A Ball-in-Play Analysis of Nine Free-Agent Pitchers

Yesterday, I published a post here examining the year-to-year correlations for a number of ball-in-play (BIP) pitching metrics. The results published there have some use on their own. As a practical application of that information, however, I’d like to take a look here at the nine potential free agent starting pitchers (besides Zack Greinke) whose teams faced a qualifying offer decision this offseason, and see how this analysis might impact their valuation.

Below, you’ll find two tables. The first, for reference, is a collection of all the year-to-year correlations from yesterday’s post. After that is a second table, featuring how each of the nine pitchers fared according to each metric. Comments regarding each pitchers in greater depth appear below that.

First, the correlation coefficients from yesterday’s post:

Correlation Coefficients, 2014-15 ERA Qualifiers
Metric Coefficient
K% 0.81
BB% 0.66
Pop% 0.53
Fly% 0.76
LD% 0.14
GB% 0.86
FL/LD 0.37
GB AUTH 0.25
ERA 0.45
FIP 0.65
TRU ERA 0.72

Remember: a 100% correlation (1.00 in the above table) is obtained when the two sets of data are totally identical. The closer to 1.00, the higher degree of correlation between the two data sets.

Now what follows are the nine pitchers in question. Stats are presented as an index, where 100 is average, above 100 is above average, and below 100 is below average. Questions about the various metrics? A more thorough explanation can be found in yesterday’s post.

2014-15 QO Candidates – Key Stats Scaled to 100
Brett Anderson 77 82 12 60 72 145 127 96 98 95 101 102
Wei-Yin Chen 97 71 153 109 96 92 98 102 100 83 104 89
Marco Estrada 91 104 168 148 74 73 70 89 74 78 110 75
Yovani Gallardo 77 118 65 84 105 111 97 108 96 85 100 102
His. Iwakuma 108 56 59 93 88 114 111 110 100 88 93 82
Ian Kennedy 121 98 104 119 108 84 114 103 121 110 116 101
John Lackey 97 79 133 98 97 100 92 111 96 71 92 91
Jeff Samardzija 90 74 111 115 101 88 90 103 102 124 105 94
J. Zimmermann 98 63 154 106 103 92 122 101 105 94 96 95


Brett Anderson (Profile)
Status: Accepted qualifying offer from Dodgers

Anderson’s key strength is his stratospheric grounder rate, the highest in the majors last year, over two standard deviations higher than league average. His walk rate was also a positive, over one half standard deviation lower than league average. Our correlation coefficients tell us that these are likely true talents, and should recur moving forward. Unfortunately, the same doesn’t apply to his low liner rate, which was over two standard deviations below league average. As much of a ground baller as Anderson is, repeating his 66.3% grounder rate would be quite a feat, and any reduction likely will translate point for point in an increase in liner rate.

His chief weakness is his low K rate, and our correlation coefficients tell us he can’t expect much help going forward. About the only way such a low-K pitcher can survive is to have a defining BIP frequency and/or authority talent; Anderson’s grounder-inducing skill is just that. It must be emphasized, however, that this elite grounder skill simply made him a league average starter. Toss in the fact that, due to injuries, 2015 was the first time in six years that Anderson was an ERA qualifier, and it’s tough to see how Anderson improves from here. He was wise to accept the QO; he needs to prove he can stay healthy before any club will invest multiple years in him.


Wei-Yin Chen (Profile)
Status: Free Agent

Chen’s chief strengths are his high pop up and low BB rates, which were both over one full standard deviation better than league average. The correlation coefficients tell us that these are true talents, highly likely to recur moving forward. In addition, Chen is one of the younger members of our group, and also one of the most reliable. He has never reached 200 innings in a season, but has consistently taken the ball every fifth day and averaged six innings per outing in his four seasons in the majors.

While Chen has yielded a large number of fly balls, he has consistently managed their authority near a league average level. FIP doesn’t like high fly ball rates, and dings Chen accordingly. “Tru” ERA gives him credit for his ability to contain damage in the air. Our correlation coefficients tell us talent is involved to some extent here, though randomness is as well. Chen would not appear to have great upside above his current performance level, but his floor and his reliability level are high. He deserves at least three years on the open market, at an annual rate a bit south of the QO level.


Marco Estrada (Profile)
Status: Re-signed with Blue Jays

Estrada inked a two year, $26 million deal just prior to the QO deadline. He was my AL contact manager of the year: his Adjusted Contact Score mark of 74 (in the BIP PRD column above) was best in the AL. His very high pop up rate is his chief strength, and it is one he has maintained throughout his career. Also, his relative production allowed mark on fly balls/liners of 70 is off-the-charts low and drove his 2015 success. As Brewer fans, and our correlation coefficients, can tell you, that figure has varied a bit from year to year.

Estrada’s liner rate allowed, like Anderson’s, is incredibly low, and should be expected to regress moving forward. Though it has now been quite low for three years running, there’s only so much talent at work here. Estrada’s rate also crossed over into below league average territory in 2015, which is a bit concerning moving forward. FIP doesn’t appreciate Estrada’s contact management skills, which are very real. That said, 2015 was a best case scenario for Estrada, a first-time ERA qualifier at age 31. The Blue Jays played his situation well; locking him for two years below the QO annual rate is a solid risk/reward play.


Yovani Gallardo (Profile)
Status: Free Agent

The most positive item here is Gallardo’s high grounder rate; this has been a staple throughout his career, and 2015 represented a career best relative to the league. The grounders aren’t going away. His other positive, on an actual production-allowed basis, was his containment of fly ball/liner damage; when adjusted for relative BIP authority, however, his performance slides much closer to league average. Unlike Estrada’s case, FIP gets it right this time: Gallardo’s good fortune on fly balls caused his 2015 ERA to be much lower than his FIP and “tru” ERA.

Now, the negatives. Steadily declining K rate over the last few years, sharp increase in BB rate in 2015. Our correlation coefficients tell us not to count on imminent reversal of these trends. Though Gallardo has answered the bell consistently over the years, his quality is unlikely to match his innings quantity moving forward. He’s a league average pitcher in a near best-case scenario in any year moving forward. Dollars per year should be a bigger concern than length of contract from a team perspective. I might be comfortable with three years, but I’m not cozy once the bidding gets past $10 million per year. He’ll likely get more than that, and it probably won’t end well.


Hisashi Iwakuma (Profile)
Status: Free Agent

There are a couple of fairly obvious strengths here: BB and grounder rates well above average, by over one full and one half standard deviation, respectively. Those are true talents, and important ones. Iwakuma’s K rate is also solid, near the top of the average range among starting pitchers. A very low liner rate also helped drive his 2015 success, but as we can see from our correlation coefficients, we should not expect this to persist moving forward.

On the negative side, Iwakuma does allow all types of batted balls to be hit harder than the league average, though his strong frequency mix kept his adjusted contact score in check at 100, the league average. Qualitatively, Iwakuma is at or near the top of this group. Nice K/BB profile, lots of grounders… good stuff. The big issue here is advancing age and reliability. Going back to his days in Japan, he has been susceptible to nagging injuries on an ongoing basis. A third guaranteed year at $15-plus million gets him done, though teams will likely try to work in a third-year vesting option instead.


Ian Kennedy (Profile)
Status: Free Agent

Ah, the guy many thought should take the QO. His positives and negatives are in-your-face obvious. His K rate is by far the best of this group, and one of only two, along with Iwakuma, measurably above league average. That’s his lone plus, but it’s a huge one, and one that our correlation coefficients tell us is likely to persist.

On the other hand, boy, do guys hit the ball hard against him. His frequency mix is far from ideal, with high fly ball and liner rates. While liner rates are usually random, Kennedy’s liner rate allowed has been higher than league average in four of the last five seasons. There has to be some true talent, or lack thereof, at work here. His 121 Adjusted Contact Score was the worst among NL ERA qualifiers in 2015. His contact management skills are bad, but not 121 bad; I’ll mark him for true talent of 110, which makes him a durable, slightly better than league average starter right now. He’ll remain so until his K rate gives way. A club in a big park with a good defense should pony up three years for Kennedy, who could be a bargain around $10-12 million per year.


John Lackey (Profile)
Status: Free Agent

There’s a lot of near-league-averageness in Lackey’s profile, aside from his one-half standard deviation better than league-average BB and pop up rates. This is probably a pretty good time to bring up the fact that, in these circles, league average is pretty good. Were you aware that the average NL ERA qualifier struck out 8.06 batters per nine innings last season? In these days of less contact, fewer walks and more ground balls, the stakes have been raised for starting pitchers. That Lackey, uh, lacks a glaring weakness at his advanced age is quite impressive.

That said, Lackey was nowhere near as good as actual ERA last season. His FIP and “tru” ERA are in agreement: he’s a near-average contact authority manager with a slightly better than average BIP frequency mix. At his age, he should not be expecting more than two guaranteed seasons, but the high quality he brings could land him $16-17 million per season. Given the Cards’ needs and Lackey’s success in St. Louis, expect the two sides to come to a meeting of the minds.


Jeff Samardzija (Profile)
Status: Free Agent

There are an awful lot of conflicting signals in Samardzija’s profile. The big positive is his low BB rate, over a full standard deviation below league average and the third best among of this group of pitchers. Other than that, his biggest pros aren’t found on a stat line. It’s his durability, raw stuff, size and athleticism that set him apart, and keep teams believing there’s more in his tank.

On the negative side of the ledger, the precipitous drop in his K rate last season stands out. Our correlation coefficients make us wonder if he can recapture his past prowess in this area. In addition, Samardzija’s frequency profile took a turn for the worst last season, as his fly ball rate spiked and grounder rate dropped. His actual 2015 performance flew in the face of projected correlation with prior year performance. The biggest final takeaway is in the last three columns of the above table, however. His relative “tru” ERA of 94 easily outstrips his relative FIP (105) and his abysmal relative ERA (124). He was killed by the poor White Sox defense. Even with the negative developments in his frequency and authority profiles, Samardzija “should” have been an above average starter in 2015. With some targeted improvement, his ceiling remains very high. Expect a four year, $60-70 million deal for the big guy.


Jordan Zimmermann (Profile)
Status: Free Agent

The 2015 version of Jordan Zimmermann was a bit like 2015 Jeff Samardzija. Zimmermann’s most outstanding trait was his low BB rate, over a full standard deviation below league average, and better than any pitcher on this list. Unlike Samardzija, this has been Zimmermann’s trademark throughout his major league career. In addition, he has developed a significant pop up tendency, also with a rate over a standard deviation better than the league.

As for negatives, Zimmermann allowed fairly hard fly ball/line drive contact (122 Adjusted F/LD Contact Score), which fortunately, our correlation coefficients tell us isn’t solely tied into true talent. After taking his relatively strong BIP mix into account, his overall Adjusted Contact Score is a much more tolerable 105. His true talent in that area is likely closer to league average, which will likely move his “tru” ERA closer to 90 than his 2015 mark of 95. He’s young, consistent, has a very high floor and a decent ceiling. He has a chance for a five-year deal, and should earn between $16-18 million per year.


An awful lot more than the numbers discussed here come into play when making a substantial free agent commitment. There’s health, makeup, athleticism, etc. Still, at a bare minimum, as a club enters this process, they must know which numbers on a free agent’s record are most likely to recur, or even improve, moving forward. Bad deals abound in free agency, where you must remember the golden rule: pay for what the player is going to do, not what he’s already done.

Newest Most Voted
Inline Feedbacks
View all comments
Ray R.
8 years ago

“[A] 100% correlation (1.00 in the above table) is obtained when the two sets of data are totally identical.”

I guess I’ll be the nit-picky jerk that points out that this statement isn’t necessarily true.

8 years ago
Reply to  Ray R.

No, the statement is perfectly true. You will always have a correlation of 1.00 when the two sets of data are identical.

However, that is not a *necessary* condition for correlation, which I assume is your argument?

Ray R.
8 years ago
Reply to  someone

Correct. The quote I supplied suggests that a PPMC coefficient of 1.00 implies that the two data vectors are identical, which is not necessarily the case. The quote should be amended to state that there are other scenarios under which a “perfect” correlation can result.

picknit of nitpick
8 years ago
Reply to  Ray R.

The quote actually says that a coefficient of 1.00 is obtained when the vectors are identical . This is ,in fact “necessarily” true (another quote ) . the post is silent on whether or not a coefficient of 1 can be obtained under other conditions.