 3 years prior is shown. The first row indicates that after 3 years, the team ERA averaged .28 runs per game lower. It also shows that of the 14 teams represented, that 12 of the 14 had a lower team ERA (relative to the league average) when the catcher was a veteran of 279-399 games, as compared to 3 years prior when he had only caught an average of 27 games in his career.

 

Table 3

 

Career Games Caught entering the latter year

Average of Career Games Caught 3 years prior

Number of Samples (Catcher year-pairs)

Average ERA difference

 (lower is better)

Number of Teams with lower / higher ERA after 3 years

279* - 300s

27

14

-.28

12 / 02

400s

82

24

-.38

20 / 04

500s

172

33

-.15

23 / 10

600s

About 280

26

-.10

16 / 10

700s

About 380

22

.04

08 / 14

800s

About 480

19

.16

06 / 13

900s – 1000s

About 630

24

-.02

12 / 12

1100s – 1700s

350 fewer

28

.07

11 / 17

           

    * 279 was the minimum number of career games caught for any catcher who also was a starting catcher 3 years ago.

           

            The data in Table 3 is pretty much in agreement with that in Table 2; significant improvement in team ERA the first few years, and a slow decrease in performance toward the catcher’s later years. The item that jumped right out at me was the first 2 rows of the right hand column. Out of 38 teams, 32 of them had ERAs that were lower with the catchers who had an extra 3 years of experience. With all of the changes that likely occurred in the team pitching staffs and other defensive changes over the years, this strikes me as remarkable that about 85% of the teams would improve their pitching.

            As I studied the 38 catcher seasons involved in the first 2 rows, I noticed that the trend was even stronger when using just the catchers who had virtually no previous major league experience. So, I organized the data one last time, using ONLY the catchers who had VERY little experience (fewer than 50 games) prior to their first full-time year, and making comparison to their “prime” years. Table 2 shows that the catchers’ prime seemed to be when he had previously caught between 400 and 799 games (this is where the cumulative ERA was the lowest). I found all catchers who        

a)                  caught at least 85 games in a season, having had 50 or fewer career games coming into that year,  AND  

b)                  caught at least 85 games in other seasons, with the same team, having between 400 and 799 career games caught before these other seasons.

There were 16 comparisons. The teams, “rookie” years, and catchers used were:

NATIONAL

 

AMERICAN

LA
Phi
SF
StL
Chi
Cin
Pit
Phi

1958
1961
1962
1963
1966
1968
1969
1973

Roseboro
Dalrymple
Haller
McCarver
Hundley
Bench
Sanguillen
Boone

 

Bos
Bal
Was
Chi
NY
Bos
Tex
Min

1952
1956
1966
1969
1970
1972
1974
1976

White
Triandos
Casanova
Hermann
Munson
Fisk
Sundberg
Wynegar

            I recorded the team ERA (relative to the league) in the rookie year, and the average team ERA of all years used in the “prime veteran” classification. Of the 16 teams, only ONE had their ERA get worse when the catcher went from rookie to veteran status; fifteen teams had better ERAs with the veteran catchers. The average improvement was .47 runs per game, or 76 runs over a 162 games season! This is very likely a larger difference than importing Ozzie Smith, Willie Mays or Bill Mazeroski in their primes to help your defense. It’s even more remarkable when you consider that the ERA comparisons are for the whole season, including the games these catchers did NOT start. Many of these catchers caught three fourths or less of their team’s games, so the improvement per game caught might be 30%-40% more! The data for these 16 teams and catchers are graphed in chart 2.

Chart 2


I went back and checked to see what each of these 16 teams’ ERA was in the year PRIOR to these catchers being rookies, just to make sure that what I was seeing here wasn’t some strange effect, such as a group of all-world defensive catchers (there WERE some mighty fine names in this bunch) helping their teams tremendously while they were in their peak years. These teams had their ERA go UP relative to the league an average of .22 runs per game in the year that they were full time rookies (the years given above in the list). In other words, in their first year, these catchers appeared to hurt their team defensively by a fifth to a quarter of a run per game. Then, over the next 3 to 7 years, their defensive skills improved enough to help their team ERA to go down by almost half a run per game, so there was some net improvement comparing their prime years to the year before they showed up.

            One of the teams in this study was the ‘58 Dodgers, who moved into a vastly better pitchers’ park in 1962, so we shouldn’t be surprised that the team ERA improved so much with Roseboro catching as he became a veteran. Still, tossing out one data point won’t make that much difference.

 

OBJECTIONS – Let’s play devil’s advocate.

1.                   Maybe this sample is too small and we’re seeing some random chance effects.

            Overruled. Already covered this; there’s too much data here. When 15 of the 16 teams improve over time…well, if you flip 16 coins, 15 of them will come up heads less than 1 time in 3,800.

2.         What if the catchers represented an anomalous group of some kind?

                        In the second grouping (comparisons over 3 years), obviously catchers who washed out of the majors didn’t factor in, since they never reached veteran status. So one could argue that maybe these were the catchers who DID learn how to call a game, and the others did not. But, in the first grouping, we used consecutive year-pairs across every level of games caught, and the same pattern was evident. Overruled again.

3.         Park factors? Moving over time to a pitching-dominated era? Great hurlers flocking to these teams for a chance to pitch to these guys? A disproportionate amount of good teams and/or catchers in the sample?

                        We compared everything relative to the league and within the same teams, to get rid of park and trend effects. These guys were good catchers on good teams…which came first, the chicken or the egg?

 

CONCLUSIONS

             A typical catcher handles a pitching staff better over the course of his first few years in the majors with a club. This is evident by the rather dramatic drop in the team ERA of about a quarter to a third of a run per game from his rookie season to his prime years with a club.

If you have a veteran catcher who has been with your team for some time, and you’re thinking of trading him and calling up the young phenom from AAA, you can expect your pitching results to get worse. Of course, you ought to call him up SOMETIME, but don’t expect the team to improve right away. How many catchers are offensively 50 runs a year better than their replacement? (Piazza begins and ends the short list)

The differences in catchers’ stolen bases allowed are apparently LESS important than his other defensive abilities. The worst throwing catchers in the majors do not allow anywhere near one stolen base per game more than Ivan Rodriguez does.

FINALE

If differences this large show up comparing CLASSES of catchers, does this not infer that there might also be large differences between INDIVIDUAL catchers? This study suggests that the measuring of catchers’ defensive contributions may be the single most important yet unanalyzed ingredient of determining team success in the game today.




This article originally appeared in the August 1999 issue of By The Numbers.


Hall Of Fame Catchers List of Catchers All-Star Catchers Post Season League Awards Batting Fielding Records Greatest Catcher 1000 Games Caught HOME - Front Page Equipment Skills & Drills Articles About Catchers Trivia Quizes Quotes Rotisserie All-American Girls Professional Baseball Links 19th Century Current Catchers Ex-Catcher Managers Feedback HOME - Front Page HOME - Front Page Miscellaneous