Groundball vs. Flyball Pitcher

There's been some interesting debate lately, on various posts and gameday threads, regarding the relative value of a groundball vs. flyball pitcher, especially at Chase Field with our current defensive alignment. This debate has become all the more interesting, as we watch the success of flyball pitchers like Ian Kennedy and Daniel Hudson, whereas groundball pitchers such as Zach Duke and Jason Marquis have struggled.

I thought I would try to take a stab at it, using numerical projections and test cases.

Background:

Tom Tango and Bill James recently made some interesting commentary on the subject. Shout out to paqs who sent me this article. Tango also presents the following run values for their corresponding baseball events, which is crucial to understanding the impact of groundballs vs. flyballs:

+1.4 FB (only HR) 
+0.3 walks, line drives 
+0.05 FB (including HR) 
-0.1 GB, FB (sans HR) 
-0.3 strikeouts, popups

It's important to note, that I believe Tango's designation of FB does not include pop ups. Otherwise, the implication would be that Tango's last line would say something like "FB (pop ups)" as opposed to just "pop ups", or he would say that the FB includes pop ups, or something similar.

Definitions and Assumptions:

In the following tables:

BIA refers to Balls in Air (includes line drives, normal flyballs, home runs, and popups)

BIP refers to Balls in Play (includes groundballs and balls in air). This was set at 800, which is a very high upper limit. Roy Halladay, who sets the gold standard for how deep into games he pitches, faces under 1000 batters per season. Once you subtract strikeouts and walks, pretty much every pitcher in the game should have fewer than 800 BIP during the course of a season.

I set Batted Ball Distribution at 55% GB for a groundball pitcher, as you will find very few groundball pitchers that throw for a higher percentage than that in the majors. Similarly, flyball pitchers are set at 40% GB.

HR/BIA is set at 7.5%. According to Statcorner, MLB average HR/BIA is 6.5%. I estimated that Chase gives up HR 15% more than league average. Which is a very very high estimate, considering that while Chase's HR park factor is 21% this year, it was only 6% and 4% in 2010 and 2009. However, as you will see the case in many of my other assumptions, I will do my best to shade assumptions in favor of groundballers to prove my point.

Popups per BIP is 7.3% according to Statcorner. I simplified it to 7.0% to make the numbers easier, and again, to give further benefit to groundballers.

Checking out the Numbers:

So what do all these numbers mean? Let's go back to Tango's run values for a second. "-0.1 GB, FB (sans HR) " basically shows us that normal flyballs and normal groundballs (ones that turn into outs, singles, doubles, or triples) tend to even out over the long run in terms of run value created. There are two major exceptions: Home Runs and Popups.

A flyball pitcher will assuredly give up more home runs than a groundball pitcher. The point is, can the flyballer counter with enough popups and strikeouts to make the groundball pitcher's extra groundballs meaningless. Assuming that two pitchers, a groundballer and a flyballer, have equal walk percentages and line drive percentages, and assuming these two pitchers have the same popup rates, then the flyball pitcher needs to strike out 33-34 more batters faced during the course of a season, to break even.

In actuality, the exact number of strikeouts doesn't really matter. The key is realizing that given these other factors remaining constant (walk percentage, line drive percentage, popup percentage), a flyball pitcher at Chase Field just has to strike out somewhere between 3-4% more batters faced than a corresponding groundball pitcher in order to compensate.

Brandon Webb:

But wait. We all remember Brandon Webb being a dominant groundball pitcher at Chase. What's going on? Groundballers are good.....right?

People tend to forget, because of just how insane of a groundballer Webb was, but the guy was a damn good strikeout pitcher as well. Healthy Webb struck out 19-20% of the batters he faced. Combined with the fact that Webb generated 63-64% groundballs (as opposed to the 55% I used in my test case), you essentially need a 5% jump in strikeout rate to match that. Trust me when I say, there aren't too many guys in baseball who strike out 24-25% of the batters they face. Webb is essentially the groundball yin to Verlander's flyball yang.

Problems with the Assumptions:

There's a lot of reason to suggest that the above is an overestimation of groundballer abilities. A lot of assumptions were shaded to give groundballers the benefit of the doubt. A lower BIP limit would mean fewer HR which would mean fewer strikeouts necessary to compensate. The Chase HR/BIA is very generous for many reasons, in favor of groundballers. First, that rate is adjusted from the MLB average, which is likely to be higher than the NL average (because the NL has the pitcher hit). Second, it would assume we play all of our games at Chase, which we don't. In fact, we play 27 games each year in the extremely HR-suppressing ballparks of our NL West rivals. Finally, groundballers tend to have higher HR/BIA rates than flyballers.

All of this, probably puts the percent of extra batters faced that a flyball pitcher needs to strike out, closer to 3%, than 4%.

Defense and Opportunity Cost:

This to me, is one of the major reasons why we need to tailor our pitching to more flyball-prone pitchers. I can't profess to be an expert on how UZR is actually calculated, so the exact numbers in the following argument may be misleading. However, I think the general point remains.

Upton's UZR/Play = 9.3/168 = 0.055

Young's UZR/Play = 7.6/216 = 0.035

Parra's UZR/Play = 16.0/122 = 0.131

Our outfield UZR/Play comes out to 0.221 runs saved per play.

Drew = 4.4/157 = 0.028

Johnson = 2.7/175 = 0.015

Roberts (at 3B) = 1.8/117 = 0.015

And we don't have numbers yet for Goldschmidt, but let's assume a best-case scenario of 5 runs saved over 100 plays.

Our infield UZR/Play comes out to an estimated 0.108 runs saved per play.

There are of course important small sample size concerns in this calculation, but the opportunity cost is obvious. If you believe that the opportunity cost, based on our defense, is 0.1 runs saved per play, then using the standard 800 BIP assumption I used above, a groundball pitcher who throws 55% groundballs would be costing his team 12 runs more than the flyball pitcher who throws 40% groundballs. That's basically equivalent to one win over the course of a season, and it's only taking into account one starting pitcher. Also note, that it almost singlehandedly completely wipes out the advantage of the groundballer vs. the flyballer at preventing home runs (see above: HR Differential run value is 12.6).

Conclusions:

And so, after this long essay comparing groundball pitchers and flyball pitchers, I came to the following conclusions.

1. Groundball pitchers can be good, but they are essentially comparable to Flyball pitchers who strike out 3% more of their batters faced.

2. Brandon Webb was really really good when he was healthy.

3. The fact that our outfield defense is quite a bit better than our infield defense, creates not insignificant opportunity costs when a groundball pitcher is on the mound.

Bonus: I didn't mention this above, but it's been shown that ERA as a statistic heavily favors groundball pitchers over flyball pitchers. This is because it's much easier to make an error on a groundball versus a flyball. So for instance, when comparing an extreme groundballer versus an extreme flyballer, if both pitchers had 4.00 runs allowed per nine innings, then the groundballer is likely to have a 3.60 ERA, whereas the flyballer is likely to have a 3.76 ERA. Despite both pitchers allowing the same number of runs, because it's harder to play defense for a groundballer, the groundballer's ERA looks nicer.