Advancing by ground

It’s good to be a ground ball pitcher. If you have a good infield behind you. Or a nicely manicured diamond with good lights. Whatever it may be, the further a pitcher advances in the professional baseball, the better he fares on ground balls. The emerging ground ball pitcher has been a recent fascination of mine, and I’m not alone in that regard.

The more I think about ground balls, the more I think about line drives and fly balls. Let me show you why, and what it means for tracking pitchers across levels.

First by ground

If you look at the lowest levels of the minor leagues, you’ll find more ground balls result in a batter reaching base (by hit or by error) than at the higher levels of pro ball, in particular the major leagues. One tradeoff that comes with that progression (presumably related to the factors noted earlier) is an increased frequency of extra base hits off ground balls. It’s not a large number of hits, so the tradeoff seems worthwhile.

Batting average and Slugging rates on ground balls, with errors treated as singles (2007-9)

League GB AVG SLG ISO
R-PIO 24713 0.305 0.322 0.018
R-GCL 34508 0.302 0.316 0.014
R-APP 25576 0.325 0.345 0.020
As-NYP 39485 0.307 0.324 0.017
As-NWL 24020 0.314 0.330 0.016
A-SAL 84488 0.295 0.314 0.020
A-MDW 72979 0.299 0.316 0.017
Av-FSL 60930 0.279 0.298 0.019
Av-CAR 42583 0.287 0.309 0.022
Av-CAL 54357 0.305 0.326 0.021
AA-TEX 42762 0.284 0.307 0.022
AA-SOU 50353 0.282 0.304 0.022
AA-EST 62303 0.289 0.307 0.019
AAA-PCL 84373 0.272 0.295 0.023
AAA-INT 73566 0.280 0.303 0.023
MLB-NL 115621 0.269 0.291 0.022
MLB-AL 101502 0.268 0.289 0.020

image

At no point does the extra extra-base hit that comes with facing stronger hitters at higher levels impact the benefits of the better field(er)s at those levels.

Before we decide ground ball pitchers will benefit from promotion, we need to consider the ground ball rates themselves. How do they vary by level/league?

Ground balls per ball in play (2007-9)

League GB%
R-PIO 50.4%
R-GCL 49.5%
R-APP 51.9%
As-NYP 49.0%
As-NWL 50.6%
A-SAL 49.0%
A-MDW 48.1%
Av-FSL 47.2%
Av-CAR 48.2%
Av-CAL 46.9%
AA-TEX 46.2%
AA-SOU 45.4%
AA-EST 45.9%
AAA-PCL 44.4%
AAA-INT 45.1%
MLB-NL 44.7%
MLB-AL 44.2%

image

Tempers the expectations, doesn’t it? As you get more out of ground balls, you get fewer of them. To gauge how that works out (or washes out), I applied some very generic linear weights to outs, errors, singles, doubles and triples at each level, using the same weights.

Ground ball linear weights

image

I know what you’re saying: Those grounders are turning into something else, like a fly ball. Thing is, fly balls are worth fewer runs each level you go up (using the generic linear weights), and, after rookie ball, the rate of change is very much the same (-0.0048 for fly balls and -0.0056 for ground balls) from level-to-level. A ground ball saves from .12 runs to .15 runs, compared to a fly ball, with the exception of Rookie ball, where the difference hops to .17. It’s really the fault of the high-elevation Pioneer League and its slightly lower cousin, the Appalachian League. The Gulf Coast League doesn’t see the long flies leave the yard as often as they do in the mountainous leagues. This makes no mention of line drives (worth more than > .33 runs) and pop-ups (almost a sure out, -.28 or so).

This all raises two questions—what are those batted ball values by league and what do those lost grounders become as pitchers advance? While I’ve already shown this above for ground balls one way, now you’ll see all batted ball types grouped by level and by league.

Stepping out of the vacuum—with one foot

Batted ball linear weights (2007-9)

League GB LD FB PU
R-PIO -0.05 0.39 0.16 -0.28
R-GCL -0.06 0.33 0.08 -0.28
R-APP -0.04 0.34 0.12 -0.28
As-NYP -0.05 0.32 0.08 -0.28
As-NWL -0.05 0.35 0.10 -0.28
A-SAL -0.06 0.37 0.07 -0.28
A-MDW -0.06 0.33 0.08 -0.28
Av-FSL -0.07 0.37 0.04 -0.28
Av-CAR -0.06 0.35 0.09 -0.28
Av-CAL -0.05 0.36 0.13 -0.28
AA-TEX -0.07 0.33 0.05 -0.27
AA-SOU -0.07 0.36 0.04 -0.28
AA-EST -0.06 0.35 0.08 -0.28
AAA-PCL -0.08 0.37 0.10 -0.27
AAA-INT -0.07 0.34 0.05 -0.28
MLB-NL -0.08 0.35 0.06 -0.27
MLB-AL -0.08 0.36 0.06 -0.27

Let’s try that again, using unweighted averages by level, just for aback-of-the-envelope look at the progression.

Level GB LD FB PU
R -0.05 0.36 0.12 -0.28
As -0.05 0.34 0.09 -0.28
A -0.06 0.35 0.08 -0.28
Av -0.06 0.36 0.09 -0.28
AA -0.07 0.35 0.06 -0.28
AAA -0.07 0.35 0.07 -0.27
MLB -0.08 0.36 0.06 -0.27

Outfield play improves almost as much as infield play—or so it appears. From the pitcher’s view, moving from rookie ball to the major leagues, ground ball values improve ~60 percent, fly balls ~50 percent, line drives ~0.2 percent and pop-ups ~4 percent. So, a ground ball pitcher should benefit from improved defensive play more than a fly ball pitcher. Since these linear weights are generic, and the by-level averages unweighted, you should be wearing your 3D error bar goggles.

Now for the second question—what happens to the batted ball distributions? This could change how we think about the answer to the first question. Again by league, and then by level. You’ve seen the ground ball numbers already.

Batted ball types (2007-9)

League GB LD FB PU
R-PIO 50.4% 15.6% 27.9% 6.1%
R-GCL 49.5% 13.0% 29.9% 7.7%
R-APP 51.9% 13.8% 27.8% 6.6%
As-NYP 49.0% 14.5% 28.9% 7.6%
As-NWL 50.6% 14.9% 27.4% 7.2%
A-SAL 49.0% 16.0% 28.1% 6.9%
A-MDW 48.1% 14.1% 30.1% 7.7%
Av-FSL 47.2% 16.1% 29.6% 7.2%
Av-CAR 48.2% 14.0% 30.6% 7.3%
Av-CAL 46.9% 13.4% 32.3% 7.4%
AA-TEX 46.2% 19.1% 27.3% 7.4%
AA-SOU 45.4% 17.9% 29.0% 7.7%
AA-EST 45.9% 14.5% 32.0% 7.7%
AAA-PCL 44.4% 18.5% 29.8% 7.3%
AAA-INT 45.1% 18.9% 28.5% 7.5%
MLB-NL 44.7% 18.7% 29.0% 7.5%
MLB-AL 44.2% 18.7% 29.5% 7.7%

By level, while ironing out important variance, you can see the shift a little more easily.

Level GB LD FB PU
R 50.6% 14.1% 28.5% 6.8%
As 49.8% 14.7% 28.2% 7.4%
A 48.5% 15.1% 29.1% 7.3%
Av 47.4% 14.5% 30.8% 7.3%
AA 45.8% 17.2% 29.4% 7.6%
AAA 44.8% 18.7% 29.1% 7.4%
MLB 44.4% 18.7% 29.3% 7.6%

Ignoring rookie ball, a five-point drop in ground ball rate appears to get picked up mostly by line drives over fly balls by nearly four to one. That’s important, as a ground ball turned into a line drive is about three times as costly as one turned into a fly ball. I know stuff is being redistributed in various directions across batted ball types, but it ends up as I’ve described.

So, a pitcher who sheds five ground balls per 100 balls in play over the course of his journey is picking up four line drives and a fly ball. Let’s take a reality-based hypothetical example.

These are the GB/FB/LD/PU rates for the stops along the route taken by prospects in the Chicago Cubs system:

Northwest       51 15 27  7
Midwest         48 14 30  8 
Florida State   47 16 30  7
Southern        45 18 29  8
Pacific         44 19 30  7
National        45 19 29  7

A prospect in the Chicago organization can, in theory, be traced from level-to-level and measured against these benchmarks. One exception: the last stop. It’s not really fair to grade major league rookies on the scale of the average major league pitcher. At that level, a given prospect is more likely to be in over his head than at any other stop along the way. In fairness, age adjusted levels should be used as benchmarks for each level, but I’m not projecting yet, just tracking.

To track pitchers, I want to first classify them. That’s a tricky problem itself, since developing pitchers are, by definition, changing. And that’s the next topic in this thread, how and when to classify or re-classify a pitcher as a “ground ball” or “fly ball” guy. We’ll return to the question of rookie performance benchmarks in the major leagues, as well.

References & Resources
Batted ball data from MLBAM. PITCHf/x data from Sportvision and MLBAM. Pitch classifications by the author.


3 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
kds
14 years ago

Harry,

Good stuff.  You seem to be assuming that it is the same population of pitchers at every level from Rookie ball to the Majors.  Might it not be that the pitchers who got great ground ball %‘s at low levels don’t make higher levels because they are bad at something else; eg. too few K too many BB.
As an alternative theory, maybe it is the population of hitters that has changed, the ones that hit more ground balls being left behind by the ones that hit more line drives.

I think it would be interesting to take a large group of ML pitchers and look at their batted ball %‘s at all minor levels to see if these ultimately successful pitchers were systemically
different from the averages in the minors.

Harry Pavlidis
14 years ago

That’s right, and that’s exactly what will emerge (or at least being to) in the next part.

One of the approaches I’m toying with, certainly not the only, is to start with pitchers who made MLB debuts in 2009 and work back. See if and how they differentiated from peers, and how those peer groups changed, as they progressed. And to repeat the entire cycle with hitters.

And, you’ve hinted at something which is, from the limited tests I’ve run, a key component for GB pitcher success – K/BB. What I’ve found so far supports your “might it” very well.

PhD Brian
14 years ago

I have heard that ground ball pitchers will strike out fewer batters on average than a flyball pitcher because high thrown balls get more swings and misses than low thrown balls.  Assuming they are equally as likely to be thrown for strikes or balls (I do not know), then flyball pitchers should have better K/BB ratios on average than groundball pitchers.  Groundball pitchers are more at the mercy of their team defense which should be weaker in the minors.  Thus flyball pitchers may get promoted more often.

However, since flyballs are far more likely to go for homers off major leaguers bat than a minor leagures bat,flyball pitchers should fall off in ability faster than groundball pitchers once they reach the majors.  So perhaps these promotions are inefficient.

Interesting to think about!