Hunting for Great Splitters 1

When the Brewers recently claimed RHP Blake Parker and RHP Steve Geltz off waivers, and signed RHP Luke Barker to a minor league deal within the span of a week, one similarity that bridged the moves was the presence of a splitter in each arsenal. Granted, Barker’s split was listed last on his self-produced scouting video, and both Geltz and Parker used the pitch relatively infrequently (Parker behind his rising fastball and curve, Geltz behind his rising fastball and slider). Even the honorable Junior Guerra, in his very brief 2015 showing with the White Sox, already flashed his splitter twice as frequently both Geltz and Parker. However, given the Brewers’ success with the splitter in Guerra’s case, and the relative rarity of the pitch at the MLB level (only 53 starting pitchers and 62 relief pitchers have thrown more than 200 splitters since the advent of PITCHf/x), it’s worth picking over these admittedly marginal moves for either some similarity or some shred of hope: what were the Brewers thinking when they acquired these players? Have the Brewers learned some lessons from what they saw in Guerra? This is important to ask because it’s easy to write off Guerra as a one-off, singular, flukey front office claim; “there’s no way the Brewers can replicate Guerra” might be a reasonable chorus. Yet it is precisely David Stearns’s job to replicate the impossible and singular where the impossible and singular have previously provided roster value, since impossibilities and singularities will be two ingredients that can help small market Milwaukee contend for an extended period of time.


Since each of these new Brewers pitchers throws a splitter, it is worth investigate the elements that create an effective splitter. First, one can question which aspects are the most important for a splitter: for instance, is success with a splitter evident in swings-and-misses, groundballs allowed, flyballs prevented, or some combination? These questions can be answered via the BaseballProspectus PITCHf/x leaderboards. For the following study, I first looked at starting pitchers that threw 200 or more splitters all-time, and I specifically considered Total Average against (TAv), swing percentage, whiffs per swing, groundballs and flyballs per Balls-In-Play (BIP), and pop-ups per BIP and fouls per swing.

First things first, here are the most effective splitter TAv for starting pitchers. I took the Top 11 (because Hisashi Iwakuma was #11, and I’ve previously looked at an Iwakuma / Guerra comp), and then added some other interesting pitchers (more on that below):

TAv Against AB TAv Career DRA
Junior Guerra 113 .160 4.43
Mat Latos 203 .162 3.85
Jeff Samardzija 668 .170 3.94
Masahiro Tanaka 573 .175 2.91
Yu Darvish 92 .182 2.40
Hiroki Kuroda 896 .184 3.68
Matt Shoemaker 550 .191 3.68
Kelvim Escobar 145 .192 4.53
Jose Contreras 341 .192 4.67
Dan Haren 1131 .193 3.70
Hisashi Iwakuma 901 .194 3.06
Nathan Eovaldi 316 .202 4.70
Alex Cobb 921 .204 3.07
[Median: 53 SP] 389 .222 -
Daisuke Matsuzaka 147 .238 4.75
Vincente Padilla 293 .249 4.25
Tsuyoshi Wada 72 .256 4.45
Taijuan Walker 311 .257 3.95

In order to judge the effectiveness of these pitchers’ splitters, I attempted to isolate contact results that could suggest lack of effective timing by batters (such as foul balls or pop ups), potentially effective BIP results (more groundballs than flyballs, ostensibly keeping the ball in the ballpark at a higher rate), and causing batters to fail to make contact. I also including swing percentage, since one might argue that a splitter is ineffective if the batter recognizes it as a ball outside the zone, or an off-speed / breaking pitch. This first step is to attempt to build an anecdotal body of evidence about the splitter.

Swings and whiffs:
In terms of drawing swings and drawing whiffs, several of the most effective starting pitchers (Top 10, or Top 20 percent) in terms of splitter TAv also drew high swing percentages and/or high whiff percentages. Junior Guerra, for instance, is successful with the splitter almost solely due to his ability to draw whiffs.

Swing Rate Splitters (Velocity) Swing Rate
Alex Cobb 2839 (86.75) 62.49%
Mat Latos 543 (81.23) 61.33%
Masahiro Tanaka 1867 (87.42) 61.33%
Matt Shoemaker 1752 (84.75) 60.62%
Koji Uehara 363 (79.93) 60.06%
Hisashi Iwakuma 2713 (85.09) 59.80%
Carl Pavano 2485 (81.40) 59.56%
Roy Halladay 1508 (83.64) 57.03%
Kelvim Escobar 437 (86.20) 56.75%
Nathan Eovaldi 990 (88.93) 56.67%
Whiff / Swing Splitters (Velocity) Whiff / Swing
Kevin Gausman 1305 (84.88) 44.31%
Junior Guerra 403 (85.79) 40.82%
Jeff Samardzija 2164 (85.85) 40.66%
Tim Lincecum 5199 (83.97) 40.26%
Yu Darvish 315 (88.85) 40.12%
Ryan Dempster 2536 (82.36) 39.60%
Roy Halladay 1508 (83.64) 38.02%
Hiroki Kuroda 3368 (87.18) 36.19%
Mat Latos 543 (81.23) 36.04%
Masahiro Tanaka 1867 (87.42) 35.55%

This should not necessarily be a controversial suggestion about the splitter. First, the splitter needs to draw swings in order to fool the batter; it’s not necessarily a set-up pitch like a fastball, or even a cutter or tight breaking pitch (or a get-me-over curve, for that matter). Furthermore, if a pitcher is drawing a high percentage of whiffs on a pitch, that limits the chances that a batter has to induce damage. Below, one will see that Guerra is absent the groundball, flyball, popup, and foul ball leaderboards; in fact, his batted ball and contact numbers on the splitter are positively pedestrian. But, since Guerra draws a whiff nearly 41 percent of swings on his splitter, he maintains strong odds of success with the pitch.


Groundballs and flyballs:
Notably, the most effective splitter pitchers are absent the top of the groundball and flyball splitter lists. Below, I’ve provided the top 25 (nearly Top 50 percent) of groundball and flyball splitter pitchers, presenting the best TAv pitchers in bold. I’ve added extremely interesting pitchers in italics, specifically those who managed to pitch in the Top 10 of groundball and flyball results without landing in the Top 10 splitter TAv. This means that these pitchers yielded arguably preferable groundball / flyball performances without producing elite TAv against on their splitter.

Tsuyoshi Wada Tsuyoshi Wada
Masahiro Tanaka Hisashi Iwakuma
Hisashi Iwakuma Taijuan Walker
Nathan Eovaldi Masahiro Tanaka
Carlos Zambrano Yu Darvish
Jake Westbrook Nathan Eovaldi
Alex Cobb Franklin Morales
Brad Penny Kevin Gausman
Roy Halladay Alex Cobb
Taijuan Walker Jake Westbrook
John Smoltz Mat Latos
Tim Hudson Jorge de la Rosa
Hiroki Kuroda Mike Pelfrey
Carl Pavano Carlos Zambrano
Alex White Homer Bailey
Jorge de la Rosa Alfredo Simon
Braden Looper Miguel Gonzalez
Brandon Morrow John Smoltz
Miguel Gonzalez Brad Penny
Yu Darvish Braden Looper
Kevin Gausman Hiroki Kuroda
Dan Haren Jake Odorizzi
Manny Parra Dan Haren
Jonathan Sanchez Kyle Kendrick
Mike Pelfrey Roy Halladay

The implication here is that the splitter is not a “weak contact” pitch. A pitcher may not necessarily look to the splitter to induce a groundball, or prevent a flyball. This hypothesis is drawn from the fact that the very best splitter TAv pitchers do not necessarily exhibit these desirable batted ball traits, meaning that their splitters are not necessarily effective because of their groundball / flyball ratios. Tsuyoshi Wada, Alex Cobb, Nathan Eovaldi, and Taijuan Walker are the outliers worth studying here.

Pop Ups and Fouls:
Like groundballs and flyballs, so too with pop ups and foul balls: once again, it does not appear that a great splitter is great because of its ability to induce weak contact.

PU / BIP Foul / Swing
Jose Contreras Alex White
Franklin Morales Vincente Padilla
Vincente Padilla Daisuke Matsuzaka
Daisuke Matsuzaka Charlie Morton
Koji Uehara Jake Odorizzi
Freddy Garcia Alex Cobb
Jason Marquis Dan Haren
Matt Shoemaker Braden Looper
Homer Bailey Matt Shoemaker
Ryan Dempster Tsuyoshi Wada
Wei-Yin Chen Kenshin Kawakami
Kelvim Escobar Jake Westbrook
Jeff Suppan Mike Pelfrey
Kyle Kendrick Ubaldo Jimenez
Manny Parra Hisashi Iwakuma
Doug Fister Alfredo Simon
Clay Buhholz Brandon Morrow
Jake Odorizzi Carlos Zambrano
Dan Haren Homer Bailey
Tim Lincecum Jorge de la Rosa
Alfredo Simon Brad Penny
Brandon Morrow Mat Latos
Jeff Samardzija Jonathan Sanchez
Charlie Morton Taijuan Walker
Kevin Gausman Masahiro Tanaka

I toyed with combining whiffs per swing with foul balls per swing, in order to create a quick-and-dirty hybrid stat that accumulates the trickiest splitters: the ones that cause batters to swing and miss, or make mistimed contact a solid percentage of the time. But Guerra is once again instructive here, as he would perform very well in this statistic because of his high strike out rate. In this case, a “Power/Speed Number” method employing the harmonic mean could find the most balanced and effective timing-limiting splitters (using a formula such as [(2*FoulSwing*WhiffSwing) / (FoulSwing + WhiffSwing)]).

Judging the new Brewers righty splitter tossers, then, one might hypothesize that focusing on the ability to induce swings and whiffs is indeed much more important than focusing on splitter-induced grounders, flyballs, fouls, and pop-ups. This may seem like an obvious conclusion, but it is worth investigating the PITCHf/x results in order to verify the assumption. In this sense one might also be able to form accurate expectations for this gang of Brewers acquisitions.


Stuff Quality
The benefit of having PITCHf/x data available is that one can compare how splitters move compared to the other major pitches in a pitcher’s arsenal. This is another qualitative investigation of a pitcher’s splitter production, as one could conceivably compare horizontal and vertical movement variations, as well as spin rate and velocity differentials, in order to determine why the best splitter TAvs are indeed the best.

With this stated, here is the list of the notable right-handed starting pitchers analyzed above. I drew information from Brooks Baseball, isolating a pitcher’s primary fastball or sinker and main breaking pitch against the splitter. This essentially builds a basic game whereby one can understand how the splitter is used:

  • Is it a “true change of speed” (slower than the breaking pitch AND fastball) or is it a “stepping stone” between the fastball and breaking ball?
  • Does it primarily move armside or gloveside compared to both the fastball and breaking pitch?
  • Does it primarily “rise” or “drop” compared to the fastball and breaking pitch?
  • Perhaps most importantly, what are the other pitches that a successful splitter ball pitcher throws? One should not rule out that a pitcher’s other offerings make a splitter great, or amplify the effects of a splitter, and one should not assume that a pitcher’s splitter produces great results only because it is a great splitter on its own. (This may seem obvious, but it is worth arguing that a pitcher’s arsenal is a comprehensive whole, and not simply a sum of parts).
Splitter Quality H / V from Main Fastball H / V from Breaking Ball Velocity (FB / Split / Break)
J. Guerra +1.93” / -5.67” -2.26” / +5.27” (!!!) 94.30 / 86.02 / 83.07 (Slider)
M. Latos Even (!!!) / Drop 11.18” -2.77” / +1.69” 93.62 / 81.33 / 85.61 (Slider)
J. Samardzija +0.63” / -4.03” -7.14” / +3.15” 95.33 / 85.92 / 85.90 (Slider)
M. Tanaka +2.27” / -3.44” +7.94” (!!!) / -0.06” 90.92 (Sinker) / 87.41 / 84.31 (Slider)
Y. Darvish -2.33” / -6.74” (!!!) -14.04” (!!!) / +5.16” 93.86 / 88.69 / 82.33 (Slider)
H. Kuroda +3.69” / -3.87” -6.01” / -1.30” (!!!) 92.60 (Sinker) / 87.21 / 84.47 (Slider)
M. Shoemaker -2.51” / -6.24” (!!!) -6.47” / +1.11” 91.47 / 84.65 / 82.29 (Slider)
K. Escobar +1.61” / +0.20” (!!!) -5.03” / +6.62” (!!!) 95.02 / 86.61 / 87.49 (Slider)
J. Contreras +4.32” / -4.42” -4.59” / -3.07” (!!!) 92.67 (Sinker) / 78.45 / 86.25 (Slider)
D. Haren 1 -5.78” / -1.09” -8.31” / +7.33” (!!!) 85.78 (Cutter) / 84.83 / 79.61 (Curve)
D. Haren 2 +1.45” / -4.82” -5.78” / -1.09” 90.28 (Sinker) / 84.83 / 85.78 (“Cutter”)
H. Iwakuma +1.55” / -3.93” -11.31” / -0.74” 89.14 (Sinker) / 85.15 / 81.40 (Slider)
K. Gausman -1.15” / -6.67” (!!!) -9.41” / +5.49” 96.14 / 85.05 / 80.67 (Curve)
T. Lincecum -2.36” / -6.65 (!!!) -4.36” / +4.20” 92.35 / 83.99 / 83.46 (Slider)
R. Dempster +1.26” / -3.98” -7.53” / +2.96” 91.37 (Sinker) / 82.48 / 85.38 (Slider)
R. Halladay -4.65” / -5.29” -12.71” (!!!) / +1.35” 91.23 (Cutter) / 83.71 / 78.22 (Curve)
Horizontal = H + = breaks gloveside - = breaks armside
Vertical = V + = “rises” - = “drops”

What does this mean? A few observations:

  • The splitter is almost universally paired with the slider. This is fascinating in and of itself — baseball has a tradition of “we stick together” arsenals, meaning that pitchers work with what is tried and true. For as rare as splitter pitchers are in the MLB, it is surprising that so many of the best splitter pitchers have comparable breaking pitch offerings.
  • Only six of the best splitter pitchers use the offering to break further armside against their fastball, meaning that the pitch would break further inside against a right-handed batter. Six of the best splitter pitchers (including Guerra) feature a split that breaks more than 1.5″ gloveside against the fastball, or “away” from a right-handed batter.
  • Everyone except for Masahiro Tanaka throws a splitter that breaks further armside (“in” on RHB) compared to their breaking ball. The vast majority of great splitters also “rise” compared to a breaking ball; obviously since pitches don’t actually “rise,” one might more effectively say that this is an optical / perceptual tool that allows a pitcher to play the breaking ball against the splitter, since they “drop” in different manners (the breaking ball much more, the splitter much less).
  • Some pitchers have “equidistant” horizontal and vertical ranges between their fastballs and breaking balls, when compared with their splitter. For example, take Guerra, who uses a split that “drops” 5.67″ from his fastball, but “rises” 5.27″ from his slider. Similarly, his split breaks to the gloveside by 1.93″ against the fastball, but moves armside by 2.26″ against the slider. So, one can see that Guerra aids his 94 / 86 / 83 fastball / split / slider stepping stone with vertical and horizontal planes that differentiate the pitches in a harmonic manner. Jose Contreras, Yu Darvish, and Kevin Gausman exhibit some similar harmonies, although not necessarily to the extent of Guerra.
  • What surprised me the most is that the pitchers are evenly divided in how they use the splitter in their velocity range; approximately half of the arsenals use the splitter as a bridge between the breaking ball and fastball, leaving the others to use the splitter as a true change of speed. Some of these pitches move into semantics, as pitchers like Jose Contreras probably threw a “forkball,” Tim Lincecum probably uses a “vulcan change up” variation, and Mat Latos throws an unclassifiable oddity.
  • Interestingly enough, having a top TAv against on a splitter does not always require a blazing hot fastball. In fact, some of the most effective splitter pitchers paired that offering with another moving fastball option, such as Roy Halladay’s gang of darting sinker-cutters, Dan Haren’s cutter, and a handful of other effective sinker pitchers (including “old” splitter pitchers like Ryan Dempster or Hisashi Iwakuma).

There is not a clear arsenal for success for projecting an unknown splitter pitcher into stardom, although one can suggest that using the pitch to elicit whiffs is much more important than any other outcome. The next question will be, “how do the new Brewers fit into these hypotheses?”

Related Articles

Leave a comment