Mound Chess: A Look at Greg Maddux’s Pitch Sequencing

A game of chess is not won with a single move but rather through the ability to see strikes and counters well in advance.  Garry Kasparov, not-so-arguably the best chess player in the world, has the ability to see dozens of moves before they take place.  This is not clairvoyance.  Instead, it is a skill learned through a multitude of experience in executing, and defending against, certain strategies.  If an equivalent to Kasparov exists in the world of baseball it would have to be Greg Maddux.  Even though the title of this article gave the identity of its subject away, the idea of succeeding through intellect as well as picking apart oppositional tendencies is most commonly associated with the man given the moniker of “The Professor.”
Tim Keown recently wrote a tremendous article on Maddux in which he discussed his chesslike approach to pitching.  Teammates remarked to Keown that Maddux can watch a seemingly run of the mill swing and deduce, with great accuracy, what will occur over the course of the next few pitches.  This specifically comes in handy with regards to foul balls headed towards their dugout.
Unfortunately, Pitch F/X data was not available during the pinnacle of Maddux’s success; hell, computers were still becoming mainstream back then and I was eight years old.  Still, there are lessons that can still be learned about his approach. 
Maddux Data
Using the available 2007 and 2008 data I decided to find a batter or two that Maddux has faced multiple times and analyze his approach in pitching to them.  Did he utilize the same sequencing?  Did he make adjustments between games?  Did he make adjustments between at-bats in the same game?  Was he able to fully exploit said batter’s weaknesses?
Fortunately, Maddux pitches for the Padres, and Petco Park was one of the original SportVision stadiums. 
I wanted to look at a hitter that Maddux faced multiple times and found him in Bengie Molina: the Padres and Giants are in the same division and Maddux has already faced them twice in 2008.  All of the circled pitches seen in the location charts are the first pitches thrown in each plate appearance.
With regards to pitch classification, Maddux’s pitches each have so much movement that the 2008 algorithm has suggested he throws about eight different pitches. Generally speaking, he throws a two-seam fastball, a changeup, and a cutter that moves a lot like a slider; due to this movement it is frequently categorized as a slider. Though it can be argued it is, in fact, a cutter, I am going to stick with the algorithm in this case because it classifies what the batter sees. Regardless of whether we call it a cutter or a slider, the batter sees a pitch moving similarly to a slider. For the purposes of this article I will call it a “cutslide.”
Understanding Molina
Before getting into the matchups I feel it is important to know a bit about Maddux’s opponent.  Molina does not walk or strikeout much as he generally puts the ball in play more often than not.  Throughout his career he sees about half of his pitches outside the strike zone.  Though he only swings at about 31-33% of these pitches he makes contact upwards of 79% of the time.   He swings at about 70% of the pitches he sees in the strike zone but makes contact around 90% of the time.  Clearly, Molina does not swing and miss much.
Of this contact, the majority comes from grounders and flyballs; in his career he has hit 19% line drives, 37% grounders, and 43% flyballs.
Maddux vs. Molina: 4/11/2007
Early last April, Maddux faced Molina twice, throwing just four pitches in the two plate appearances.  The first, a PA lasting three pitches, resulted in a double.  In the second PA, Molina grounded out on the first pitch.  All four of these pitches were two-seam fastballs and their locations can be seen below:
madduxmolina4607.JPG
In the first plate appearance, Maddux started Molina off with a fastball, letter high, resulting in a called strike. He then challenged him on the outside part of the plate; Molina fouled it off. Never one to throw a waste pitch (something I’ll explore next week) Maddux threw an 0-2 fastball towards the inside corner. It came in a bit too high and closer to the plate than he would have liked and Molina lined a double.
Molina adjusted his approach as he is generally not one to swing at the first pitch. The common creed with Maddux is that, since he is deadly accurate and despises wasting pitches, the first pitch you see may be the best one. He abides by the “best pitch is strike one” motto and so by taking the first pitch a batter is usually seceding the ability to be ahead in the count.
Maddux, however, did not make an adjustment between at-bats: His first pitch to Molina the next time they faced off was put in virtually the exact same location as the first pitch from before. Perhaps Molina knew this would happen as he swung at the high two-seamer, grounding out to shortstop in the process. Here is what the movement of these pitches looked like:
movement1.JPG
Unlike the other matchups seen in this analysis, the charts from this game are separated by plate appearance rather than pitches; he threw the same pitch in both so it is more interesting to see how the PAs differed. Though the fastballs in the first plate appearance differed in location they all seemingly had very similar movement. The lone fastball the second time around slightly differed. Whether or not Maddux held the ball with a different grip or if this movement discrepancy occurred by chance, it resulted in a groundout.
 

Maddux vs. Molina: 8/3/2007
The two once again squared off on June 27, 2007, but for whatever reason the Pitch F/X system did not record all of the pertinent data required to analyze the sequencing. A little over a month later the teams met again. Molina continued his solid production against Maddux, going 2-3 at the dish with two singles and a groundout.Unlike in April, when Maddux “pounded” him with fastballs, Molina saw a nice mix of pitches throughout these three plate appearances. Below is a location chart showing the pitches thrown and locations:
madduxmolina8307.JPG
Maddux started Molina off with a two-seamer that just barely missed the outside corner for a ball. He then decided to throw to the same location albeit with a changeup. Again, Molina held, but this time the pitch was called a strike. From watching the game it appeared this was some form of a makeup call. Though the strike zone in the chart would contend his first pitch missed the corner, this was clearly a pitch usually given to Greg Maddux. With a 1-1 count Maddux got a bit lucky; he threw a two-seamer essentially right down the middle of the plate but induced a groundout.
Sticking with what worked, Maddux began the next plate appearance with a two-seamer on the outside corner, in an almost identical position to the called strike changeup in the second inning. Molina held for a called strike one. Again, Maddux followed an outside fastball with an outside offspeed pitch, this time reverting to his cutslide. Molina reached out and laced a single to left. As you can see in the above location chart Maddux hung this pitch.
The next time up, Maddux adjusted his approach but had trouble locating. Instead of hitting the outside corner with a fastball he started Molina off with an up and in cutslide that caught more of the plate than I imagine he desired. Due to his accuracy and first-strike mindset, starting Molina off with a non-fastball suggests that Maddux anticipated Bengie would break habit and jump at the first pitch, just as he had in April. Molina took the called strike. Showing him a breaking ball in Maddux then went back to the outside corner with a changeup. The pitch sailed too far outside and evened the count at 1-1. Maddux attempted to throw a two-seamer on the inside corner–his patented pitch that starts inside to lefties but tails right back over the plate–but badly missed his spot, throwing it right into Molina’s wheelhouse. Again, the slowest player in baseball history laced a single to left.
Here is how his pitches moved throughout these three plate appearances:
movement2.JPG
Maddux vs. Molina: 4/7/2008
Maddux has faced off against the Giants, and specifically Matt Cain, twice in the early going of 2008, allowing us to see if his approach to Molina has changed over the offseason. Facing Molina three times he threw a total of 13 pitches; Molina singled and grounded out twice. Here are the locations:
madduxmolina4708.JPG
The first time up Molina saw six pitches (three two-seamers and three cutslides). Maddux started him off with a two-seamer down and away that missed for a ball. He followed it up with a cutslide way off the inside corner. Behind 2-0 he threw a picture perfect, down and away two-seamer. Unfortunately, the umpire did not agree with my assessment and called it ball three. Maddux suddenly found himself in the quite-unusual-for-Greg-Maddux situation of a 3-0 count.
Last year, Pitch F/X recorded just 14 instances of a 3-0 count for Greg; in 12 of them he threw two-seamers while the other two were changeups. Never one to give in, Maddux challenged Molina to swing on 3-0, with a cutslide literally right down the middle of the plate. Molina held for strike one. Maddux followed with another picture perfect, down and away two-seamer, that Molina fouled off.
With the count full, Maddux opted to throw the cutslide on the inside corner. It caught a bit more of the plate and Molina again laced a single to left, this time bringing in a runner to score. At this point, Maddux had adjusted his selection and location yet Molina had lined singles to leftfield in three consecutive plate appearances.
Molina led off the fourth inning, taking an outside two-seamer for strike one. Maddux went right back outside with another two-seamer, throwing an even better pitch than the first, but somehow did not get the call. With the count even 1-1 Maddux again went outside, this time with a changeup. The change in velocity combined with the very similar movement to fool Molina; he swung early and grounded out to shortstop.
In the sixth inning, Molina came up with two outs and nobody on. Maddux started him off with a cutslide down and away. His modus operandi with Molina thus far seems to favor the down and away location. He came right back with the same pitch but located it very nicely slightly off the plate towards the inside corner. Somehow, and I have no idea as to the “some” of that “how,” this pitch was called a ball. With a 2-0 count he went back outside, this time with a changeup. Molina took for strike one. Maddux then went right back outside, this time with a fastball. Again, the change of speed but similar movement fooled Molina into grounding out weakly to shortstop.
I’ve mentioned the similarities in movement between his fastball and changeup in Molina’s plate appearances and here is how it looked:
movement3.JPG
In theory, changeups with the same movement and release point as fastballs will be extremely effective because it will look like the same exact pitch. If the only difference is velocity then hitters will be seeing virtually the exact same pitch coming towards them.
Maddux vs. Molina: 4/23/2008
Maddux also faced the Giants two weeks later, providing three more plate appearances against Molina with which to analyze. In the first inning, Molina came up with a runner on second and two outs. He grounded out to shortstop to end the inning. He then led the fourth inning off with a groundout to third base, and ended the sixth inning with another groundout to third base. This would be the only Pitch F/X documented game in which Molina did not victimize Maddux via a hit. Here is the pitch location breakdown:
madduxmolina42308.JPG
Maddux started Molina off with two straight two-seamers his first time up. The first just barely missed the outside corner for a ball; the second badly missed. With a 2-0 count Maddux threw a cutslide, letter high, that Molina took for a called strike. He went back outside, throwing a changeup to the same location as the initial two-seamer. Molina fouled it off to even the count. Despite hitting a similar location, the movement of his changeup differed from that of the fastball.
Greg followed that offering with a two-seamer, up and in, that brought the count to 3-2. He then threw a picture perfect, down and away two-seamer that Molina put on the ground to shortstop.
The second plate appearance proved to be much easier for Maddux. After throwing a two-seamer taken by Molina for strike one, he threw another slightly off the plate. Molina reached out and pulled a grounder to third base.
The final time up, Maddux started Molina off with a cutslide and followed it up with another outside two-seamer that Molina again grounded towards third base. Here is the movement chart:
movement4.JPG
Overall
Of the 10 overall plate appearances documented here, Molina put the ball in play every time, hitting nine balls to the left side of the field. The only ball he hit towards the right side was a weak grounder to second base that resulted in a fielders choice. Otherwise, he grounded out to shortstop three times, to third base twice, laced three singles to leftfield and got a good hold of a ball, doubling to deep leftfield.
Maddux missed his spots a few times and did not get certain calls from the umpires that we would generally expect to be given to a veteran with pinpoint accuracy.
Given that Molina is very selective with pitches outside of the zone it made sense for Maddux to utilize his strike zone accuracy as much as possible. Since he is against throwing 0-2 waste pitches it becomes very likely that the pitches far out of the strike zone were mistakes rather than the results of strategic planning. Molina seemed genuinely fooled when Maddux’s changeup movement matched that of his two-seamer.
Though Molina has experienced recent success in the form of hits, if Maddux’s goal involved directing balls towards the left side of the field then he succeeded. It will be very interesting to see how Maddux pitches to him next time as this most recent game consisted of three groundouts.  Though the only discernible pattern in the pitch sequencing here can be found in his desire to hit the inside corner it will definitely be interesting to see if he approaches Molina with the same sequences that induced these groundouts in the next Maddux-Giants matchup.
For next Saturday we will look at Maddux vs. Dave Roberts.  Instead of something as lengthy as this I would prefer to analyze the difference in approach from righty to lefty.  Next Thursday I will look at his supposed distaste of 0-2 waste pitches to see what he likes to throw, where he likes to throw it, as well as how often this anti-conventional-wisdom strategy hurts him.

Advertisements

13 Responses to Mound Chess: A Look at Greg Maddux’s Pitch Sequencing

  1. Eric – Cool stuff. Looking forward to part 2

  2. I sure hope so since this took a LONG time to do, haha.

  3. Pizza Cutter says:

    This here is the new frontier in what pitch f/x can do. I’m looking forward to part 836…

  4. It’s something I’ll be able to explore more going into next season, when much more data is available. Ideally I’d like to be able to combine video footage (I may be able to via Redlasso for certain matchups) and the data to show a mixed method of what really happens. We can see a lot from the graphs and data but we can gain even more from watching game as well – we can see Molina’s body language, if Maddux missed targets, etc.
    I only watched one of these games but that’s definitely something I’m going to explore for a future article. Embed videos of each at-bat as well as show the PFX data.

  5. John Walsh says:

    Great stuff, Eric. Breaking down a series of at bats like this is really fascinating.
    Just a couple of comments: AT&T Park was not one of the original PITCHf/x parks, but came online in the course of the 2007 season. That is why the June 27 game is not available.
    Also, I’m pretty sure you’re confusing inside and outside on the location plots. Your plots are from the catcher’s viewpoint and since Molina bats righty, the pitches towards the right of the plot are outside, not inside, as you have it.
    Look forward to the next installment.

  6. John, thanks for the info. That definitely explains the June 27th info. I had been under the impression it was.
    Yeah, I realized the inside/outside as soon as I published and am working now to change the text.

  7. All inside/outside fixed.

  8. Geoff Young says:

    I’m late to the party, but this is fascinating and fantastic analysis. Nicely done, Eric, and thanks!

  9. Thanks, Geoff. I’m doing something on Maddux for Thursday, too, looking at his 0-2 and 1-2 pitch selection since he hates waste pitches.

  10. […] * * Eric Seidman at Statistically Speaking takes a good, hard look at Greg Maddux’s pitching strategy. Eric has broken down several confrontations between […]

  11. I’m even later to the party, but I’d like to compliment your efforts too. Here’s a question, how many outs has Maddux gotten on that tailing fastball you mentioned to lefthanders? 1,000 maybe? Heck, he has over 7,000 groundouts in his career.

  12. Guillermo says:

    Great job.
    Love the analysis.

  13. Guillermo, thanks. I have another on Maddux up right now, and then the final one in the series of three will be up Saturday afternoon.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: