Earlier this spring, Pittsburgh Pirates pitching coach Ray Searage told the Tribune Review’s Rob Biertempfel that Jameson Taillon “has got to pitch at the top of the zone when he’s ahead in the count.”

As far as the Pittsburgh Pirates go, that was unusual.

For years the club has stayed fast and true to its mantra of inducing ground balls by working down in the zone. Sinkers, two-seam fastballs and sliders were all the rage as the Pirates backed up that philosophy with innovative – and more importantly, sound – defense behind their hurlers.

Could the Pirates be starting to inch towards more activity up in the strike zone? And if so, would that be an organizational philosophy, or one tailored to individual pitchers and their unique repertoires?

We will answer that question. Eventually.

But first, we’d love to show you a new way to look at zone effectiveness, borne completely by accident. In trying to find out why the Pittsburgh Pirates did not like to pitch up in the zone, we may have developed a new way to look at strike zone effectiveness. We’re calling it “ZEF.”

Editor’s note – this is a rather large post. We would highly recommend reading what ZEF is, but after that you can fast forward to specific sections by clicking the links below

| What is ZEF? | The Pirates and The Upper Third | Pirates ZEF Highlights | Team-level ZEF |

…ZEF?

Zone Effectiveness Factor, or ZEF, aims to develop one singular metric to gauge a pitcher’s effectiveness of pitching in any part of the strike zone. It is a weighted formula that relies on coefficients in vital pitching metrics to properly gauge effectiveness.

 ZEF=x(EV)+y(Whiff%)+z(DF)

 

To help me explain ZEF futher, I’m going to bring in Jeremy Wyne. You can read more about Jeremy below. His help has been instrumental in developing this metric.

The Core Components of ZEF

ZEF is comprised of three distinct components.

  • The average exit velocity of batted balls (EV)
  • The Rate at which batters miss when they swing at a pitch (Whiff%)
  • A “Deception Factor” that looks to accurately determine a pitcher’s ability to “fool” hitters. (DF)

Deception Factor is the number of called strikes divided by the total number of pitches taken by the batter.  A higher DF is obviously favorable for a pitcher and it quantifies his ability to deceive the batter into taking pitches that will land for called strikes. As with many pitching statistics, DF may be significantly influenced by catcher framing.  Umpire bias can also have an influence over small samples, but should even out over larger samples.

For each component, a normalized score is calculated to rank each team or pitcher such that a score of zero is MLB average.  If a pitcher has positive scores for each component, then he will have a positive ZEF which means he has performed favorable to the league average and the opposite can be said if he has negative scores for each component.

ZEF relies on the assumptions that lower exit velocities, higher whiff rates, and higher deception factors all contribute towards pitching effectively, either overall or within certain zones.  However, ZEF does not assume that each of the components are equal and have the same effect.  To assign weights, ZEF uses the least squares regression method so that the coefficients seen above (X,Y,Z) are chosen in a way that minimalizes the residuals between ZEF and WAR.

CORE ZEF COMPONENTS by weight

  1. Deception Factor
  2. Whiff %
  3. Exit Velocity

For 2016,  the X and Y coefficients were close in value while Z was approximately double the value of x or y.  Therefore, ZEF weighs deception factor highest followed by whiff rate and exit velocity.

We gave thought to determining the coefficients using metrics other than WAR.  For instance replacing WAR with xFIP would yield a similar result, but since FIP metrics only consider the true outcomes, the exit velocity component is de-valued since balls in play do not matter.  Likewise, the whiff rate component is exaggerated by the weights calculated with xFIP because strikeouts make up a large portion of the true outcomes that are considered.  For these reasons, it was determined that the coefficients derived from WAR would be more representative of overall effectiveness.

Just like each component score, the league average ZEF at the MLB level is zero.  ZEF does not consider the number of pitches thrown so small sample sizes can lead to extreme values, but over the course of a season’s worth of pitches the values level out.  Similarly, ZEF calculated for teams’ pitching staffs fall within a tighter range compared to ZEF calculated for each individual pitcher in the league.

ZEF can be easily read in much the same way as WAR, albeit with a ‘look and feel’ that reads more like AVG. Figures can be either negative or positive, with zero representing replacement level for WAR and MLB average for ZEF.

A quick note on sample size and type of pitches. While pitching in the upper third more often may be employed as a strategy by some teams, the sample size of pitches there still lags far behind other zones. We also cannot discount the presence of “mistakes.” Often, pitches end up in the upper third when they were clearly not meant to. However, we can point to the difference in fastballs and offspeed pitches as an example of why this may not occur as much as one might think.

At an individual level, the average percentage of fastballs landing in the upper third – 11.56 percent –  was nearly double the amount of offspeed pitches – 5.96 percent – landing there. While mistakes happen on fastballs too, a counter argument could be made that more purposeful pitches such as moving fastballs, cutters, etc. result in less mistakes.

The Pittsburgh Pirates, the Upper Third, and ZEF

To bring this back to a Pittsburgh Pirates focus, we developed ZEF partially to determine how well the Pirates could pitch in the upper third of the zone, with the upper third being defined as zones 1,2,3 from pitch F/X. One important note – Pitch F/X uses the called  strike zone – not the rule book strike zone.

As we said at the top, the club has been reticent to throw in the upper third, as evidenced by their ranking of 29 out of 30 MLB teams in terms of percentages of their pitches that land in the undesirable zone in 2016. Statcast queries were not available for years prior to 2016 at the time of this writing to compare year-to-year.

The story behind the Pirates and the upper third is an ugly one. For 2016, the team ranked dead last of all 30 MLB teams with a ZEF of -.829. The team that utilized the upper third most efficiently was by far the Los Angeles Dodgers – with an overall U3 ZEF of .514.

As luck would have it, the Pittsburgh Pirates and Los Angeles Dodgers are mirror images of each other not only in overall ZEF, but also find themselves on the ends of the spectrum in fastball and offspeed ZEF, as you can see on the sidebar.

Are the Pirates going through a philosophical change that will ripple throughout their entire pitching staff, much like the groundball revolution did several years ago?

Probably not. However, the club might just be looking to add a new wrinkle. For a team like Pittsburgh, a modest increase in any part of the strike zone would have amplified effect after years of just the opposite.

We also cannot discount that a potential change in Upper Third philosophy could be tied to the defensive outfield changes the club has put in place for 2017. One could easily point to the Tampa Bay Rays – and their all-world defensive CF Kevin Kiermeier – as taking a similar approach.

Pittsburgh Pirates ZEF Highlights

Looking beyond the ZEF figures for the Pittsburgh Pirates, we can start to pick out a few oddities, a few trends, and one undeniable truth.

Oddities

  • Raise your hand if you thought that Antonio Bastardo would have the best overall ZEF rating of any pitcher with the club in 2016. Not so fast, everyone. Bastardo is one of only two Pirates pitchers with positive scores on both his fastball (.562) and offspeed (.120) offerings. With a U3 rate of 10.22 percent, he is very effective.
  • Going back to the original impetus for this post, Taillon was not very effective at all with his offspeed pitches in U3, carrying a -.425 ZEF. Taillon carried a fantastic DF of .833, but was penalized by inducing just 11 whiffs out of 105 total swings at U3 pitches. With an exit velocity of 90.1 mph, his ZEF rating suddenly makes sense. It will be very interesting to see what Ray Searage can do to maximize his off speed upper-third effectiveness.
  • Generally speaking, fastballs play better in the upper third than off speed, but not so for Felipe Rivero. The young fireballer threw 6.02 of his off speed pitches in the upper third, which is actually a high rate for the slow stuff (more on that later). He rode a solid .846 DF to a .246 off speed ZEF.

Trends

The most telling trend in looking at Pittsburgh Pirates ZEF data comes from players that dont’ even play for the team.

The club’s two most strongly linked offseason  targets – Jose Quintana and Derek Holland – had interesting ZEF figures.

Pirates Offseason Targets' ZEF

 U3 ZEFU3 Fastball ZEFU3 Offspeed ZEFUsage
Quintana.818.741-.0619.99%
Holland-.365-.512.20210.67%

Holland had a DF of just under .900 on his offspeed, which carried him to a solid score there. His fastball did not come out well in the wash, but tht was mainly due to a lack of whiff percentage, at just 17 whiffs out of 119 swings.

Quintana, on the other hand, is a more exciting U3 pitcher. A whopping 51 out of his 55 takes in the upper third were for strikes. He had a decent whiff rate of 21 percent in U3, and he topped it off with just an 85.6 exit velocity. Using 13 percent of his total fastballs in upper third, he also had a high usage. His offspeed stuff is belied by an unusual whiff percentage – zero (no whiffs on 20 swings), but a fantastic DF of .913.

It is simply no wonder why the Pittsburgh Pirates pursued him with such vigor.

An Absolute Truth

If there’s one undeniable truth in all of this, it is that Gerrit Cole should stay far away from the upper third.

Cole had the worst overall ZEF of any pitcher in 2016. Okay well maybe that’s not true. He was clearly the worst pitcher in U3 with any serviceable sample size at 9.85 percent usage. He ended up with a -1.93 ZEF for his efforts.

Interestingly enough, Cole’s U3 offspeed (-.469) played a bit better tan his U3 fastballs (-1.63). Cole’s high heat was hit at just about average with 90.6 mph, but he had no deception or ability to generate swings and misses. He carried just a .50 DF to go along with 14 whiffs out of 126 swings (11.1 percent).

One thing we can say about the Pittsburgh Pirates’ number one starter is that he is capable of better. In 2015 – by far Cole’s best season – his fastball played much better in the zone. It was tagged at just 87.9 mph, carried a 21.4 percent whiff and had secent DF at .735.

We’ll wrap up by focusing on some team-level, MLB-wide findings.

Team-Level Upper Third Findings

When we look at the MLB-wide, team-level Upper Third ZEF ratings for 2016, we can start to see some trends emerge.

First, we will take a look at upper third usage. Generally, teams ranged from a percent usage of around eight percent to just under 12 percent.

Of the MLB’s 30 teams,

  • Eight teams had usage of 10 percent or more
  • 12 teams had usage ranging from 9-10 percent
  • 10 teams had usage below nine percent

With a ZEF rating of 0 being considered league average, 21 teams found themselves with an above-average ZEF rating for pitches thrown in the upper third in 2016. This leads us to our first substantial finding – only one club with above-average ZEF in U3 had less than nine percent usage.

That team – the New York Yankees – had an 8.38 percent usage rate, and was barely better than average at .0013.

From our research, all teams with usage of 10 percent or more found themselves in the top 18 of MLB in ZEF. Two of the three teams that had usage with 11 percent or more joined the Dodgers in the top three by ZEF.

Those were the Detroit Tigers – third in usage (11.01 percent) and second in ZEF (.360) and the Boston Red Sox – First overall in usage (11.81 percent) and third in ZEF at .323.

 

A Magic Number?

Aside from the New York Yankees and their barely-under-nine percent usage, each team that had less than nine percent usage found themselves with a below-average ZEF. It would be tempting to say that we can now state that nine percent seems to be a de-facto cutoff for effective upper-third pitching. Alas, that is not a perfect conclusion.

For one, the team with the second-most usage – the Tampa Bay Rays at 11.59 percent – ranked just 13th in ZEF with a .214 rating.

Second, based on a few real-world examples, we can conclude that effective team-wide pitching in the upper third – and any part of the strike zone, really – can be seen in more than a few ways.

Our first example comes from those same Tampa Bay Rays. As the indefatigable Travis Sawchik recently wrote on FanGraphs, the Rays have encouraged pitchers such as Jake Odorizzi to use the upper third more. The 27 year-old right hander can tell us a lot about what drives ZEF and therefore, effectiveness.

Odorizzi is something of a ZEF enigma, as his overall figure is sub-par at -.432 with a 15.24 percent usage overall.   Odorizzi’s sample size of pitches in U3 for 2016 is significant, in fact it is the highest overall usage rate by any pitcher in 2016.

Despite that, Odorizzi did get tagged for a 91.7 MPH exit velocity on those pitches that were put in play, and had a lower deception factor of 76.6 percent.

So, why do the Rays continue to have him throw there?

A quick look at a scatter chart of team’s batting average against in the upper third might help us figure out why.

Upper Third Usage x Exit Velocity, with Batting Average Against

Here we see team usage from 2016 for zones 1,2,3 (U3). We also see team’s exit velocity on the left. The colored dots represent batting average against (blue = low batting average; red = high).

We picked out a few teams here to illustrate how the Rays approached their value of Odorizzi’s U3 usage. Simply put, they don’t care if Odorizzi’s upper third pitches get offered at because they are usually converted to outs.  In this way, we shouldn’t really care about the Rays’ ZEF rating, because they care more about producing outs than strikes.

Still, this gives off a “playing with fire” vibe. Overall – regardless of zone – the Rays had a BABIP of .297, good for ninth-highest in the American League. With a lower-than-most deception factor and whiff rate among their top 15 ZEF brethren, the Rays might start to see more of those fall in for hits.

On a side note, though the Dodgers have the best ZEF rating, Boston’s usage of the upper-third is equally impressive. Not only did they throw the most pitches in U3 by percentage, but they also induced the softest contact. They boasted solid DF (.869) and Whiff Rates (20.7 percent) as well. They might have been the total package if the Dodgers did not have a better DF (.889).

Here now is a chart showing the ZEF trends against percentage usage by teams.

So, in 2,000 some odd words prior to this sentence, that is Zone Effectiveness Factor, or ZEF.

Future Applications of ZEF

It is our sincere hope that others may join us in helping to dissect this ZEF data. On our part, we are going to be watching this throughout the 2017 Pittsburgh Pirates season. We’ll be looking at it from all angles, compiling data in many different scenarios, from certain pitch counts, to runners on base, ahead & behind, home & away and much more. For each pitcher. All season.

If you would like to take a look at all of our compiled Upper Third ZEF data, including stats for all MLB pitchers who registered 25+ pitches in that part of the strike zone in 2016, CLICK HERE for our publicly available spreadsheet.

Acknowledgements

If you’re still reading, I can’t thank you enough, but I will start my thanks with you. So, Thank You!

As stated Above, this post would not have been possible without Jeremy Wyne, and a mere thanks would not be enough. Here now is a bit more about Jeremy, who we hope to rope into a couple of large projects like this more often!

Jeremy Wyne (Twitter: @WaynoTexino) has been a baseball fan since listening to games on the radio with his great-grandmother as a preschooler.  Using data to explain and better understand the game has always been of interest to him.  He is an alumnus of Penn State and Carnegie Mellon Universities.  Jeremy works as a software engineer in the South Hills of Pittsburgh where he resides with his wife and son.

 

 

Jay Obstarczyk – brother of our own Ethan Obstarczyk – is a graphic designer living and working in Pittsburgh. He provided the sidebars seen in this post as well as many other graphics for PBD, including our podcast logo and branding. Major thanks to him for taking on a project at short notice. Please visit his portfolio.

Jason Rollison

Jason Rollison has been analyzing baseball and the Pirates in one way or another for 4+ years. Jason's previous stops include rumbunter.com, Pittsburgh Sporting News, Call To The Pen and several print publications. He also covers the State College Spikes for the Centre County Gazette (State College, PA) When it comes to analyzing baseball, he likes to take a middle-of-the-road approach, with one foot on the analytics side of the fence and the other on the old-school side. Having said that, he is a sucker for pitchf/x. Jason has appeared as a phone-in and in-studio guests in numerous outlets, including Trib Live Radio and 93.7 The Fan (CBS Sports Radio)