Can Pitchers Prevent Solid Contact?

Dallas Keuchel has seen his velocity drop and groundball rate drop this season. (via Arturo Pardavila III)

Dallas Keuchel has seen his velocity and ground ball rate drop this season. (via Arturo Pardavila III)

Is it possible pitchers have a trait to give up hard contact or prevent hard-hit contact? It was a considered a pitching skill for many years until Voros McCracken discovered batting average on balls in play and showed that quite a bit of luck occurs with balls in play. Since 2002, more data has become available — be it from Baseball Info Solutions, Brooks Baseball, Statcast, Inside Edge, PITCHf/x, and many more. With more and more data becoming available, it seems like we should be closer to the answer of determining if a pitcher can give up harder or softer contact. The problem is that we are not. Truthfully, I don’t think solid contact will ever be usefully projectable for the simple reason that pitchers change too much too quickly.

The reasons pitchers change are many. They add or subtract a pitch. They throw their pitches with a different mix. The throw hard or softer. They change the shape of their pitches, or the location. They change a lot, is the point. Let’s dig into some details.

Let’s start with some terminology. For the sake of limiting confusion, I am going to call the trait I am looking for “Solid Contact.” Both Baseball Info Solutions (BIS) and Inside Edge (IE) track “hard contact.” So Solid Contact will mean the overarching idea, while Hard Contact will reference BIS data.

The next major hurdle is defining Solid Contact. I constantly hear and read about a pitcher being hittable and allowing line drive/hard-hit/solid contact. Before I wrote this article, I asked several sources for a definition, and I got a ton of answers. Line drives. Home runs only. All extra-base hits. BIS’s or IE’s Hard Hit rate. Even just BABIP. None of the answers seemed right, but none seemed wrong either. I always thought the perfect answer was out there once the right data could be collected. I decided to look at the following categories: BABIP, LD% (BIS), Hard Hit% (BIS), ISO (2B*2*3B+3*HR) and HR/contact.

Here are some thoughts on each.

  • BABIP: If I hadn’t asked around, it wouldn’t be included. Too many of its inputs are based on infield defense and runner speed to be a good measure. I will examine it for others, but I don’t like it as a Solid Contact measurement.
  • LD% (BIS): At one time, line-drive rate was thought to be the Holy Grail to measure pitchers. Once it was measured and reported, it became the one of the slow stats to stabilize along with BABIP.
  • HardHit% (BIS): What better to measure Hard Contact, than with Hard Hit rate? If it only correlated at all from one season to the next. It takes just as long as LD% and BABIP to stabilize. Comparing even and odd seasons, I found it takes around 2800 BIP for a hitter’s hard-hit rate to stabilize with an r-squared at .50.
  • ISO (2B+2*3B+3*HR): Now we are getting to the two underutilized stats for pitchers, ISO against and HR/contact. If the goal is to find Hard Contact, what stats better shows the results than the pitcher’s ISO? While I do like it, I will give it a nice facelift later.
  • HR/BIP: We know it usually takes quite a hit to knock a ball out of the park. I like using home run per batted ball as it shows the likelihood of contact causing the home run. Home run per nine is influenced by strikeout and walk rate, while home run per fly ball also needs flyball rate to find the overall rate.

There’s also the new StatCast data. I didn’t forget about it, but Russell Carleton beat me to the punch, and I had to cut off part of my work and throw it away like a moldy pizza slice. Recently at Baseball Prospectus, he looked to find if the StatCast data can stabilize and how quickly. Our numbers were similar, but he was able to do the work with a cleaner data set. My recommendation is to go read the entire article, but he came across the same smoking gun I found with the with the other Solid Contact data.

I’d suggest that we’re seeing something similar for pitchers and the time that it takes for that true talent to wander around is a lot less than you might think. We know that pitchers do get better and worse as they develop and age, but maybe those developments are less linear and more rapid than we thought. It is possible – and according to these numbers, common! — that while a pitcher might have had a good April, by June he could be a different pitcher. What’s strange is that we’re not seeing that these numbers are unreliable. In that case, we might say that exit velocity allowed is all chance, sorta like BABIP. In fact, in small doses, exit velocity is quite reliable.

This might be a little breakthrough in understanding the mystery of DIPS. Maybe the problem was that we were conceptualizing the problem wrong. We assumed that pitchers should be the same throughout a year and that more data were better. If performance wasn’t correlating, then it must be a function of luck, rather than rapid, but real fluctuation in talent level. These findings suggest we’d do better looking into understanding how a pitcher changes within a season – maybe within a month – if we really want to understand him.

I came to the same conclusion using the more common pitcher stats.

I tried to combine various variables, giving each one a different weight. I spent way too much time looking for an answer we may not be able to find without adjusting our thinking. The more I look at pitcher batted-ball data, the more I feel this final frontier may never be conquered. After examining too many pitchers, I have a good idea why it takes so much information. As I said before, pitchers just change too quickly. Mr. Carleton was right on track.

Let’s assume a starting pitcher enters the league for his age-24 season. It will take him eight seasons to have his BABIP stabilize. Over those eight seasons, he will likely see his fastball lose two mph. Additionally, normal aging dictates that a pitcher’s ground ball rate drops 1.5 percent.

These are the effects of “normal” aging. Beyond that, pitchers are constantly messing with new pitches, changing the mix of pitches, and the location of the pitches. For now, let’s concentrate on the effect of two the factors — groundball percentage and fastball velocity. I like to use velocity and groundball rate because they stabilize quickly, and their values can help to predict the changes most likely seen in a pitcher.

Since 2002, which is the first season for which FanGraphs has both groundball percentage and fastball velocity data, here is the percentage of times when both values are constant from season-to-season.

GB +/- FBv +/- Rate of Occurance
1.0 0.5 7.1%
2.0 1.0 21.8%
3.0 1.5 38.3%
4.0 2.0 52.4%
Minimum 20 innings pitched in each season.
Sample size = 4,445

Almost half of all pitchers see their fastball velocity change by more than 2.0 mph or see their groundball percentage change by four percent. These changes can really mess with a pitcher’s Solid Contact stats. First, here is a table of the average Solid Contact values for various groundball rates. The first table is the actual historical rates. For the second table, I found the best-fit line for each set of data and calculated the ideal value. Also, the r-squared values for each of the best-fit lines are included. The lines are curved with a peak. This is to see if a pitcher can be more productive as they move to the flyball/groundball extremes.

R-squared values

  • BABIP: 0.74
  • HR/contact: 0.98
  • wContact: 0.98
  • Hard%: 0.90
  • LD%: 0.86

27% 0.275 4.9% 0.229 27.8% 21.4% 36
28% 0.274 4.8% 0.216 28.2% 20.3% 33
29% 0.285 4.7% 0.223 30.1% 21.6% 49
30% 0.269 4.5% 0.215 27.9% 19.4% 66
31% 0.286 4.4% 0.214 29.1% 20.2% 92
32% 0.286 4.4% 0.216 28.8% 21.0% 106
33% 0.277 4.5% 0.215 28.1% 20.5% 145
34% 0.291 4.2% 0.209 29.2% 21.2% 149
35% 0.289 4.5% 0.216 28.7% 20.5% 201
36% 0.291 4.2% 0.209 28.9% 20.5% 211
37% 0.290 4.2% 0.205 28.7% 20.7% 228
38% 0.290 4.3% 0.209 28.9% 20.5% 260
39% 0.292 3.9% 0.197 28.2% 20.8% 273
40% 0.298 4.0% 0.201 28.4% 20.8% 291
41% 0.295 3.9% 0.196 28.4% 20.7% 292
42% 0.297 3.9% 0.197 28.2% 20.3% 285
43% 0.298 3.7% 0.188 28.3% 20.3% 296
44% 0.297 3.7% 0.190 28.1% 20.0% 298
45% 0.297 3.6% 0.186 28.4% 20.2% 318
46% 0.295 3.4% 0.175 27.7% 20.0% 287
47% 0.297 3.3% 0.177 27.7% 19.9% 269
48% 0.294 3.3% 0.170 27.6% 19.6% 236
49% 0.299 3.2% 0.171 27.6% 19.7% 241
50% 0.299 3.2% 0.171 27.0% 19.8% 196
51% 0.298 3.2% 0.166 26.9% 19.3% 199
52% 0.299 2.9% 0.163 27.1% 19.1% 145
53% 0.297 3.0% 0.160 26.9% 19.2% 140
54% 0.298 2.9% 0.157 26.2% 18.8% 119
55% 0.287 2.6% 0.143 25.8% 18.6% 99
56% 0.302 2.8% 0.150 26.6% 18.0% 114
57% 0.295 2.7% 0.144 25.7% 17.6% 77
58% 0.294 2.5% 0.140 26.3% 18.5% 68
59% 0.302 2.5% 0.138 25.8% 18.8% 43
60% 0.292 2.3% 0.131 25.5% 17.7% 45
61% 0.303 2.4% 0.139 25.0% 17.7% 42
62% 0.293 2.1% 0.125 24.2% 17.6% 38
Minimum 20 innings pitched in each season.
Sample size = 4,445

27% 0.275 4.9% 0.225 28.5% 20.7%
28% 0.277 4.8% 0.223 28.6% 20.7%
29% 0.278 4.7% 0.221 28.6% 20.7%
30% 0.280 4.7% 0.219 28.7% 20.8%
31% 0.282 4.6% 0.218 28.7% 20.8%
32% 0.283 4.5% 0.216 28.7% 20.8%
33% 0.285 4.4% 0.214 28.7% 20.7%
34% 0.286 4.4% 0.212 28.7% 20.7%
35% 0.288 4.3% 0.209 28.7% 20.7%
36% 0.289 4.2% 0.207 28.7% 20.7%
37% 0.290 4.1% 0.205 28.6% 20.6%
38% 0.291 4.1% 0.203 28.6% 20.6%
39% 0.292 4.0% 0.200 28.5% 20.5%
40% 0.293 3.9% 0.198 28.5% 20.5%
41% 0.294 3.8% 0.195 28.4% 20.4%
42% 0.295 3.8% 0.193 28.3% 20.3%
43% 0.296 3.7% 0.190 28.2% 20.3%
44% 0.296 3.6% 0.187 28.1% 20.2%
45% 0.297 3.5% 0.184 28.0% 20.1%
46% 0.297 3.5% 0.181 27.9% 20.0%
47% 0.298 3.4% 0.178 27.8% 19.9%
48% 0.298 3.3% 0.175 27.6% 19.8%
49% 0.298 3.2% 0.172 27.5% 19.6%
50% 0.298 3.1% 0.169 27.3% 19.5%
51% 0.298 3.1% 0.166 27.2% 19.4%
52% 0.298 3.0% 0.163 27.0% 19.2%
53% 0.298 2.9% 0.159 26.8% 19.1%
54% 0.298 2.8% 0.156 26.6% 18.9%
55% 0.298 2.8% 0.152 26.4% 18.8%
56% 0.298 2.7% 0.149 26.2% 18.6%
57% 0.297 2.6% 0.145 26.0% 18.4%
58% 0.297 2.5% 0.141 25.7% 18.2%
59% 0.296 2.5% 0.138 25.5% 18.0%
60% 0.296 2.4% 0.134 25.2% 17.8%
61% 0.295 2.3% 0.130 24.9% 17.6%
62% 0.294 2.2% 0.126 24.7% 17.4%

These tables help to start clarifying the confusion with batted-ball data. The first key is to notice that each of the Solid Contact measurements have a peak value, and usually not at the extreme ends, ISO excepted. The second table is a great way to compare a pitcher to other pitchers with similar groundball rates.

Moving on, here are the groundball percentags when each of the five values peak.

GB% at peak

  • HR/contact: 27%
  • ISO: 27%
  • LD%: 31%
  • Hard Hit %: 33%
  • BABIP: 52%

For most measurements, a high groundball rate is desired, but it must be understood the pitcher will likely allow a high batting average against and could get BABIP’ed to death. Also, some unique numbers will exist as the line-drive and hard-hit rates increase, while BABIP drops. Normally, people think BABIP should follow the other two stats, but it is heavily influenced by the number of ground balls (medium BABIP) and fly balls (low BABIP).

The values for pitch speed are easier to follow since the changes velocity causes are linear.

Fastball velocity changes +1.0 mph

  • LD%: -0.18%
  • Hard%: -1.1%
  • BABIP: -0.00365
  • ISO: -0.010
  • HR/BIP: -0.0006

So on average, when a pitcher sees a velocity drop, he should expect worse results. There are always exceptions to the rules. As a pitcher loses velocity, he may move from a straighter four-seam fastball to one with more movement, like a splitter or sinker. Or they can just continue to try to throwing the same way as they did in the past and fail (see Lincecum, Tim). In most instances when a pitcher loses velocity, he will continue with his old arsenal until he figures out it won’t work. Then he begins experimenting to find a new mix of pitches that will. Then, the pitcher may or may not stabilize at a new workable talent level. He usually will keep using this new pitch mix until it fails him again and he needs to adjust again. With all this adjusting going on, no wonder hard-hit rates never stabilize.

For some examples of how pitchers change, here are some 2016 pitchers who are seeing some Solid Contact changes because of velocity and groundball changes.

Edinson Volquez (FB velocity +0.2 mph, GB% from 46% to 53%)

The change is fastball velocity is negligible, but the change in groundball rate fits ideally with his changes. The increase in the groundball rate would lead to a higher BABIP but a lower number of home runs, line drives, and Hard Contact. All four of these things have happened, but to a larger extent than predicted — his BABIP has increased 23 points, and his line-drive rate is down six percentage points and his Hard% is down 8.5 percentage points. While HR/Batted Ball isn’t widely available, his home run rates are down, with his HR/FB going from 8.0 percent to 5.6 percent, and HR/9 from 0.72 to 0.48.

Dallas Keuchel (fastball velocity -1.6 mph, GB% from 62% to 56%)

Keuchel has both his velocity and groundball rate pull him around. All his batted-ball numbers should be getting worse with the drop in velocity. Additionally, since his groundball rate was so high in 2015, the drop caused all his batted-ball numbers to worsen. His BABIP, LD%, and Hard% have all gotten worse, but his home runs allowed actually has dropped. After looking over too many different pitchers, I find this mixed, but weighted, bag of results. The pitcher gets hit hard in several categories, but not all. Also, the pitcher is trying different pitches and locations to help get him out of the funk.

Gerrit Cole (fastball velocity -0.6 mph, GB% from 48% to 42%)

With Cole, the velocity drop should increase all his Solid Contact while the drop in groundball percentage should cause an increase in all values except BABIP, which should drop. His home runs are down. His BABIP was probably going to stay even, with the two variables pulling it in different directions. Instead, it is up along with both his Hard% and line-drive percentage. Again, not every item is going in the predicted direction, but most are.

As can be seen from this trio of pitchers, the predicted changes happen more often than not. A loss of velocity usually leads to worse results, while a change in groundball rate is not as important unless at the extreme ends of the spectrum.

I hate to say it, but I don’t think we will ever be able to definitively predict if a pitcher will continue to give up Solid Contact. The measures are in place to measure the solid contact, including the StatCast batted-ball velocity data. Carleton did a great service by proposing the idea of the ever-changing pitcher. Hopefully, I helped to clarify the issue by looking at how two ever-changing variables, groundball rate and velocity, can significantly change a player’s Solid Contact profile. Knowing the Solid Contact historic rate of change, a person can help understand the changes a pitcher is most likely to experience. A wormhole another researcher can climb down is to look at pitch-type usage and results to help determine the possible changes in specific pitchers.

Determining Solid Contact for pitchers has become one of the last major sabermetric questions to answer. Even as more and more data have become available, the answer has been evasive. The issue is that pitchers morph significantly and often. The answer to how to handle ever-changing pitchers may lie with a different time frame to look at pitchers, like half seasons to maybe even a month. Maybe each pitcher should have a change value associated with him to show how much they are different than the month/half season/year before. The key to finally answering the question of Solid Contact starts with how the pitcher is performing now and then moving slowly backward.

References & Resources

Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won three FSWA Awards including on for his MASH series. In his first two seasons in Tout Wars, he's won the H2H league and mixed auction league. Follow him on Twitter @jeffwzimmerman.
newest oldest most voted

I really like the framework you’re working with, given the hypothesis about the ever-changing pitcher (Yogi might ask whether we cut each season into 6 or 8 slices 😉 Using your combination of Retrosheet-level and PitchFX data gives a much deeper historical basis to test the issue. I’ve wondered whether there is any negative correlation between HR/FIP (or HR/FB) and what I’ll call ISOBIP, which is the difference between SLG on non-HR balls in play and BABIP. There ought to be a lot of luck-based volatility in the rate of balls just clearing the fence vs. just staying in the… Read more »


There are so many variables. For example, the era you measure is known as the “PED” era. Batters were “enhanced,” shall we say. Some pitchers were also received “help.” New ballparks with higher elevation, hence, lighter air, were paid for by We The People, not the “owners.” New bats using different wood, which splinters, maiming players and fans came into use. The way the game is played has also changed dramatically. Players swing for the fences and strikeouts have gone up exponentially. That fact must have an effect on “solid contact.” Players “back in the day” played for the TEAM.… Read more »


There are so many variables. For example, the era you measure is known as the “PED” era. Batters were “enhanced,” shall we say. Some pitchers were also received “help.” New ballparks with higher elevation, hence, lighter air, were paid for by We The People, not the “owners.” New bats using different wood, which splinters, maiming players and fans came into use. The way the game is played has also changed dramatically. Players swing for the fences and strikeouts have gone up exponentially. Players “back in the day” used to play for the TEAM. Now they play for the CONTRACT. MLB… Read more »


Buried in all this is actually one very interesting question: how much does the size of the strike zone affect pitchers’ ability to manage the quality of contact? I would think the huge strike zone of the 60’s would put the hitters more on the defensive, settling for weak contact vs. no contact at all. The mailbox sized strike zone of the 90’s let hitters wait for their perfect pitch, giving more incentive to drive the ball for power.


How about exit velocity?


Interesting analysis. I recently defined a measure for a pitcher’s intrinsic quality of contact which might be of interest at

and also found a correlation between a pitcher’s FB velocity and his ability to control contact according to this measure. The full list of intrinsic quality of contact values for batters and pitchers for 2014 appears at the end of

Stevie Y
Stevie Y

Mike Pelfrey another guy who’s seen a 1.7 MPH velocity decline from last year and with that a 14% increase in hard hit ball percentage (not counting Saturday’s start)

Stevie Y
Stevie Y

and David Price another good example as he’s had over a 2 MPH decline in FB velocity but he’s getting hit hard.

That’s been the big difference with him this year as most keep looking at an elite FIP and Siera and thinking he’s fine but the hard hit balls are the problem and have to wonder if that’s coming from decreased velocity.

jim fetterolf
jim fetterolf

Hitters do adjust, then pitchers adjust back. Then there’s variations among home plate umpires, some have zones that help pitchers, some have zones that hurt them.