Frightened to Swing? The Odd Approach of Batters to facing Carlos Marmol

Carlos Marmol is an awesome pitcher.  His astronomical K rate has been talked about much at length, as has Marmol's wildness that leads to his high walk rate.  And yet there's an interesting fact about Carlos Marmol that you might not know:

Marmol's rate at hitting a wide strike zone (58.14%), while below the league average in 2010 (61.65%), is actually not that low or even in the top 50 lowest rates of hitting the zone!  (For the five lowest, see last week's article on Derek Lowe)  Meanwhile, it's not that the rate that batters swing at pitches out of the zone against Marmol is really low (or high) either.  These things make Marmol's walk rate look a little inexplicable.  What's causing that?  

The answer:  Batters aren't swinging at Marmol pitches when they're IN THE STRIKE ZONE.  Well to be exact, the swing rate of batters on pitches IN the strike zone (also known as his Z-Swing Rate) against Marmol has been the lowest in the League for the last two years.  You can see that in the graph above.  So while this does result in Marmol having the highest rate of called strikes in the league, it also extends at bats against Marmol, resulting in more walks and strikeouts.  

After the jump I'll examine why the question of why Marmol's Z-Swing is so low and how it's affecting things.

As you should know, Carlos Marmol, throws two pitches:

A fastball that averages 94 MPH with 6 inches of tail in on right-handed batters and 8 inches of "rise."  The fastball thus doesn't have great movement, but the velocity more than makes up for it.  Marmol hits the zone with this pitch only 59% of the time against left-handed batters, though he gets a good rate of 67.8% against right-handed batters.

A slider that averages a little over 83 MPH with 4.74 inches of tail in on LEFT-handed batters and 2.2 inches of sink.  The horizontal movement on this pitch is really good relative to the fastball, and Marmol uses the pitch actually more frequently than the fastball.  Meanwhile, he gets it in the strike zone a surprisingly high amount of the time: 51.9% of the time against left-handed batters and 56.50% of the time against righties. This pitch is, as you should know, absolutely devastating, getting a whiff  50% of the times that left handed batters swing and 38% of the times that righties swing at the pitch.  That's utterly ridiculous.

Oddly enough, in 2009, when Marmol had a higher rate of walks, his rate of hitting the strike zone with each of these pitches was HIGHER than it was this year.  

Now one thing you might think is that, well: his lower z-swing is probably because of simply the fact that Marmol uses the slider more often than the fastball.  But the fastball's Z-Swing rate, while higher than the slider's, is lower than 50% in itself!  As you can tell from the graph up top, basically every pitcher who qualifies gets swings at all pitches in general over 50% of the time...but Marmol doesn't even get that from his fastball! Note also that the slider's Z-swing rate (48.6% against left-handed batters, 45.3% against right-handed batters) is far below the average Z-swing rate of 60.1%.  So even just using the slider more often probably doesn't account for Marmol's extremely low rate of batter swings at pitches in the zone.

Okay so maybe Marmol's Z-swing rate is caused by most of his pitches coming in the outer edges of the strike zone (after all, he's known to be wild), and thus places where batters are less likely to swing.  The heat maps below show where Marmol's pitches to left and right-handed batters most frequently end up: 

Man, that's an awful lot of pitches in the middle areas of the strike zone.  So this explanation for Marmol's low Z-swing rate doesn't quite fly.

Well perhaps Marmol's Z-Swing Rate is due to him getting more frequently into counts where batters don't swing as often (presumably high ball, low strike counts).  To see if this is the case, lets look at the Z-swing of Marmol by the count and compare that to the MLB Average in 2010.  The graph below compares the two rates (Note:  the graph below includes Marmol data from 2009 in addition to 2010):

As you can see, the count doesn't seem to explain this totally....Marmol's pitches in the strike zone are swung at less frequently than the major league average no matter what the count. Still, this effect is MUCH more noticeable as the count becomes less and less favorable to Marmol, to an extreme point once he enters two ball counts. 

Let's take a look at how batters swung at Marmol's pitches over the last two years in each ball-strike count, with a graph for each count (except for 3-0). 

All of the below graphs show are read from the catcher's point of view: right-handed batters would be on the left side of the graph while left-handed batters are on the right.  Red dots represent fastballs while blue represent sliders.  The DARKER dots of each color are pitches that batters swung at, while the LIGHTER dots of each color were watched.

You can see here that batters take a large amount of pitches that are in the strike zone, including many fastballs that seem to be in the middle of the strike zone.  Batters do seem to swing at some of the middle-high fastballs and a few of the low and away-from-righties sliders, but a large majority of the time feel free taking pitch 1 from Marmol (Z-Swing is 29%).  This is not uncommon (the league average Z-swing on 0,0 counts is only 34.5%), but it is still lower than the league average by a decent bit. 

On 0-1 counts, we of course notice an increase in the amount of times that batters swing against Marmol, and we start to see a whole bunch of sliders out of the strike zone that batters have swung at.  We also see that what seems to be a good majority of fastballs in the zone on this count are swung at by batters.  Still a decent number of sliders are ignored by batters on this count, but notice the location of these pitches - the inside part of the plate for right-handed batters and frequently in the higher parts of the zone.  This is almost certainly the result of batters being unable to pick up HOW MUCH the slider moves in on lefties/away-from-righties.  In other words, these are the areas of the strike zone where the movement of the pitch takes the pitch from OUTSIDE of the zone to just inside the strike zone - resulting in backdoor strikes.  Similarly, the high-in-the-zone sliders are pitches that look at first to batters like they're going to wind up high, but drop into the zone.  Thus batters opt not to swing, thinking these will miss the zone (for the same reason, batters are swinging at Marmol pitches that are below the zone, as these look like strikes before they drop out of the zone).  It's these pitches, it appears, that results in Marmol having a lower Z-Swing, by 8.6%, than the Major League Average.

Here naturally the fastball disappears and makes way to the slider, and the number of pitches in the zone naturally drops.  Regardless, batter swings go up on 0-2 counts, and that's the case against Marmol as well.  Essentially the only pitches not swung at here are low fastballs that hit the corner (surprising batters) and a few high sliders.  The gap between Marmol and the average MLB pitcher's Z-Swing on this count is tiny: a difference of only 1.3%.  Batters simply have to swing on these counts. 

In this count, there's a major difference between Marmol and the MLB Average Z-Swing: batters swing at the average pitcher's in-zone pitches 52.3% of the time, but swing at Marmol's in-zone pitches only 34.4% of the time!  As you might expect, we see batters are taking a pass on high-in-the-zone sliders here.  But batters are swinging at high-in-the-zone fastballs, but are letting go a bunch of fastballs that are basically in the MIDDLE of the strike zone!  Similarly, while batters are swinging at some sliders in the low and away-from-righties part of the strike zone, they're still passing on a good amount of these pitches!  These are pitches which we don't expect to fool batters (they look like strikes all the way), but batters are laying off them anyhow!

One more thing:  you should also start noticing here that Marmol actually pitches BACKWARDS - where nearly all pitcher throw the fastball more often as they get into worse counts, Marmol leans on his slider.  Since Marmol has a pretty good accuracy rate with the slider, that's not a terrible idea.  Still it's odd to see more sliders in counts like 1-0, or 2-0, and this might explain to some extent why Marmol's Z-swing in these counts is so much lower than the major league average.  Still, given that batters, as stated above, aren't swinging at fastballs or sliders which they should pick up as strikes, this doesn't seem to be a complete explanation. 

The difference between MLB Average Z-Swing and Marmol's Z-Swing on this count is still pretty high (10.6%).  Interestingly, batters seem on this count to be swinging at more sliders in the upper half of the zone.  Of course, there are a whole bunch of sliders taken in the top left corner of the strike zone (in-on-righties).  Still there are other sliders (and a few fastballs) taken all over the zone.  Given that top left corner, we can chalk of at least a good bit of the difference in Z-Swings once again to batters being fooled on backdoor pitches.

Once again, on a 2 strike count, the difference between Marmol and the average pitcher's Z-Swing drops significantly: batters swing at 75.9% of these pitches from Marmol inside the zone, while they swing at 80.6% against the MLB average pitcher. And the small gap would once again seem to result in Marmol's pitches being hard to pick up in the left part of the strike zone where they wind up being backdoor sliders (and fastballs).  

Here's where things start to get interesting.  First, as mentioned above, despite being in a count where Marmol really could use a strike, Marmol's fastball is rarely seen on this count.  Second.....batters are almost never swinging on this count, swinging at just 16.4% of his pitches, when the major league average on this count is 50.4%!  That's a difference of over 34%!  Now some of this may be the fact that Marmol is throwing far less fastballs on this count than the average pitcher, but many of Marmol's sliders are in the middle of the zone and are simply watched!  Surely this can't be the full explanation for this huge difference!

My belief is that the answer is in fact quite different:  Batters in favorable counts (starting actually in 1-0 counts, but really noticeable from 2-0 counts on) have simply decided: "Ya Know what? This guy walks a ton of people, and I'm in a good count, I'm not going to try and swing and hit his crazy slider!  I'm just gonna let this pitch go by!"  In other words, this effect is caused mainly not because of any deception in the pitch's movement or his delivery, but by batter "cowardice."  (Alternatively, you could call it strategy). 

Once again, on 2-1 counts there's a huge difference between the major league average Z-Swing (73.9%) and the Z-Swing of Carlos Marmol (51.1%).  It's a small sample size, but it seems like batters are generally swinging at the fastballs in the zone on this count.  What they're frequently not swinging at are the sliders in the zone.  Once again I suspect that while some of the difference is caused by the deception of Marmol's slider, a good bit of Marmol's low Z-Swing is explained by batters deciding ahead of time not to swing if it's the slider. 

On 2-2 counts, like the other 2 strike counts, things are more normal: Marmol's Z-Swing is 76.1%, compared to a league average of 84.4%.  This of course makes sense: if batters are taking pitches in 1-0, 2-0, 2-1 counts just because they're hoping for balls, they're certainly going to stop doing so when a taken strike will result in a K. 

Here, once again we have batters opting not to swing at pitches in the zone at a far greater rate than they do against the average pitcher.  Marmol's Z-Swing in this count is only 37.5%, while the major league average is 66.9%.  The reason is quite easy to see as well: on 3-1 counts, batters did not swing at a single slider that was inside the strike zone against Marmol (though they did swing at two that were low); these batters were clearly looking dead red.  Once again it seems that batters have decided that they're not going to chance swinging at the crazy slider and that they'll hope it winds up called a ball.

I've skipped 3-0 counts (for the obvious reason that it doesn't tell us anything)

And of course, once again in a 2 strike count, we see a small difference (9.6%) in Marmol's Z-Swing (77.1%) as compared to the Major League average (86.7%). 

Conclusion: 

The fact that Marmol gets the most called strikes of anyone in the majors (the inevitable result of having these pitches taken by batters) should not be over-emphasized.  It is not the main reason for his insane K rate.  Remember, Marmol has one of the highest (#9 out of all qualifying pitchers) swinging strike rates in the league, and there's a high correlation (R^2=.666) between swinging strike rate and K%.  Meanwhile, the correlation between a pitcher's called strike rate and K rate is zero (R^2=.006).  This shouldn't be surprising if you've been following along...after all, the swing rates against Marmol on 2 strike counts are still quite high and close to normal, showing that the vast majority of these called strikes come on counts where they can't result in an at-bat being over. 

That said, it does explain why the K and BB rates of Marmol are so extreme, especially for a pitcher who actually does hit the strike zone not at a super infrequent rate.  Batters when they find themselves ahead in the count, particularly in 2 ball counts (but starting in 1-0 counts), are opting to take pitches rather than to try and hit Marmol's elusive slider.  The end result is that batters are putting less balls in play than they would otherwise and thus are having more opportunities to take walks or to strike out.