Deconstructing the Catcher: Valuating the Five Tools of the Backstop

“’A catcher works hard, . . . You should not bat him too high in the order.’” (Bang the Drum Slowly, Mark Harris, 1956) ^a

A Full Plate

The average major league catcher bats around seventh in a lineup. (6.8^th in 2019)

This is not surprising. Catcher is typically regarded as a poor-hitting position. 2012 was the most offensively productive year for catchers in the last 25 years, and catchers were still 5% below league average in hitting. The average batting order position for a catcher rose no higher than 6.3.

There are other reasons to bat a catcher lower in the order. It reduces his workload. The catcher is already involved in every pitch of the game. He squats and stands 150 times per game and throws the ball more than anyone else. He gets struck by 100 mile-per-hour foul balls, and even by the bat itself on errant backswings. He deals with managing the umpire, managing his pitcher, and keeping baserunners in check.

Stationed behind home, the catcher is the only player on his team with a view of the whole field. He has a lot on his plate, if you will excuse the pun.

The Five Tools of the Catcher

Scouts routinely refer to the five tools of position players: Hitting for average, Hitting for Power, Fielding, Throwing, and Running. Catching is a different beast compared to the rest of position players.

I present to you the Five Tools of the Catcher:

• Hitting
• Baserunning
• Pitch Framing
• Throwing Out Runners
• Blocking Pitches

It is hard to find a catcher who excels at all five of these skills.

After seven years with defensive-minded Jason Castro behind the plate, the Houston Astros have explored many different options as their backstop in the past three years. Each has had different strengths and weaknesses.

Brian McCann came into 2017 as a perennial 20 home run hitter, but he struggled to throw out baserunners. Evan Gattis also possessed admirable power at the plate, but was a defensive liability and better suited as a designated hitter.

In 2018, thanks to a McCann injury, Max Stassi saw increased playing time. He led the majors in pitch framing, but was at best average in everything else.

Later in 2018, looking to control the running game, the Astros traded for Martin Maldonado, who led the American League in Throwing Runs the year before. Maldonado was, however, a liability in the batter’s box.

For 2019, the Astros signed free agent Robinson Chirinos. Chirinos could hit well, and led the AL in pitch blocking in 2017. So far this year, he leads the majors in pitch blocking, and his bat has made him an All-Star Game voting finalist, but he remains poor at framing and throwing out baserunners.

So which tools are the most important? How much more important is each skill when compared to the others?

Fortunately, these five categories also happen to be the five categories considered when calculating WARP and fWAR, Baseball Prospectus’ and Fangraphs’ respective versions of Wins Above Replacement (WAR), a metric for evaluating the absolute value of a player’s performance to his team’s win total.

By breaking down these versions of WAR into their individual components, we can get an idea of each of these skills’ relative importance to team success.

The Ground Rules (Methodology, Part I)

This part is a little boring (or possibly, more accurately, “more boring”). But if you have come this far in article about catcher skillsets, you are probably on board for boring.

The formula for Baseball Prospectus’ WARP is:

(Batting Runs + Baserunning Runs + Framing Runs + Blocking Runs + Throwing Runs + Positional Adjustment + Replacement Level) / Runs per Win

Every player we are considering is a catcher, so the positional adjustment (which accounts for differing difficulty of different positions) will be the same. Replacement level will also be the same for all catchers. Runs Per Wins is used to convert the runs value of a player into team wins, by a ratio of runs to wins that varies by year. This is also a constant, when considering data from a single year.

Since these are all the same for every catcher, we can ignore these and focus on the part of the formula that varies from catcher to catcher:

Catcher Runs = Batting Runs + Baserunning Runs + Framing Runs + Blocking Runs + Throwing Runs

Fangraph’s fWAR is calculated in a similar manner. The sole difference is that since 2015, Blocking Runs is either no longer factored in, or it is has been incorporated into Throwing Runs. This nuance will be discussed later.

Baseball Reference’s bWAR was not considered for this analysis, as bWAR does not yet incorporate Pitch Framing.

The Catcher Spectrum (Methodology, Part II)

It is difficult, if not impossible, to say which skill is more important than another absolutely. However, if we establish what the spectrum is for being good at a skill vs. being poor at a skill, we can calculate the impact of the difference between having a catcher who is at the better end of the spectrum at a specific skill versus having a catcher who is at the poorer end.

Where do we set the goalposts for each skill? The limits of what is humanly possible? Should the upper limit be what has been historically possible in baseball? A theoretical catcher who hits like Ted Williams or who runs like Rickey Henderson would certainly be immensely valuable, but that just isn’t what’s out there.

This is what is out there:

We consider 2018 rather than 2019, so that there is a full season of data to work with. In 2018, 35 catchers had at least 250 plate appearances and caught at least 500 innings. Why 250 and 500? It is arbitrary. Generally, catchers have 1 plate appearance for every 2 innings caught (IC) (These 35 catchers averaged 2.06 IC/PA). If you increase the PA and IC requirements, you have fewer catchers to work with to establish the skill spectrum. If you decrease the PA and IC requirements, data points suffer from small sample size, particularly for offensive production.

We now have 35 catchers with whom to make a spectrum, sorted from the best to the worst. It is important to note that for this exercise, we are interested in each catcher as a data point representing a skill level, so we adjust any counting or cumulative statistics for playing time.

For example, in 2018, J.T. Realmuto led all catchers in batting runs above average, according to Fangraphs, with 16.6 batting runs above average over 531 PAs. But in this methodology, the highest data point belongs to Wilson Ramos, who had 15.9 batting runs above average in 416 PAs. Ramos in 2018 averaged 26.2 PAs per batting run above average, compared to Realmuto’s 32.0 PAs per batting run above average.

With the 35 data points, we could arbitrarily take the best and worst of the 35 catchers at a skill and mark them as 100 and 0, the upper and lower ends of the spectrum. But the #1 and #35 catchers at each skill can easily throw off the curve if they are aberrantly better or worse than the next best or next worse catcher.

An extreme example of this would be the 1984 NHL season. Wayne Gretzky was the #1 scorer with 208 points. The next 4 highest scoring totals were 135, 130, 126, 121. Dale Hawerchuk’s 130 points is excellent, but if The Great One’s scoring is set as the upper standard, Hawerchuk seems like a player who was just a little above average.

To account for this, rather than using the #1 and the #35 catcher as the upper and lower standards, we use the 90^th percentile and 10^th percentile catchers. Those catchers are assigned 100 and 0 spectrum scores. Thus, there are 3 catchers rating above 100, 3 catchers rating below 0, and 27 in between.

Once the is spectrum set, we calculate the difference in runs between a 90^th percentile catcher and a 10^th percentile catcher, as applied to a theoretical 400 PA, 825 IC Catcher season. (This ratio was determined by the aforementioned 2.06 IC/PA.) This is the difference is between a catcher good at a particular skill versus a catcher who is poor at it. By comparing the 90^th percentile to 10^th percentile difference for each skill and comparing them to the differences for other skills, we have an idea of the relative value of the skill, given the available catchers.

Jake Roth-USA TODAY Sports

TOOL #1: Hitting

Hitting is the easiest skill for a fan to appreciate. Hitting makes the highlight reels. Hitting makes for great baseball card action photos. Hitting gets Mike Piazza in the Hall of Fame. Hitting gets you in the All-Star Game. In fact, MLB doesn’t even list any other statistics besides hitting on the voting ballot.

Baseball Prospectus measures Batting Runs by BRAA (Batting Runs Above Average). The site does not go into a whole lot more detail on how they calculate it, but define it as “the number of runs a hitter produces relative to an average hitter, adjusted for park.”

It is likely calculated similarly to wRAA (Weighted Runs Above Average), which is how Fangraphs measure Batting Runs, as they correlate pretty well. wRAA is derived primarily from wOBA (weighted On-Base Average).

By dividing a player’s PAs by their BRAA or wRAA, we arrive at Plate Appearances per Batting Run (above average). Negative values indicate the number of PAs for that player to be worth one batting run less than average.

The difference between a 90^th percentile catcher and a 10^th percentile catcher in hitting over 400 PAs is 31.0 runs using Baseball Prospectus’ metrics, and 31.1 runs using Fangraphs’.

Photo by Michael Reaves/Getty Images

TOOL #2: Baserunning

Catchers are slow. Everyone knows it. When you make a kids’ movie about a baseball team, the fat kid is invariably the catcher, which plays to the idea that the catcher is the slowest player on the team. (But there is probably some additional thinking that a fat kid probably blocks pitches better, by taking up more space.) If you have a fast catcher, it is newsworthy. When Mitch Garver led off a game with an infield single on April 23, it was the first time a catcher did that in 9 years.

Baseball Prospectus measures Baserunning Runs in BRR. It is the sum of various baserunning components including GAR (Ground Advancement Runs), AAR (Air Advancement Runs), HAR (Hit Advancement Runs, SBR (Stolen Base Runs), and OAR (Other Advancement Runs)

Fangraphs measures Baserunning Runs in BsR. It is also a combination of several factors: wSB (Weight Stolen Base Runs), wGDP (Weighted Grounded Into Double Play Runs) and UBR (Ultimate Base Running.)

Although PAs are not an exact measurement of baserunning opportunities, it does measure offensive playing time in general. So like hitting, we calculate plate appearances per baserunning run (above average), by dividing PAs by BRR or BsR.

The difference between a 90^th percentile catcher and a 10^th percentile catcher in baserunning over 400 PAs is 4.8 runs using Baseball Prospectus’ metrics, and 4.4 runs using Fangraphs’.

Photo by Bob Levey/Getty Images

TOOL #3: Pitch Framing

Pitch framing is the skill of receiving a pitch by positioning one’s body and glove in a way that the home plate umpire calls pitches in the strike zone as strikes, and even some borderline balls as strikes as well. It is easily the least flashy of the tools. When a fan watches the game, if a pitch outside the strike zone is called a ball, the fan is more likely to attribute blame to the umpire than credit the catcher for framing the pitch well. If there were a framing competition in an All-Star Game skills competition, it would be the most boring event imaginable. Photographers do not take “action” shots of framing. It was rather difficult trying to find a photograph of a catcher “framing.” (In the end, I was not even successful, as you can see Lonnie Chisenhall has hit the ball in the photo, so there was nothing for the catcher to frame.)

But the skill is real and reproducible; the same names pop up on framing leaderboards each year. It is surprisingly valuable as well. The difference between a ball and strike can dramatically alter the outcome of an at-bat. Max Stassi’s elite pitch framing skills single-handedly vaulted him to the top of 2018 fWAR leaderboards despite being decidedly average in all other aspects of a catcher’s game.

Pitch framing is not going away anytime soon either. Fans and players may clamor for robo-umps, but there are a number of hurdles before electronic strike zones can be instituted.

Baseball Prospectus measures CSAA (Called Strikes Above Average). In combination with the total number of Framing Chances, BP calculates Framing Runs above average. The 35 catchers in this analysis averaged 7.34 Framing Chances per inning which is what we use when calculating the framing run production of the theoretical 825 IC catcher.

Fangraphs measures FRA (Framing Runs Above Average). It does not report framing chances, so we adjust FRA by IC instead.

The difference between a 90^th percentile catcher and a 10^th percentile catcher in framing over 825 Innings Caught is 25.1 runs using Baseball Prospectus’ metrics, and 26.9 runs using Fangraphs’.

Photo by Dilip Vishwanat/Getty Images

TOOL #4: Throwing Out Baserunners

Throwing out baserunners is what most people think of when they think about catcher defense. The stolen base is an exciting play. It even has an exciting name. A player is trying to “steal” second base? But the heavens have declared “Thou shalt not steal.” That is not an unwritten rule of baseball. It is a carved-in-stone commandment from Yahweh. The catcher who can punish the sinner is deserving of God’s glory.

The most commonly used statistic for throwing out baserunners is CS% (Caught Stealing Percentage), the percentage of times a catcher successfully throws out a would-be base stealer. This statistic does not consider that more than the catcher is involved in throwing out a runner. The pitcher’s delivery time and receiving ability of the second baseman or shortstop come into play as well.

Baseball Prospectus measures Throwing Runs by combining TRAA (Takeoff Rate Above Average) and SRAA (Swipe Rate Above Average). TRAA factors in how often baserunners attempt stealing while a catcher is behind the plate, whether they succeed or not. A catcher with a better reputation for throwing out runners will have less baserunners testing his arm. SRAA measures actual stolen bases taken on the catcher’s watch, but adjusting for factors such as the pitcher, runner, stadium and inning.

Fangraphs measures Throwing Runs as rSB, a DRS-based Stolen Base Runs Above Average. Similar to how Baseball Prospectus factors in TRAA, rSB also considers decreased stolen base attempts to be an indicator of good stolen base prevention.

To quantify each catcher’s stolen base prevention skill, we calculate IC per Throwing Run or per rSB.

The difference between a 90^th percentile catcher and a 10^th percentile catcher in stolen base prevention over 825 IC is 1.0 throwing run using Baseball Prospectus’ metrics, and 5.0 runs using Fangraphs’. The difference between BP and Fangraphs’ assessments is more marked than in any of the other catcher’s skills. This may be explained by Fangraphs’ lack of a separate Pitch Blocking metric (see next section.)

Photo by Joe Sargent/Getty Images

TOOL #5: Blocking Pitches

When you can’t catch, block. The difference between a wild pitch and a passed ball is often difficult to discern. Both cases result in baserunners being able to advance an extra base, but was it the pitcher’s or catcher’s fault? By convention, any pitch that hit the dirt and gets past the catcher will be considered a wild pitch. That does not mean the catcher was helpless though. Some catchers are better than others at keeping a pitch from getting past them to the backstop, in the same way some first basemen are better at picking bounced throws and saving would-be errors from their shortstops.

A walk with the bases empty is functionally no different than a single. It does not matter if the batter reaches first as a result of his hitting skill versus the pitcher’s inability to throw strikes. Similarly, advancing a base on a wild pitch or passed ball is functionally no different than a stolen base to the runner. It does not matter if the runner gains an extra base as a result of his base stealing skill versus the catcher’s inability to block pitches.

Baseball Prospectus measures Blocking Runs using EPAA (Errant Pitches Above Average), which measures the rate of wild pitches and passed balls above average. The lower the rate, the better a blocker the catcher is.

The 35 catchers in this analysis averaged 4.90 Blocking Chances per inning which is what we use when calculating the blocking run production of the theoretical 825 IC catcher.

Fangraphs used to measure RPP (Runs on Passed Pitches) and factor that into fWAR. They no longer do; it has not been recorded since 2014. No statement was made as to why RPP was removed, so it is unclear whether Fangraphs no longer factors in pitch blocking into fWAR, or if pitch blocking has simply been incorporated into rSB, Fangraphs’ Throwing Runs. rSB appears to approximate the sum of Throwing Runs and Blocking Runs on Baseball Prospectus, so this is possible.

The difference between a 90^th percentile catcher and a 10^th percentile catcher in blocking over 825 Innings Caught is 5.0 runs using Baseball Prospectus’ metrics.

Deconstructing the Catcher

We have now quantified the difference between the better catchers and the worse catchers at each of the five catcher skills. We can compare those differences relative to each other and arrive at a tidy, visually pleasing pie chart:

Some observations:

• Defense is nearly half the pie. Catcher is the most defensive position on the field, apart from pitcher, and this bears that out.
• Hitting and Pitch Framing are far and away the most impactful skills of a catcher. Having a catcher who is at the higher end of the spectrum versus the lower end for these skills contributes more to a team’s success. Hitting is probably not a surprise. These results, however, demonstrate how important Pitch Framing is. Its relevance is on the same order of magnitude as hitting. For as much attention as a fan or team should pay to a catcher’s OPS, xWOBA or wRC+, they should place just as much focus on their framing ability.
• The difference between having a Salvador Perez cannon arm who throws out baserunners and a Brian McCann who does not is minimal. Over the course of 825 innings, a good throwing catcher saves only 1 run over a poor throwing catcher. Throwing out a baserunner is still a significant play, but there is more to the play than the catcher. The runner may have been thrown out even if McCann were behind the dish; the runner may have been safe even with Perez catching.
• Pitch blocking is a more impactful skill than throwing. A catcher can prevent extra bases by the runner more effectively by focusing on blocking as opposed to practicing their throws to second.
• Except for J.T. Realmuto, catchers are generally poor baserunners. The “better” baserunning catchers are still pretty bad, so it does not make a great deal of difference if a team has a catcher at the higher end of the catcher baserunning spectrum.

Using Fangraphs’ metrics, the deconstructed pie chart looks very similar:

The main difference here is that more impact is assigned to throwing out baserunners. However, as discussed earlier, Pitch Blocking might simply be incorporated into Throwing out Runners.

House of Pies

With defensive skills occupying half of the pie charts shown, it is clear that having a better defensive catcher make a large impact on team success. For comparison, this is the pie chart demonstrating the run difference between the 90^th percentile and 10^th percentile 2018 shortstops at hitting, baserunning and fielding:

mhatter106

Shortstop is generally considered to be the next most defensive position after catcher. Yet the impact of having a good offensive shortstop is nearly 4 times that of having a good defensive shortstop.

We can also compare 2018 Catchers to 2017:

2017 and 2018 are fairly similar regarding orders of magnitude. Hitting and framing are again the most impactful skills. Baserunning, throwing and blocking are minor. Defense is well over half the pie in 2017, with framing having even more impact with a larger discrepancy between the better and worse framing catchers.

Reconstructing the Catcher

Ideally, you would like a catcher who excels at all five skills. That catcher is not in the league right now. In 2018, only one catcher ranked in the top half for all five skills: Buster Posey. Although Posey did not rank lower than 17^th out of 35 for any skill, he also did not rank higher than 12^th.

You do not need to be good at all five skills. Just like the five tools of the position player, not all tools are of equal value. For catchers, there is hitting and framing. Everything else combined does not have even half the impact of either of those skills.

We can take the spectrum scores for each skill for each catcher and create a weighted average of their scores, based on how impactful each skill was found to be for that year. Here are the top 5 catchers by weighted overall spectrum score:

(2019 data is based on half season collected on June 23, 2019. This will be discussed later.)

2017 Tyler Flowers stands out in particular. Not only does he rank at the top, but his score is over 40% higher than second place. This incredibly high rating comes on the back of a phenomenally good framing season (139), while at the same time having an excellent hitting season (108).

We began this article with a tour of Houston Astros catchers since the departure of Jason Castro. Let us take a look at their weighted overall spectrum scores:

(2018 Jason Castro and 2019 Max Stassi did not accumulate enough playing time to generate meaningful scores. 2017 Evan Gattis and 2018 Brian Gattis fell just short of eligibility requirements to be considered in the formation of the 2017 and 2018 spectrums, but scores were generated for them based on their limited playing time. 2019 results are based on half-season data.)

One thing that stands out is the volatility of the baserunning score. How was Brian McCann an 80 score for baserunning in 2017 and a -28 in 2019? Again, all catchers are terrible baserunners. As a result, the spectrum from the 90^th percentile to the 10^th in baserunning is quite narrow. With a narrow spectrum, small fluctuations can result in appreciable changes to a player’s spectrum score. The volatility does not affect the weighted overall spectrum score very much, because the same spectrum narrowness responsible for the volatility also causes baserunning to be weighted minimally.

[Full spectrum score data and rankings can be published later in a separate piece if there is demand (and possibly even if there is not), but this article has gone on long enough as it is, and I am fairly certain I wore out my welcome 6 or 7 sections ago.]

(UPDATE 7/14/2019: Full spectrum score data and catcher rankings for the 2016, 2017, 2018, and 2019 updated through the All-Star break, are now available at the following link: 2016-2019 Catcher Skills Spectrum Scores)

(UPDATE 11/12/2019: 2019 Catcher Skill Spectrum Scores and Rankings for the 2019 seaosn are now available at the following link: Garver, Grandal, Realmuto Top the 2019 Catcher Skill Spectrum Scores )

The Mystery Sixth Tool: Game Management

There is one other skill that I have not mentioned: Game Management. The ability to call a game by knowing what type of pitch to call from the pitcher at which location in a given situation or count is certainly important. Good chemistry between a pitcher and catcher is paramount. The skill of acting as a psychotherapist to a rattled pitcher during a mound visit to call him down and re-center him is undeniable.

The problem is there is no way to measure these things. Fangraphs calls it “the black box of catcher defense.” It is real. It is important. It is not yet quantifiable, so we ignore it.

2019 – The Return of the Hitting Catcher?

You may have noticed 2019 was included in the lists of top 5 catchers. Here is the 2019 deconstructed pie chart compared to 2018:

At first glance, hitting looks like it has a dramatically larger impact in 2019, but it is a work in progress. Because only half a season of games had been played (roughly 79 for each team on June 23, 2019 when the data was collected), eligibility requirements were reduced to 120 PAs and 240 IC, rather than 250 and 500. With the smaller sample sizes, wider variance is expected, particularly for categories like hitting. Jason Castro had just 129 plate appearances to base his hitting on, but had already accumulated in 2,149 framing chances to base his framing on.

As 2019 goes on, the sample size increases for hitting, we can likely expect the variance between the 90^th percentile hitting catcher and the 10^th percentile hitting catcher to decrease, and the pie chart will look more similar to 2017 and 2018.

That is unless the offensive onslaught of top end catchers continues. Gary Sanchez, who is having a blistering year at the plate, is the 90^th percentile catcher in 2019. His wRC+ in 2019 would have made him far and away the #1 hitting catcher in 2017 or 2018, yet in 2019 there are 3 other catchers with even higher hitting production rates (Garver, Grandal and Contreras). If that remains the case, the difference between catchers at the higher end and lower end of the spectrum and would continue to be great, and the impact of having a better hitting catcher would remain increased.

It is important to remember that this analysis considers the impact of each skill, based on the skill gap within the available catchers. 2018 was the worst hitting year for catchers in 17 years. If five Mike Piazzas enter the league, this would shift impact towards hitting. Halfway through 2019, we may be seeing this happen, but it may be prudent to reserve judgment until the season is over and a full season of data is available. (Update 11/12/19: Full 2019 Season Data is now available.)

Last Off the Field

If you skipped everything and scrolled down to the end, I don’t blame you. This was exhausting to write; I’m sure it was exhausting to read. For those of you who stuck it through, I hope you have found this analysis amusing, if not intriguing. Here are the Cliff Notes:

• A catcher’s WAR value is divided between five skills: hitting, baserunning, framing, throwing, and blocking
• The difference in impact between the best catchers and worst catchers is not equal for each skill.
• For 2018, the difference between having a good catcher and bad catcher in hitting and framing had much more impact than in any other skill.
• Good framing vs. poor framing has the same order magnitude of impact as good hitting vs. poor hitting.
• Skill at throwing out baserunners comprises only a small fraction of the catcher’s defensive impact
• The impact of having a good defensive catcher is accentuated by the fact that catchers are a weak hitting position. The relative impact of the hitting skill may increase if there is an influx of good hitting catchers at the top end, as we have seen so far in 2019 in a small sample size.
• I am obsessed with catchers.

---

^a The quote at the start of the article comes from Mark Harris’ 1956 novel “Bang the Drum Slowly.” It is not only my favorite baseball novel, but one of my favorite books of all time. It is less about baseball than it is about people and life, as a clubhouse learns over the course of a season that their catcher is diagnosed with Hodgkins’ lymphoma.

---

UPDATE 7/14/2019: A follow-up to this piece with full spectrum score data and catcher rankings for the 2016, 2017, 2018, and 2019 updated through the All-Star break, are now available at the following link: 2016-2019 Catcher Skills Spectrum Scores)

UPDATE 11/12/19: Completed 2019 Season Catcher Skill Spectrum Scores are available here