Catcher Fielding

One of the things you learn in a basic statistics class is the difference between the kinds of data you can get. While this is rudimentary, it is easy to get them confused when you are looking at fairly complex data. No one really confuses nominal data. These usually take the form of yes or no and tend to be binary in nature. The confusion usually comes when we look at the difference between ordinal data and interval data. This happens when we do something like this. When data isn’t scaled in an identical (or even similar) then the combination becomes ordinal because we simply cannot trust the combination. This is even if each individual source was interval in nature.

Ordinal data is what we use when we want to simply rank one player over another. In our case, we are ranking the ten catchers in the Hall of Fame in terms of fielding. The problem is when we start looking at multiple sources we start looking at wins in some cases and runs in other cases. We are then looking at each one in isolation and standing them next to another that we look at in isolation. When we start combining these we start to get into some serious trouble. This is especially true if we attempt to make any conclusions about the actual value of a player. The differences in value are what we would call interval data. Interval data is the holy grail of data because it not only says who is more valuable than another, but by how much.

When we look at players from different eras and compare their fielding numbers we need to eliminate two words from our vocabulary: greatest and best. Those are conversations best left for the sports bar. We also get into trouble when we start using terms like difficulty and importance. When criticizing fielders, some passionate fans will tend to get defensive and say something like, “well, how much baseball do you play?” or “I’d love to see you go out there and catch.” I’ll admit right now. I never played catcher. When I was a kid I played in the outfield so I am saying nothing about the importance or difficulty of the position. At least I’m not saying it directly.

That being said, the data does say something about the relative importance of the individual skills a catcher must possess. Rating catchers is always more difficult than any other position because they must call the game, block pitches in the dirt, and control the running game of the other team. In the current game we also include pitch framing in that equation. As strategy has changed, the relative importance of all of these skills fluctuates. For instances, controlling the running game was huge throughout the 1960s, 1970s, and 1980s. So, catchers from that period could accrue more value if they were excellent at that particular facet of the game. Conversely, during the live ball era (1920s – early 1940s) hardly anyone stole bases. So, one could be brilliant or not and you wouldn’t see a range of value there.

So, I would have a hard time saying that a Johnny Bench or Gary Carter was a greater fielder than a Gabby Hartnett based on the numbers. Greatness implies that the skills were superior and while that may be true, it is not true based on the numbers. Players are always better understood when they are compared with their own generation. We know more about Hartnett when we compare him with Mickey Cochrane and Bill Dickey then when we compare him with the catchers from the 1970s. Those that saw both could weigh in, but that would be a more qualitative analysis. This becomes increasingly problematic when we include new analysis like pitch framing. The modern catchers become much more understood when we include that, but we can’t include it for the earlier catchers. While we have play by play data going back to the beginning of the century, that play by play data won’t tell us if an individual pitch should have been called a strike or a ball if the catcher were better at framing it.

You undoubtedly did not open this up to hear my caveats about fielding, but it represents the very best reason why I am not comfortable naming a best catcher of all-time. At least, I’m not using the data here to define it. Give me a beer and a good game on television and I’ll throw down on any player, but I’m leaving the charts and graphs at home. We will split the fielding data into two categories. First, we will look at the major career data for our three sources. Keep in mind, since some use wins while other uses runs we will not combine them. Our WAR and WS from the previous article has already embedded that in their totals. Here, we are distilling out the numbers to get a sense as to who was the most valuable fielder of all-time.

Career Fielding 

  DWS/5 dWAR TZ
Johnny Bench 18.7 19.3 97
Yogi Berra 19.9   8.7 33
Roy Campanella 13.7   5.7 17
Gary Carter 24.1 25.5 106
Mickey Cochrane 16.3   4.4 -2
Bill Dickey 19.0   7.6 20
Carlton Fisk 22.4 16.4 30
Gabby Hartnett 21.9   6.6 12
Mike Piazza 12.0   1.0 -25
Ivan Rodriguez 30.1 26.7 135

The above represents my best attempt to normalize the data. The win shares we simply divided by five to make it roughly equal to defensive WAR from baseball-reference. Unfortunately, I struggled to normalize the total zone runs from Fangraphs because it would change the rankings. Typically, we would add ten runs per season to the totals because that would convert from a comparison against the average to a comparison with the replacement level player. In the end, the desire to present the data as those sites/sources wanted ended up winning out.

However, this is where our discussion of ordinal versus interval data comes in. The above numbers still don’t make a whole lot of sense when you put them together because all three sources have considerably different opinions of how much value individual events (or skills) have. Those change between eras as they change between platforms. However, if we show the same table and replace the win totals with their simple rankings in those systems we see something quite remarkable

Fielding Rankings 

  DWS dWAR TZ
Johnny Bench 7 3 3
Yogi Berra 5 5 4
Roy Camanella 9 8 7
Gary Carter 2 2 2
Mickey Cochrane 8 9 9
Bill Dickey 6 6 6
Carlton Fisk 3 4 5
Gabby Hartnett 4 7 8
Mike Piazza 10 10 10
Ivan Rodriguez 1 1 1

I would surmise that we are actually seeing more variance here than at any other position. Different sources treat each individual skill a catcher must possess differently and with the different weights can come wildly different values. Still, only two players are more than two places apart across the board. When win shares is dropped that moves to zero. Four of the players are exactly the same including the most valuable and least valuable.

That being said, this is one of the areas where we have to be careful. Notice how Carlton Fisk drops when compared to the average as opposed to replacement level. He caught for over 20 years, so he managed to accumulate a lot of value. That’s not the same thing as being the greatest or the best, but it is important to note. This is especially true when we start looking at things like Gold Gloves, Win Share awards, and total zone awards. Lacking in those awards does not mean you lack value. It just means you were never the best in any one particular season.

Moreover, this distinction can be seen more at the bottom of the scale than the top. Mike Piazza was universally regarded as a poor defensive catcher (although he rated highly in pitch framing), but depending on the source he still had value when you compared him to the replacement level catcher. This is where we include the remarks of our critics when they say, “I’d like to see you go out there and do it.” Why yes, we aren’t saying any of these guys were a complete buffoon. It’s all comparative in nature and unfortunately you are stacking each up against the very best to play the position. Some aren’t going to come out looking good in every single category.

All that being said, we will finish our journey by looking at individual season honors. We do have some limitations here. Total Zone awards aren’t an official honor, but we only have data going back to 1953. The Gold Glove awards were awarded in both leagues for the first time in 1958. So, some of our players will not have any chance to win anything and others only a limited opportunity. However, win shares awards were added back through all of our players careers. When all things are considered, these are fairly meaningless, but I know many of you will be interested, so we close with this table.

  GG WS TZ
Johnny Bench 10 4 6
Yogi Berra 0 5 3
Roy Campanella 5 1
Gary Carter 3 8 5
Mickey Cochrane 5
Bill Dickey 5
Carlton Fisk 1 0 0
Gabby Hartnett 6
Mike Piazza 0 0 0
Ivan Rodriguez 13 7 9

Win shares is the great equalizer thanks to James’ hard work in breaking down the data for us season by season. When we see how evenly distributed the awards go we can see that greatness is really more evenly distributed than the basic numbers would indicate. It definitely is more evenly distributed than the Gold Gloves themselves. However, we return to our original caveat: greatness and value are two entirely different things. Just because Carlton Fisk won no awards doesn’t mean he has zero value. It’s just that he had the misfortune of catching at the same time as Johnny Bench, Gary Carter, Bob Boone, Jim Sundberg, and Ivan Rodriguez at the end of his career.

Where you can find me

The Hall of Fame Index

Most of the same topics we are covering on this site were covered here originally back in 2011. However, I also offered commentary on players elected by the Veterans Committee and also gave a complete list of who should be in and who should be out. I’d honestly recommend both. Please go with the paperback option. Tables and graphs don’t translate well to e-books.

Combating Ignorance: Inside and Outside the Classroom

These are a collection of stories from my first ten years in the classroom. The names, dates, and places have all been changed to protect the innocent, but the stories or all true. They are weaved in a way that will explain the world of education to those that live outside of it. Of course, fellow teachers know how true and believable all of the stories really are.

The State of Baseball Management

This is an earlier book from my writing career where I looked at the decision making of the best five teams and worst five teams between 1993 and 2003. The idea is that money plays a role in team success, but the biggest factor in success and failure are the quality of decisions teams make in trades and free agent signings.

Checks and Imbalances

This was my first book back in 2002. This book looked at the economics of the sport and the history of competitive balance (or not) from history. In the end, I made some suggestions on how the sport could gain more competitive balance.

The Fantasty Fix

I currently write for the Fantasy Fix where I share my thoughts on fantasy baseball. I am currently working on my rankings for players in advance of the 2018 season, but I also comment on trades and signings and how they affect the typical fantasy baseball player.

 

 

 

 

 

The New Catcher Index

A lot has changed with the index, but we’ve almost come full circle since the book was published. The inclusion of win shares was controversial at the time because wins above replacement had leapfrogged Bill James contribution and that hasn’t really changed in terms of popularity. However, James recently criticized WAR because it is based on runs scored and allowed instead of the number of games a team actually wins or loses. For most fans, their eyes gloss over when hearing this kind of talk, but for the mathematicians out there this is a huge deal.

Teams usually finish within five games one way or another of their “expected wins”, but occasionally there are wider ranges in play. This happened last season in the American League when the Yankees finished nearly ten games worse than they were expected to. So, some of the positive run differential may not have been as valuable as WAR would make it out to be. In short, WAR did not account for the clutch element. In terms of a career these things usually even themselves out. From season to season there could be significant variance. The goal of the index is not to come down one way or another, but to gather a collection of the top sabermetric minds in the industry. So, we include what most people know as bWAR, fWAR, and win shares. For those new to this, bWAR stands for the WAR configured by baseball-reference.com, and fWAR stands for the WAR configured by Fangraphs.

James configured his win shares to be three times the number of wins a team earned in a season. So, a 100 win team would have 300 total win shares. Even with that multiplier, win share totals end up being considerably higher than WAR totals. Part of this is based on the fact that win shares always begin at zero whereas players can have negative WAR totals. James would later develop loss shares, but we aren’t going to confuse the process further. Instead, we will divide the win share totals by five in order to scale win shares similarly to the two WAR formulas.

Admittedly, we are creating a bit of a Frankenstein monster in terms of math. This is usually what happens when a liberal arts guy starts crunching numbers. However, the goal is not to determine who the best catcher was or pinpoint exactly who is where. The goal is to measure a player’s fitness for the Hall of Fame. We do this by determining what the industry standard is for guys that are already in. Occasionally, we find outliers (or players that shouldn’t be in), but most of the time we are simply using what the BBWAA has already done to determine what they should do in the future.

We will try to define things as we go, but if you want a more in-depth overview of why we are doing what we are doing you can take a look at the home article that describes the process. In short, there are ten catchers in the Hall of Fame that were voted in by the BBWAA. We are ignoring the Veterans Committee selections because they are rife with so much bias that it ends up muddying the waters beyond recognition. We will start with career value and then break that down first. Then, we will move onto peak value and go from there. In a book format we would look at offense, fielding, and other considerations all at once. At the website we can look at those things individually.

Career Value 

WS/5 bWAR fWAR Total
Johnny Bench 71.2 75.0 74.8 221.0
Yogi Berra 75.0 59.5 63.7 198.2
Roy Campanella 41.0 34.2 38.2 113.4
Gary Carter 67.4 69.9 69.4 206.7
Mickey Cochrane 55.0 52.1 50.6 157.7
Bill Dickey 62.8 55.8 56.1 174.7
Carlton Fisk 73.6 68.3 68.3 210.2
Gabby Hartnett 65.0 53.9 53.7 172.6
Mike Piazza 64.8 59.4 63.7 187.9
Ivan Rodriguez 67.6 68.4 68.9 204.9

There are two major differences between the index and most other rating systems. First, it combines a career and peak value element. The idea is that career value can sometimes be misleading on its own. Some players are solid players for a very long time but are rarely ever great. You don’t want a Hall of Fame of stat compilers. Using these kinds of formulas does limit that to a certain extent. A player could be replacement level for a considerable time and add nothing to their career value. However, usually players offer some minimal value even when they are merely average.

The second major difference is that there are no hard and fast rules as to what is Hall of Fame worthy and what isn’t. It’s more about the players’ proximity to each other. When we see separation we begin to take notice. Roy Campanella is obviously separated from his peers, but there are clear extenuating circumstances there that we will get to in another article. The rest all fall above 150 combined wins. The beauty is that you don’t have to know what that means in order to offer sound analysis.

Mickey Cochrane is also lagging behind the others in career value because his career was cut short due to injuries. However, his peak value might mitigate those differences and some and he might close the gap. So, we will look at their peak values. In this case, we take the top ten year stretch for each player. Campanella only played for ten seasons, so his peak value will be the same.

Peak Value 

WS/5 bWAR fWAR Total
Johnny Bench 54.6 59.9 60.4 174.9
Yogi Berra 55.2 47.2 50.8 153.2
Roy Campanella 41.0 34.2 38.2 113.4
Gary Carter 52.4 60.8 59.0 172.2
Mickey Cochrane 49.6 47.0 45.4 142.0
Bill Dickey 46.0 42.4 42.7 131.1
Carlton Fisk 38.0 41.9 40.2 120.1
Gabby Hartnett 41.0 35.8 35.2 112.0
Mike Piazza 54.6 54.0 59.1 167.7
Ivan Rodriguez 43.0 51.8 50.2 145.0

Bill James once described good statistics as having the qualities of language. While that kind of description is compelling, I’ve always likened it to either a painting or photograph. Good numbers paint a more accurate picture of a player. When you compare career and peak values you get a sense as to what a player must have been like during his prime. Someone like a Carter, Piazza, or Bench was truly dominant throughout the decade. Players like Hartnett not so much.

What is interesting is that all three statistical sources begin from the same premise. They are comparing players with a replacement level player. In other words, we don’t compare them with the average because even an average player has value. We compare them with the triple AAA equivalent. However, what that looks like varies differently depending on the source. What specific value specific events have also varies slightly. This is particularly true when we evaluate the fielding side of things. We will focus on fielding in a subsequent article. The purpose of this article is to give us a baseline for catchers where we can jump off on as many tangents as we deem fit.

The Hall of Fame Index 

Career Peak Total
Johnny Bench 221.0 174.9 395.9
Yogi Berra 198.2 153.2 351.4
Roy Campanella 113.4 113.4 226.8
Gary Carter 206.7 172.2 378.9
Mickey Cochrane 157.7 142.0 299.7
Bill Dickey 174.7 131.1 305.8
Carlton Fisk 210.2 120.1 330.3
Gabby Hartnett 172.6 112.0 284.6
Mike Piazza 187.9 167.7 355.6
Ivan Rodriguez 204.9 145.0 349.9

It’s at this point where I must remind the crowd of my usual disclaimers. First, the index is not meant to identify the greatest player at a particular position or in history. There are a number of problems with it that we will get to later when we cover individual categories like fielding. So, if people want to use it to justify that Bench is the greatest catcher of all-time be my guest, but that isn’t my intent. Secondly, there are no hard and fast cutoffs of who should be in and who should be out. It seems silly to set 300 as a line of demarcation and suggest that Cochrane doesn’t belong in. At the end of the day, we are talking fractions of a win even in the case of Hartnett. Remember, we have three different systems that are divided into career and peak value. The differences are magnified.

Additionally, identifying outliers is not the same thing necessarily as identifying mistakes. Roy Campanella is not a mistake by any stretch. He is a player with extenuating circumstances that override the statistical exercise. As an outlier, we place him aside and consider the rest. When a player gets north of Hartnett then he becomes a viable candidate. That doesn’t mean he definitely deserves to be in. It means that he should be a very compelling candidate. We will get those examples as well in other articles. Still, we will always use this article as a reference point for every other catcher we consider from here on out.

Welcome to the Hall of Fame Index

As some of you know, The Hall of Fame Index was published back in 2011. The idea at the time was to use a complex statistical system to see if there were patterns in the way Hall of Famers were selected by the Baseball Beat Writers Association of America and the Veterans Committee and then use that to compare it to modern players and players that had been overlooked. Through the process, we discovered that there was no rhyme or reason to how the Veterans Committee selected players. Some players fit and some players didn’t.

 

Whenever, someone publishes a book, there is always that moment when you wish you could have it back to make quick corrections. The Hall of Fame Index was several years in the making because I had those moments before publishing. Then, I’d have to start over. With this blog I can correct and update as I go. Suddenly, a static thing like the index can move constantly with the changing of the times.

 

The most significant change is that we will ignore the Veterans Committee selections and stick to the BBWAA. The idea here is that we want to determine a baseline for enshrinement that makes sense so that we can look at the players outside the Hall of Fame and determine which ones should get in. Doing this in blog form allows me to look at individual players and ask more focused questions that might be more interesting to those reading. Of course, the best part of the blog is that it is free to produce and free to read.

 

How does the Index Work?

 

As you’ve already heard, we are taking WAR data and Win Shares data and combining it into one number. The index has two equal aspects. The first is what I like to call “career value.” Simply put, I add the WAR together from the two sources and Win Shares (WS) divided by five into one number. Bill James calculated win shares to be three shares per win. However, he had a different calculation point than the makers of WAR, so his scores skew higher. We are dividing the total by five to make sure that it does not inadvertently account for more of the formula than the other two sources. All three take players from the replacement level and go from there. It helps us avoid what I like to call the Harold Baines problem.

 

See, the problem with Baines is that he was never really great. Greatness has always captivated the BBWAA. After all, it is called the Hall of Fame and not the Hall of Stats. So, I’ve added an element I call “peak value.” The trouble with peak value is that it has a number of different definitions across the board. For some it is five seasons. Others go seven season. Some count years consecutively while others do not. I have chosen ten consecutive seasons because that is the minimum needed to get into the Hall of Fame. Career value and peak value are added together to come up with a player’s Hall of Fame index.

 

How is the Index Used?

 

This part is unchanged from the book. My opinion on statistics has not changed remarkably since then. Comparing a catcher with a first baseman or center fielder makes very little sense. It also makes very little sense to pick out an arbitrary number and make that the dividing line for Hall of Fame fitness. The numbers will tell you where that line should be. There almost always is a gap between the bottom of the list and those that have been on the ballot and not gotten in. Naturally, there are exceptions to that rule and we will evaluate them one by one as we go on.

 

So, determining fitness is actually pretty simple. For instance, there are ten catchers that were elected by the BBWAA. Roy Campanella doesn’t really fit the mold for historical reasons we will address in a later article. So, you can reasonably compare any current catcher against the other nine to see if they belong. As we will see, a couple of catchers that have been retired for a few decades also belong in that group.

 

The purpose of the index is really two-fold. First, it aims to make sense of the selections that have already been made. The BBWAA usually has done a good job of selecting worthy candidates over the years, but they have made the occasional mistake either way. Secondly, it refines the debate. That is an important distinction. I’m not one of those that aims to replace the BBWAA or turn the Hall of Fame selection process into a simple statistical formality. Debate is good for the sport and one of the more fun activities in November and December when hardcore fans are waiting for Spring Training. So, by all means, join in the fun.