Anatomy of a player: Carlos Marmolby Josh Kalk
July 08, 2008
|Fastball? Slider? Or is Carlos Marmol's arm falling off? Cubs at Blue Jays. June 14, 2008 (Icon/SMI)|
Carlos Marmol has an electric arm. Just watch him pitch once and you will come away impressed. Apparently his manager, Lou Piniella, is impressed too; he has used Marmol in more games and more innings than any of his other relievers. Marmol is tied for the sixth most appearances by a reliever this year.
Recently, Marmol has had some troubles on the mound leading to suggestions that he may be fatigued. Because of the wealth of PITCHf/x data this year, we can track him throughout the year and see if fatigue really is the problem. But first, here is some background on Marmol that I think will prove useful.
Carlos Marmol, from the Dominican Republic, signed as a free agent with the Cubs in 1999 when he was 17. He is a right-handed pitcher who made his minor league debut in 2002, throwing one game for the Arizona Cubs. In 2003, as a 20-year-old, he spent the entire year in Arizona racking up strikeouts and the walks over 65 innings.
The next year, he was promoted to Single-A in the Midwest League and inserted into the rotation. At Lansing, he threw 154 innings and had a 3.20 ERA with nearly a strikeout every inning. At 21, he was a little old for the league so he didn't surface much on the scouts' radar.
In 2005, Marmol headed to high-A ball and pitched well in the Florida State League, earning a midseason promotion to Double-A. That year he posted a solid ERA under 3.50 but walked just over 4.50 per nine innings. His name was bandied about and he made the Baseball America Cubs top 10, but control and his smallish stature were marks against him.
Finally, in 2006 as a 23-year-old, Marmol busted out. He started the year at Double-A, reduced his walk rate to less than four a game and raised his strikeouts to more than 10 a game. That performance got him a look with the Cubs.
Sadly, Marmol flopped in his first exposure to The Show. He appeared in 19 games, starting 13, with an ERA above six and just as many walks as strikeouts. Even then, you could see the life Marmol had on his pitches, but his inability to throw the ball over the plate kept him from success. Worse, Marmol had to spend time on the DL at the end of 2006 with shoulder fatigue. He was dropped from Baseball America's Cubs top 10.
Finally, last year, Marmol put everything together. He started the year in the rotation at Triple-A Iowa and after 41 innings was called up to the bullpen for the Cubs. This time Marmol, used in low leverage situations at first, hit the ground running. By the time the year ended, Marmol was the main setup man for then-Cubs closer Ryan Dempster. Marmol ended the year with a cool 1.43 ERA and more than 12 strikeouts per nine innings. The walk total was still relatively high, at 4.53 walks per nine innings, but if you are striking out that many batters it really doesn't matter. His performance was even good enough to earn a MVP vote.
In 2008, Marmol has been the key setup man for the Cubs, leading the relievers in WPA and with a very high leverage index and boosting his strikeouts to more than 13 per nine innings. Piniella went to Marmol early and often, having him throw 16 innings in April and 18.2 innings in May. This put Marmol on pace for more tha 100 innings, but he responded with an ERA under two and the league lead in holds.
In June, Piniella pulled back on the reins and Marmol threw only 11 innings, but ended the month with a 7.36 ERA thanks to a couple of home runs and "only" 12 strikeouts and nine walks (!) during that span. July has started shakily as well; he gave up a three run lead in San Francisco though eventually getting the win. The next night he came back and threw but three pitches, giving up another long ball in a Cubs loss.
He has somewhat righted the ship with two solid outings in St. Louis, but concern about Marmol is mounting. There were questions about whether Marmol might be tipping his pitches or possibly going through a "dead arm" period, but the Cubs dispute both those claims. If that isn't what is going on with Marmol, then what is?
Before we tackle what is ailing Marmol, we need to look at what he throws. Marmol is a two-pitch pitcher with a fastball that tails in to right-handed batters and a slider (?) curve (?) slurve? Let's look at his movement chart to find out.
Here you can see what kind of movement Marmol is getting on his pitches. Whatever you want to call that off-speed pitch, it is absolutely devastating, with more than six inches of horizontal movement and nearly two inches of vertical drop. When Marmol starts that pitch on the outside corner, it falls off the table and is unhittable.
What should we call it, though? John Walsh and my pitch classification algorithm call this pitch a curve because of that vertical drop. You don't see a drop like that from a slider. Marmol, however, calls it a slider, though slurve is probably the best description. From here on, I'll call it a slider, since that is what Marmol calls it.
His fastball bores in to right-handed batters, moving nearly eight inches horizontally while still getting about eight inches of vertical "rise." Coupled with the nearly 95 mph velocity of the pitch, that makes it death to right-handed batters. Putting it together, you get two pitches that move more than a foot difference horizontally and nearly a foot difference vertically with a speed difference of nearly 15 mph. Because of his low arm angle, only a minimum hump is formed, deceiving the hitter even more.
What makes things even worse for the hitter is that Marmol throws each pitch almost exactly 50 percent of the time, with little variation due to the count or the handedness of the batter. This keeps batters on their toes and makes guessing a very dangerous game for the hitter.
Okay, so we've seen what kind of stuff Marmol has. The question is how it's changed over the season. To look at that, I am going to make what I call a wear pattern for Marmol. That's a look at a single pitch type as it varies as the season goes on. For each game Marmol threw, I calculate the mean and standard deviation of the pitch type for the speed and horizontal/vertical movement and put everything together in one plot. Here is the wear pattern for Marmol's fastball.
The x axis is the number of the day of the year, so 91 is March 31, 121 is April 30 and so on. The velocity of Marmol's fastball is the black dots and corresponds to the y axis on the left. The horizontal and vertical movements are in red and blue respectively and correspond to the y axis on the right. This way we can look at all three variables together and see how they are changing as the season goes on. This shows each game up to July 5. I also want to point out that this is corrected PITCHf/x data, so park variations have been removed.
Let's start by looking at the speed of Marmol's fastball. Marmol began the year near 95 mph and stayed there until a dip around day 145 to day 160 (May 25 to June 8). This was at the tail end of the 18.2 innings Marmol pitched in May and then the start of June, where Piniella backed off and used Marmol much less. His fastball speed dropped 93 mph at the tail end of a Cubs home stand and their west coast trip to San Diego and Los Angeles. After a few days off back home, Marmol's velocity came back and he has been hitting 95 again. So it appears Marmol was tiring, then when giving a lighter workload his velocity came back and everything is fine, right?
Sadly, the picture isn't quite that rosy for the Cubs. If you look at the horizontal movement, an alarming feature pops out. Marmol began the year with his standard huge horizontal movement in to right-handed batters (the more negative, the more the ball moves in to a right-handed batter). As the year has progressed, however, he has slowly lost some of that movement right up until May 25, when the trend is reversed and his horizontal movement comes back for a while only to have it disappear once again. This transition happens at exactly the same time Marmol lost a few mph on his fastball. So what is going on here?
Pitchers have a fine line to walk as far as movement and velocity go. As a pitcher lowers his arm angle on his fastball, more of the backspin he imparts on the ball turns into sidespin. It is that sidespin that moves the ball inward to a similarly handed batter. In Marmol's case, because he throws nearly sidearm, a big percentage of that spin is sidespin and results in huge horizontal movement.
So why don't all pitchers throw like this? Because, generally, the lower the pitcher's arm angle the less speed he has on his fastball. The fact that Marmol can throw 95 mph and bore his fastball in so much puts him in elite company. Many pitchers can do one or the other, but only a few can do both.
Notice how correlated the horizontal movement and speed plots are, especially after May 25. It is almost as if at that point Marmol or the Cubs realized that he wasn't getting the same bite on his fastball and he consciously altered his arm angle back to what it was in the beginning of the year to get that movement back. What happened, though, was the velocity suffered so he now he has decided to go back to the higher arm slot, regaining the speed at the cost of the movement. This points to Marmol possibly tiring in mid-May hiding it by slightly increasing his arm angle until May 25, when he tried to go back. It also seems to show that even after a light workload in June, Marmol hasn't fully recovered.
What about his slider? Do we see the same things there?
Marmol's slider took a little while to warm up, but peaked in early May (day 120 to 135). Again we see a large drop in horizontal movement around day 140. That has somewhat stabilized recently, though not quite to peak levels. Again, you can see good correlation between the speed of Marmol's slider and the horizontal movement. Notice that around day 160 he started throwing his slider faster and the movement suffered as a result. All in all, though, it looks like Marmol's slider is pretty much back to normal, with the large speed and movement difference from his fastball.
For more on Carlos Marmol's struggles, check out Harry Pavlidis' excellent blog post.
The $430,000 question is which Carlos Marmol the Cubs will have for the rest of the year. While Marmol threw more than 100 innings last year, more than a third were as a starter in Triple-A and some were in low-leverage situations when he was called up. He has a history of arm fatigue and it looks like he is fatiguing some now.
Marmol began the year on fire and had a terrific April and May thanks to incredible speed and movement on his fastball and his electric slider. As the calendar turned to June, Marmol's stuff began to slip and he paid the price. While his slider appears to have regained form, his fastball still isn't back. While a much straighter 95 mph fastball still is an effective pitch, it isn't anything like what he was throwing previously.
I don't expect Marmol to regain the elite fastball the rest of the year because he didn't seem to regain it after a light June and it isn't likely he will have more rest than that from here on out. He is going to have to learn to get big league hitters out with just a very good fastball. Obviously that can be done, but it will be a transition for him. Expect his strikeout rate to drop back down closer to 10 per nine innings than the 13 he has posted.
The walk rate will be the big key. In the past month Marmol has put on nearly a batter an inning via walks and hit-by-pitches. While I expect that to go down, I do expect him to continue to walk a batter every other inning or more. That is playing with fire and with his straighter fastball he is more than likely going to continue to get burned with the long ball from time to time. This combination of walks and more home runs will likely cause Marmol to be very erractic, throwing up a few zeroes, then exploding and give up three or four runs, then going back to throwing goose eggs. June will likely be a typical month where he looks great in the middle of the month then struggles at the end of it.
So how should Piniella use Marmol, assuming the above is correct? As strange as it sounds, Marmol should be used in even higher-leverage situations than before to get the best results. While this may seem counter-intuitive, here is an example of why this is the case:
Imagine it is the bottom of the eighth in a tie game and the Cubs are on the road. They have about a 40 percent chance of winning, according to Win Expectancy. If they can get through the eighth without any runs scored, their WE goes up to 50 percent. If they give up one run, their WE goes down to about 16 percent. That obviously is a huge difference, but what if they give up three runs in the bottom of the eighth? Now their Win Expectancy goes down to about 3 percent.
So giving up any runs is a huge blow but giving up three isn't terribly worse than giving up one. What if instead the Cubs are up by three runs? Now their WE going into the bottom of the eighth is nearly 92 percent. If they put up a goose egg it goes up to about 97 percent, but giving up one run still increases their chances of winning and they would sit at 93 percent going into the top of the ninth. Conversely, if they give up three runs they would be 50/50 going into the ninth.
So if you have a shaky reliever like Marmol, the time to use him is in high leverage situations when putting up a zero makes a huge difference but the difference between giving up one and say three runs is small. The worst time to use him would be a situation where you were up three in the bottom of the seventh. In situations like that he is more likely than a steady reliever like Bob Howry or say Neil Cotts, depending if you want a right- or left-handed pitcher. Those guys aren't as likely to throw a perfect inning but they are also less likely to blow up and give up the entire lead.
As long as Marmol is used when the game is on the line, he still can be a very effective relieve for the Cubs. If they lose confidence in him and put him in more middle-leverage situations, when he does blow up it is only going to look worse. Fortunately for the Cubs, it appears that Piniella generally realizes this: He used Marmol in a one-run game in the eighth inning over the weekend.
The Cubs have to be careful not to overuse Marmol. It seems pretty clear that when he is worked hard his stuff suffers, so Piniella needs to find that sweet spot where Marmol is still effective but not overused, or else we might see even more degradation in his pitches.
The Cubs are very likely going to make the playoffs, and while Marmol is a key member of the bullpen, he is far from the only one out there who can get outs. When the playoffs roll around, if Marmol is in peak form he could be a dominant weapon for the Cubs, shortening games with the many off days in the playoffs. The key for them will be getting him to the playoffs in that peak form.
I have been very impressed with Lou Piniella as the Cubs manager. He seems to have pushed all the right buttons so far, so it will be interesting to see if Piniella can keep pushing the right buttons with Marmol.