clock menu more-arrow no yes mobile

Filed under:

Josh Johnson, Carlos Zambrano, and Defense-Independent Pitching

MIAMI, FL - JUNE 11:  Josh Johnson #55 of the Miami Marlins pitches during a game against the Boston Red Sox at Marlins Park on June 11, 2012 in Miami, Florida. The Marlins defeated the Red Sox 4-1.  (Photo by Sarah Glenn/Getty Images)
MIAMI, FL - JUNE 11: Josh Johnson #55 of the Miami Marlins pitches during a game against the Boston Red Sox at Marlins Park on June 11, 2012 in Miami, Florida. The Marlins defeated the Red Sox 4-1. (Photo by Sarah Glenn/Getty Images)
Getty Images

As we got into the season, i have pulled back on sabermetrics tutorials, but I figured with the occasional slow news day, I would introduce a few more primers and tutorials on sabermetrics and the sorts of numbers and analysis that we use on this site. Every once in a while, when a concept seems relevant, I will touch upon it and use the current Marlins situation as a basis for discussion.

Note: This is a fairly lengthy discussion, but a good one and one that does not involve math. Readers have been forewarned. Please proceed.

Today, the Fish will attempt to halt their June slump by throwing Carlos Zambrano out to the mound against the Tampa Bay Rays, as the Fish start a nine-game road trip with three games right up I-75. Up until his last start, Zambrano had been everything we could have asked and more for the Marlins. Even after his worst start of the 2012 season, Zambrano still owns a 3.55 ERA, which is better than any of us could have expected.

Meanwhile, Josh Johnson's latest start only recently signified the potential of his return. Even with his dominant start, he still has just a 4.27 ERA through 13 starts. Prior to that, his pitching had been questioned, in part due to a lack of velocity.

These results may be opposite of what one would expect going into the season. But the truth is that they actually have performed almost to what we guessed, at least when you look at their peripherals.

2012 18.3 10.3 1.9 3.55 3.97 .244
ZiPS Preaseason 19.7 10.9 2.2 4.20 4.12 .307

2012 20.1 7.0 0.9 4.27 2.79 .360
ZiPS Preaseason 23.2 6.7 1.3 2.66 2.71 .296

You can see that in both playes, their expectations are being closely met by their peripheral performance. And the reason why Johnson has such a high ERA compared to his projection and Zambrano such a low one? A matter of hits on balls in play, the sort of thing that pitchers have little control over.

DIPS Theory

The defense-independent pitching theory is simple in construction and revolutionary in impact. It may be one of the most important things that sabermetrics has ever contributed to baseball. Originator Voros McCracken found this simple concept many years ago and summarized it cleanly in this sentence:

There is little if any difference among major-league pitchers in their ability to prevent hits on balls hit in the field of play.

It sounds like blasphemy, doesn't it? All of our lives, we have been trained to credit the pitcher for almost everything good or bad on the field, even though we still talk about amazing defensive plays being made or mishaps in routes leading to balls falling into gaps that should not have existed. I mean, there is little difference between how the best pitchers in baseball and the worst allow hits? That cannot be right!

But before you rant, allow yourself one little mental exercise. Think of the best pitchers in baseball in the 90's. Maybe you thought of Greg Maddux. Maybe Roger Clemens. Maybe Randy Johnson. Maybe Pedro Martinez. Those are all great names right? Now, think of some of the worst in the 90's. Now, this may be a little more difficult, so I enlisted some assistance. I found the four worst ERAs among pitchers with at least 1000 innings from 1990 to 1999, getting names like Ricky Bones (4.86 ERA) and Willie Blair (4.94 ERA). In that same timespan, by the way, Maddux, Clemens, Martinez, and Johnson were four of the top five ERAs. How did they fare in batting average on balls in play (BABIP) for their careers?

Roger Clemens 3.12 .286
Randy Johnson 3.29 .295
Greg Maddux 3.16 .286
Pedro Martinez 2.93 .282
Willie Blair 5.04 .306
Ricky Bones 4.85 .288
Mike Moore 4.39 .286
Bobby Witt 4.83 .307

That is an amazing finding. The difference between the lowest BABIP from one of the best pitchers in baseball and the highest BABIP from one of the worst is a matter 22 batting average points. In 400 balls in play, that is a difference of 10 hits. Pedro Martinez averaged about 500 balls in play per 200 innings. The difference between his BABIP and Bobby Witt's .307 BABIP in 500 balls in play is 12 hits. Twelve hits in 200 innings equates to about eight runs of difference between the best and worst in a given season based on balls in play.

That above example shows you the power of defense-independent pitching. The truth is that, as McCracken noted many years ago, there is not a lot of difference in major league pitchers in terms of their ability to control hits on balls in play. For whatever reason, most major leaguers are pretty much as good as each other in that department. Pitchers like Josh Johnson, who were elite starters in the game, have allowed high numbers of hits on balls in play in one season and low numbers in another. Worse starters like Carlos Zambrano, who is not Johnson's league in terms of ability, can post seasons in which they hold batters to low averages on balls in play one year and allow too many hits the next. No pitcher is immune.

Now, this is not meant to say that pitchers do not control their outcomes on balls in play. One need only imagine him or herself on the mound before realizing that there is no way he or she would allow a .300 average on balls in play. This effect likely stems from a simple, understandable fact: as you players progress through various levels of play, they are selected heavily for this trait. That is, the pitchers who allow too many hits on balls in play fall by the wayside at earlier levels, whether they be high school or college or professional levels. This is likely because this skill is integral to success in the majors. By the time they reach the big leagues, the selection of major league pitchers has been so strong that no pitcher can differentiate himself from the others in that regard.

So there is skill involved in allowing hits on balls in play, that much is obvious. But at the major league level, that skill difference is so small that it is hardly visible amid the other aspects affecting balls in play. Remember, what is affecting a pitcher's balls in play? Not only is it the pitcher's performance, but also the hitter's, the defense's, and external environmental factors. That is four different aspects clashing to determine the result on a ball in play. It should be obvious that, with all those factors, such a number can fluctuate from year to year, and that over time, pitchers will end up receiving very similar results because those other factors around them will even out. Defenses come and go, pitchers face a variety of different hitters of different skill levels, and you play all over the country in different environments, and all of that evens out. What you are left with in the long term is the pitcher's skill, which as you can see is not able to differentiate Greg Maddux from Ricky Bones.

How do pitchers differentiate themselves? How do we get from Bones to Maddux? The factors of pitching that do not involve defense. Why are strikeouts so revered and walks and home runs so reviled? Because those are things over which pitchers have the most control. In those situations, only the pitcher, hitter, and environmental factors are involved, as the defense is completely shut out of those outcomes. As a result, we see that strikeouts and walks are highly correlated year-to-year, and home runs also correlate well, though less than the other two.

Johnson, Zambrano Heading in Other Directions

What does that mean for Josh Johnson and Carlos Zambrano? It means that, when we see such discrepancies in BABIP, we can safely assume that these will normalize on their own over the long run. Johnson's .360 BABIP is expected to go down, simply because we know he is good enough to allow a league average BABIP based on his career and the fact that few major leaguers ever are capable of consistently outperforming or underperforming in that skillset. Similarly, we would expect Carlos Zambrano's BABIP to go up as we go forward. This means their respective ERA should head towards their FIP, as FIP is an estimate of ERA based on just strikeouts, walks, and home runs.

Indeed, we have already begun seeing this phenomenon happen this season for Johnson. In the month of April, batters hit .446 against him, but since then they have hit just .310 on balls in play. Compare that .310 mark to Johnson's career .302 BABIP. Quite similar, right? For Johnson, regression has already begun, and it should be on its way for Zambrano, much to the chagrin of Marlins fans.

This does not meant that Josh Johnson will necessarily be amazing going forward or that Zambrano will be terrible. It simply means they will be returning to their norms very soon, as all pitchers do in terms of balls in play. Zambrano has performed well in terms of his peripherals and is much improved from last season. Johnson's strikeout rate has dropped, but his walk rate has fallen too, so he should be close to his marks from the last two seasons. But the ERA you see for both pitchers right now is a combination of good or bad fortune and good or bad pitching. Over time, that fortune and that pitching evens out pretty well for starters, and thus we can expect them to approach their expected pitching performances sooner rather than later.

To be able to say that, you need the power of defense-independent pitching. It is a revolutionary finding that deserves attention and credit even to this day. This idea changed the way baseball is viewed, and should be treated as such.


- There were plenty of resources used, but the basis was from this 2007 David Gassko article from The Hardball Times and this 2001 Baseball Prospectus piece by Voros McCracken. Thanks to both for the important information.