A Tale of Two Cain's
April 29, 2008 3 Comments
When Brian Sabean signed Aaron Rowand this offseason it became clear that the San Francisco Giants had no clear direction. Signing an average hitter coming off of a career year in a hitter’s park to spend five years in a pitcher’s park just did not make much sense. This lack of direction put even more pressure on the dynamic young arms in their starting rotation. With Barry Zito on pace to break the single-season losses record a rotation featuring Tim Lincecum, Matt Cain, and Jonathan Sanchez needed to step up.
Lincecum and Sanchez, by most accounts, have been better than expected. Cain, however, is somewhat of a different story. He looked brilliant in three of his first five starts and terrible in the other two. Now, most of the readers here know I am a big fan of his; the fandom began as pity for his extreme lack of luck and blossomed into just really enjoying watching him pitch. Since he has had a topsy-turvy season I decided to get my feet wet with Pitch F/X data to provide a scouting report on him.
Carlos Gomez, former writer for The Hardball Times and current scout for the Diamondbacks, wrote a tremendous article breaking down Cain’s mechanics just about a year ago today. Cain has a windup very similar to that of Daisuke Matsuzaka, one that derives its power from balance, torque, and hellish momentum. As Carlos points out, the lower body completely uncoils his upper body; this, along with the loading of his shoulder, helps increase his velocity.
One of the reasons Cain can occasionally struggle with control deals with a problem at his release point: He tends not to meet the glove with his body but rather brings the glove towards him. This happened to be something I noticed from watching some starts this year, prior to reading the mechanics column, and was very happy to read something supporting my eyes. Cain’s release point has seemingly been off this season which may have contributed to his 18 walks through five games.
Since there were some issues with classifying pitches within the PFX algorithm I decided to mix my methods. Basically, anything that went unidentified or seemed fishy was corrected through watching the games and making due corrections. The incorrectly classified pitches mainly tended to be changeups. The algorithm recorded some of these changeups as splitters; however, from watching Cain pitch as well as reports on his repertoire, it is evident a splitter is not part of his arsenal. His changeup is more of a circle-change, though, one that offers significant sink; probably the main reason it was classified as a splitter in some instances.
This brings up an interesting point that I want to explore more in the coming weeks but essentially, the algorithm classifies the pitch based on what pitches similarly classified will do; regardless of what we call it, the pitch looks like a splitter due to its sink, break, velocity, and spin. If I call it a chanegeup the pitch is not any easier to hit. Regardless, Cain throws a fastball, slider, and changeup, with the curveball used sparingly. He has significantly reduced the usage of a curveball since his minor league tenure and early major league upbringing. He has thrown so few curves, in fact, that I’ve actually excluded them from many of the charts to come later. He has always relied on his fastball and this year is no different. Here is his location breakdown to lefties and righties:
And here is a look at how he has distributed the three major pitches amongst these hitters:
He clearly favors the slider against RHH and the changeup against LHH. With a circle-change with as much as sink as his it makes sense to use it primarily against lefties; it tails away from them. He appears to throw his fastball a bit more often against lefties but the discrepancy is much closer in the early going than the graph would suggest.
Last year, as documented by Josh Kalk’s card as well as Chris Quick’s article at Bay City Baseball, Cain threw his fastball at least 50% of the time in all counts except for 1-2. In the early going of 2008, Cain has thrown his fastball much more, with a minimum of 58% in the same 1-2 counts. Reverting to his slider more often in these counts he has also recorded the highest percentage of his strikeouts. Here is his current graph of pitches by count:
As mentioned, he has upped his fastball usage thus far, starting 80% of batters off with one. He has virtually cut the curveball out of his diet, failing to throw it on 0-2 counts and any count with two or more balls on the batter. Yes, that line sounds funny in retrospect, but no more innuendos.
The area of pitch data I am going to focus on the most involves sequencing, IE, what does a pitcher throw after a certain pitch? After a series of pitches? Does this differ between batter types? Between handedness? Once the 2008 data is corrected I would love to explore this more in-depth via location quadrants: up and in, down and in, etc. For now I will look solely at the pitches thrown. The following charts show the percentage of pitches that follow a given pitch, broken down by lefties and righties. To further explain, for anyone unfamiliar, in the LHH chart, in the ‘CH’ row, the number 4.5% refers to the fact that, against lefties, Cain has followed a chanegup with a slider 4.5% of the time so far:
As expected, he follows all of his pitches with a healthy dose of fastalls and, when he differs, the results come in the form of more changeups to lefties and more sliders to righties. Another aspect of sequencing that fascinates me deals with determining the likelihood of a pitch based on the previous two pitches. The charts below do just that:
This shows somewhat of a break in the standard mold, as Cain has followed each sequence to LHH with more fastballs than against RHH. Now, sequencing raises various questions involving the independence of pitches, mainly whether or not each pitch should be considered an independent commodity. Based on all of the planning, technology, and strategy in the game today it is naive to think that pitches are merely being thrown in random fashion but these results definitely further his extreme reliance on the fastball. Luckily for Cain–something I never imagined I would ever write–he has a plus-fastball due to taking advantage of a mechanical trick as outlined in Carlos’s breakdown. When he drives towards the plate, his lead foot appears to be stepping over something; this adds a few miles per hour.
Power vs. Contact
I remember reading a comment left on one of Mike Fast’s old articles calling for splits against power or contact hitters. When determining who qualifies as a power hitter it would be irresponsible to solely look at current results. Even though one could argue that Rafael Furcal, on his hot streak this year, will not be pitched to the same way as years past, the usage of past performance in determining power/contact offers a concrete baseline with which to base judgments. Using TangoTiger’s Marcels I classified everyone with a .440+ SLG as power and anyone under that to be contact. The baseline can be debated and I’m open to changing that but, for the sake of this article, it works for me. Below is a chart detailing his pitch breakdown vs. each type of hitter:
Against power hitters he has incorporated his slider more at the expense of his fastball while the adverse can be said for contact hitters. Here are his results against each type of hitter:
Neither group swings and misses much at his fastballs and his percentage of balls and called strikes are similar as well. Power hitters are fouling more of them off, though, and putting less in play than the contact hitters.
Jekyll and Cain
Going back to sequencing, the following two charts show the likelihood of a pitch after a given pitch, first against contact hitters and then against power hitters. Each chart is broken down into whether or not his start was good or bad. Before Monday’s game against Colorado his starts had followed the pattern of: good, bad, good, bad, good. Though small sample sizes are at work we can still look at what has happened thus far. We cannot make concrete judgments on what he is doing wrong but rather can evaluate what has happened as potential causes for concern down the road:
Against contact hitters, his good starts have seen him incorporate more changeups than sliders while throwing a higher percentage of fastballs. When dealing with power hitters, though, his good starts are seemingly counterintuitive to the commonly thought of approach. When he mixed up his sequencing he ended up with worse starts than when he stuck to his heater. Again, we cannot draw conclusions but rather explore potential concerns down the road.
Cain Meets Duncan
Another interesting aspect of Cain’s season to date is that he has had two starts against the Cardinals and two against the Padres; against each he had one good and one bad start. I wanted to see if adjustments were made and analyzing Chris Duncan made the most sense. On April 12th, Duncan grounded out and struck out in his last two at-bats vs. Cain. on April 18th, in his first two, he hit a ground-rule double and a two-run homer. Below is a pitch chart for those at-bats. The home run came on the upper slider while the double came on the hanging changeup in the middle of the zone:
None can be made so early into his season but it seems from the early data that he does not have a ton of confidence in his breaking pitches and forcing himself to incorporate them in order to adjust to the Cardinals hitters proved detrimental. He made adjustments between the two starts against the Padres and shut them down in his April 23rd start, one that saw him throw predominantly fastballs, upwards of 88%. In five or six more starts more conclusions can be drawn but, for now, there are some interesting aspects of sequencing and pitch selection from Cain to explore down the road. Hopefully by then his season will have ended its topsy-turvy nature and followed a pattern of mainly good starts.