Meet the New Boss, Same as the Old Boss

Pundits nationwide have decried that the Rays have taken a step back going into 2013 in order to improve the team down the road via the James Shields for Myers and Friends trade.  There is no doubt that the team is better situated over the long run with Myers than without, but the first part is where the Rays fan is likely to take umbrage.  It’s very easy to see the team come back sans Shields and Upton and assume that the team is less talented, but digging deeper shows us that it’s likely the Rays already have a suitable replacement for Shields.

Alex Cobb has thrown 189 innings in his career staring down 793 batters and pitching to the tune of a 3.86 ERA and 3.82 FIP.  This equates to roughly a season’s worth of better than average work.  He has struck out 18% of batters faced while walking 8% and giving up an average against of .247.  We know that he gets ground ball.  A lot of ground balls.  We know that he throws mostly a three-pitch mix consisting of a curve ball, change up, and fastball.  We think the fastball is around league average.  We know the change up is really good.  We think the curve ball is pretty good.  Let’s leave the sun behind as we leave the surface and amble down where the gems are found.

First stop is where we decide to take a look at Cobb’s pitch usage and run values over each start of his career.  All data is courtesy of the now defunct Joe Lefkowitz Pitch F/x database and I have computed the run values using this great table:

Fastballs are the green area, the change ups are in red, and the breaking ball is blue.  The usage stuff lines up with the primary axis on the left and the RV/100 lines coincide with the secondary axis on the right.  Each vertical slice shows that individual start number and I have data for all starts except for one on 7/21/12.  For instance, Cobb’s sixth start of his career saw him throw around 12% breaking balls, around 35% change ups, and the rest were fastball variants.  For this particular start, like many others, his change up was his best pitch, though the fastball was also pretty good and his breaking ball was around league average, which is denoted by the horizontal black bar.  The vertical black bar shows how the calendar moved from 2011 to 2012 to give the reader a frame of reference.

The most obvious takeaways are that Cobb seems to use his breaking ball more as he has become more acclimated with the Show.  This has come mostly at the expense of his change up, though on any given day he may mix it up differently depending on what he has a good feel for that game.  We can see that his breaking ball seems to have the most volatility in results as it’s quite often his worst pitch (and it’s not particularly close in many instances) or his best pitch, particularly towards the end of last year where it could fluctuate to both ends of the spectrum on seemingly a start by start basis.  An interesting takeaway is that his change was a Pagan beast who’s name we do not mutter aloud over the first half of the sample, but over the back half we removed it’s mask to discover it is a mere mortal much like us.  By drawing claret from the offspeed stuff, however, it seems that the fastball showed better results and perhaps that is the reason for the breaking ball’s improvement as well.  Having a great pitch is one thing, but it’s a whole other animal to figure out how to optimally mix it in with the other two that aren’t half bad either.

Speaking of the fastball(s) we might have another reason why it seemed to become a better pitch over the course of the season:

Again, the vertical line connotes when we moved from 2011 to 2012.  On August 7, 2011 Cobb was put on the DL for shoulder pain that ultimately proved to be Thoracic Outlet Syndrome.  Though the condition is serious, and the surgery involves taking out either an entire rib or at least part of it, it is not nearly the death sentence common with most shoulder issues like a torn labrum or rotor cuff or bursa sac issues.  It certainly gives us a good reason for why Cobb’s velocity was down to start the year while almost continuously trekking northward.  It’s common for pitchers to increase velocity over the course of a season as the arm strengthens and stretches, but it should not be a wild hypothesis to think that Cobb will begin the year throwing faster than 90 MPH and improve on that as the year goes along.

Using the Pitch F/x data we can take a solid look at how his pitches move:

For the uninitiated, the above chart plots how each pitch moves relative to a pitch with no spin.  We know that gravity rides everything so we can adjust for that.  Thus, his fastball doesn’t actually rise 6-9 inches, but it does so compared to a ball thrown with no spin.  Conversely, his breaking ball is dropping around 10 inches compared to that same spinless pitch.  You’ll notice the area of convergence between the fastball and change up.  Part of this is that Cobb is likely throwing some two-seamers in there that trade rise for run, similar to his change up, and another part is that the algorithms that classify these pitches are more good than great.  Please keep that caveat in mind at all times.  Most of these pitches are correctly classified, but this chart gives a good example that sometimes pitches end up in the wrong camps.  The big takeaway here is to see where pitches cluster.  It seems like he can add and subtract movement to his fastball that can see the pitch move anywhere from 3 to 10 inches laterally.  The spread on the change is smaller, but he seems to show an ability to make these pitches dance.  This can be seen with the breaking ball to an extent, as well, but the pitches out of the dense mass are more like stragglers than purposefully different.  At least to this analyst’s eyes.

Let’s look at this again, but bring velocity into the equation:

Let’s first take a quick look at the horizontal movement and velocity spreads.  The area of convergence is smaller here once we take into account velocity, though it still exists somewhat.  This is important for one really big reason.  Pitches are at their best when they are distinct and different.  It’s one thing, and a very good thing at that, to look like another pitch before revealing it’s true nature, but it’s a very different thing to have two pitches acting alike, especially when it’s at a reduced velocity.  The change up that moves exactly like the fastball without much of a difference in speed is a pitch that generally gets abused.  If you find the approximate midpoints of these clusters you’ll see that Cobb’s fastball is around 91 while moving to the arm-side around six or seven inches.  The mid-point of the change is around 85 with similar movement.  That six MPH difference is enough to fool even very good batters, but when that number is more like three or four, well sir, most good batters are going to adjust.  It sounds minute, but that is the pin’s head margin that these guys are playing with.

In this next chart we’re again looking at velocity, but this time it is on the x-axis to give a more intuitive feel to the vertical movement differences between these pitches.  Our area of convergence is smaller here and we begin to get a notion that some of those fastballs should really be considered change ups.  Something interesting is that the spreads are smaller here.  Cobb, and perhaps all pitchers, have more ability to move the ball laterally than they do up and down.  Ask any old woman, but gravity is a bitch that is difficult to overcome while side-to-side movement is a more attainable goal.

Now that you have an idea of how his pitches move, let’s take a look at where he is throwing them from:

The picture is included to give an idea of what is being shown here.  This is a start by start look at his horizontal release spread for each pitch type.  Again, there is a divider to separate the seasons.  Also included are a few trendlines to show how these release points have drifted over time.  The fastball release point has moved further away from his body over time more than the other pitches and it really shows up towards the end of 2012.  Coincidentally, Cobb’s fastball also ticked up in speed over the course of 2012, but coincidence is all that’s going on here as the correlation between horizontal release and pitch velocity for the fastball was merely .035.  The change up similarly was released from further out from his body over time, though not as drastically as the fastball.  Meanwhile, the breaking ball was released from closer to the body over time.  This could be an avenue for future study to see if other pitchers experience this same evolution or if Cobb is on his own here.  Until then it’s difficult to draw more conclusions.

Looking at the vertical release point over each start seems a bit more intuitive.  Here we see that, over time, each pitch was gradually released from a higher point and that over the entirety he released his breaking ball from a higher point than his other pitches.  It’s not exactly tipping and we’re talking about a few inches from sixty feet away, but there could be something here.  Moving on from a possible tell, it’s interesting to see that his release point moved outward in both directions, for the most part, over time.  This would seem to indicate an arm that is getting stretched more and more.  Centrifugal force tells us that the longer the leash the more velocity we should expect, but certainly there must be a counterbalancing biomechanical break point here.  Is Cobb getting stretched out closer to optimum mechanics or is a DL stint the only thing that he’s actually getting closer to?  Again, here’s an area for future study, but for now it’s interesting to see him exhibiting this trend.

Now that we’ve become accustomed to our subterranean lair, what say we go down to another floor where we can look at strategy rather than tactics.  We’ve seen that Cobb has three pitches at his disposal and each has been effective, some more than others.  Let’s look at how he puts it all together:

There’s a lot going on here, but learn it now since the format will be used extensively throughout the rest of this.  The primary axis, on the left, is the percentage of each outcome (Ball, Called Strike, etc…) and the right-hand, or secondary, axis shows the Run Value per 100 pitches (RV/100).  The boxes in the main part of the chart show the raw number of pitches that fit each section.  The boxes at the bottom show the total number of pitches, but also the percentage that each type was used.  As an example let’s break down the change up.  He threw 908 in total with 29% or 259 being counted as balls.  Another 78, or 9%, were registered as called strikes.  The discerning reader might think of these two sections as the pitches that were taken.  So overall, the change up was taken 38% of the time, meaning batters swung 62% of the time.  Moving on, he had 23%, or 206 pitches, fouled off, 231 put in play, and garnered swinging strikes on 15% of all change ups.  Per 100 change ups he allowed almost two runs less than the average pitcher.  That’s really good.

The change is clearly his best pitch, though the deuce isn’t bad either.  His curve is better than league average, though closer to his peers than it is to the change.  Batters are taking  the curve 65% of the time with 25% of all curveballs being taking as a strike.  This has been an effective way for him to steal strikes from time to time and he’s getting another 7% of this pitch to go down as a swinging strike.  He’s throwing the breaking ball 19% of the time so it’s not something he’s stashing away, but he’s not leaning on it as heavily as he does the change, either.  Lastly, the fastball is basically league average at -0.01 RV/100 and he throws it roughly half the time.  It profiles similarly with the breaking ball in regards to percent of pitches taken and also the likelihood of a taken pitch going for either a called strike or a ball.  The difference between the two, at least from a results perspective, is that the fastball gets close to one half the number of swinging strikes with the majority of the difference manifesting as foul balls.  Counterintuitively batters were more likely to put the ball in play on the fastball than the breaking ball, though it’s not a large difference.

The purple regions shown above indicate the percentage of balls put in play.  Sometimes this is a good thing, sometimes it’s not, and the difference usually lies in how the ball was put in play.  Pop ups are great and liners are not, while ground balls and fly balls are all over the middle and have a lot to do with aim and velocity on whether the pitcher is happy with the outcome or not.  Here’s a look at what we’re talking about:

Similar stuff here as above with a couple of tweaks.  The boxes on the right-hand side show the wOBAcon for each ball in play.  This isn’t a commonly seen measure, but we apply the linear weights of single = .90, double = 1.24, triple = 1.56, and homer = 1.95 to combine the best parts of batting average and slugging percentage to get an idea of how well each trajectory or pitch is getting hit.  We total up the base hits and divide by all balls in play, such that, the lower the number the better.  At the bottom the right-hand box shows the overall wOBAcon for that pitch type.

Batters hit the fastball the best, which shouldn’t be a total surprise.  It’s Cobb’s worst pitch and big leaguers feast on fast balls.  The interesting thing is that they don’t do as much damage on the fly balls that were hit as compared to the other pitches even though they hit more of them.  They also hit more pop ups and liners so they don’t have much difficulty elevating the fastball.  Many ground ball pitchers get a lot of their worm-burners by having a fastball with heavy sink, but Cobb gets his grounders with his secondary stuff.  The breaking ball is particularly adept at inducing a ground ball, but the problem comes in with the batters that don’t top it, because they end up hitting holy hell out of the ball when it gets up in the air, whether a fly ball or liner.  Still, he’s getting nearly 3/4ths of his balls in play as grounders so he’s doing something right.  The change up comes in between the two still generating a ton of ground balls of a very weak variety, but the rest of the time batters are able to do some damage if they can get the ball elevated at all.

It’s useful to know these things at the aggregate level, but we can also take a look at various important counts that have a statistically useful number of pitches.  Let’s start off with first pitch:

Same format as the overall here.  Again, we see that his change up is a beast, but he’s only using it 13% of the time to start a batter off.  This follows the old adage that you don’t want to throw a change first pitch because what, exactly, are you changing up from?  Still, it’s nice to see him not completely eschew his best pitch.  He’s not throwing it under the bus, either.  The reason to throw first pitch change is to destroy the souls of first pitch fastball hitters.  He seems to be doing just that with his 13% swinging strike rate on top of another 28% of called strikes.  A sort of flag would be the 18% balls in play, but I’d hesitate to label it of the red variety since his RV/100 is so strong it’s likely that he’s getting a lot of weak contact turning into easy outs.

He’s using his curve ball much more than normal throwing it around 30% of the time first pitch.  This is a central tenet to the idea of pitching backwards.  What batter other than Manny Ramirez is going up to the plate saying to himself, “C’mon Meat we’re gonna sit on that curve ball and then crush it out of the yard.”?  Nobody.  That is the answer to the question.  Nobody goes up there looking for a curveball and that plays out with the fact that 81% of his curves were taken with a nearly even split between those called a strike and those called a ball.  I love this.  He’s also throwing the fastball quite a bit to keep batters honest.  He’s seeing a similar take rate as the curve with 78% not inducing a swing.  Uncannily, he threw the exact same number of called balls and strikes at 179 apiece.

First pitch might feature the most number of instances since every batter gets one, but 1-1 might be even more important.  In 2012 American League batters had a slash line of .346/.347/.575 in a 2-1 count.  In a 1-2 count they only hit .165/.172/.245.  To this analyst there is not a more important pitch than 1-1.  It sets up the entire rest of the at bat.  Get behind and the pitcher must burn more fuel than he’d like to battle back.  Take the lead and he can control the situation cruising through without all the stress.  Here’s what Alex Cobb did on 1-1:

On this all important pitch we see that Cobb favors his change.  This is a man that grasps the importance of the situation.  When your back is to the wall you don’t pull out a knife you reach for your revolver.  Not only does he throw the change the most, but he does it effectively getting 16% swinging strikes and saving 1.6 runs per 100 pitches compared to the average.  Batters are amped to swing taking only 34% of these pitches, but he steals his fair share by drawing a foul ball another 22% of the time. He threw the curve nearly a fifth of the time, which only amounts to 59 pitches, but it was basically a league average pitch that wasn’t swung on all that often and put in play even less.  He used the fastball less than the change, but only by around 17 pitches with batters taking around half and fouling off another fifth.  It’s likely that the fastball was being hammered, though, as his RV/100 was over 1.0.  The difference between a good fastball and an ok one is that the former can still get by batters when they’re expecting it while the latter can’t always say the same.

Next we can take a look at what he throws when he does proceed past 1-1 into the next gauntlet:

First the bad news.  When Cobb gets to 2-1 he’s basically a one pitch pitcher.  He can’t throw the curve ball because if it’s taken for a ball then he just went from the frying pan to the fire and he can’t throw the fastball because it gets shellacked at roughly 32% balls in play and another 28% taken for a ball and an RV/100 over 4.0.  This means that batters can keyhole the change.  If Cobb throws it for a ball, which he did 31% of the time, then he’s in an even worse situation, but his RV/100 of 0.76 shows that even a very good pitch can be hit when the element of surprise is removed.  Let’s look at the other side of the coin:

With the tables turned Cobb doesn’t really go to his curve all that much more which is kind of surprising, but with that RV/100, albeit over 34 pitches, there’s probably a reason he avoids it.  His fastball isn’t a whole lot better and he’s throwing it almost half the time, mostly for a ball, though he also got 12 frozen strikeouts unlike with the other pitches.  This is another facet of reverse pitching.  You want to be able to throw your secondary stuff for strikes and be able to nibble with your fastball.  If you’re going to miss you’d rather it’s off the plate for a ball, but if you can steal one then even better.  The change really shines here, though.  He got a swinging strike 18% of the time and only threw a ball 30% of the time.  All of this led to an RV/100 of -3.81.  Striking guys out and inducing weak contact is a great combo.

The plan was to take you down one more floor, but after compiling the following table it seemed a lot less necessary:

The idea was to look at how he pitches differently to lefties and righties, but he really doesn’t.  His pitch mix is essentially the same.  Whereas many pitchers will only throw their change to opposite-handed hitters, Cobb has the competence and confidence to throw it to both.  This means that he doesn’t need to appreciably change his pitch mix based solely on where the opposition stands relative to the plate.  The interesting thing, however, is that there is a bit of a platoon opportunity here for the opponent and it’s not news to them.  Cobb has thrown nearly 1,700 pitches to lefties and only around 1,250 to righties.  Managers seem to be stacking their lineup a bit with lefties in order to take advantage of Cobb’s fastball.  That probably makes sense when you see how much righties have struggled with Cobb, but by stacking with lefties they make his change up even better.  Seeing these numbers help me gain a lot of respect for Cobb’s fastball.  It’s not going to blow up a gun, but it’s his best pitch against righties and it forces opposition management to pick their poison.

Alex Cobb has basically one season under his belt.  He’s still under the radar for most, though his name is starting to pop up on a lot of sleeper lists for fantasy and even more folks will be trumpeting his outstanding spring training stats, which admittedly is better than trying to justify a poor performance, but doesn’t mean a whole lot either way.  If you started off thinking Cobb had one good pitch and wasn’t much more than a Sonnanstine-clone then the hope is that you’ve gained some newfound respect.  The rest of us will continue to smirk and smile at yet another product of the Rays pitching machine that can take guys like James Shields and Alex Cobb and turn them into ground-balling maestros. Not saying it will go seamlessly.  Tight zones will give Cobb more trouble than most, but with this infield behind him and Jose Molina stealing strikes Alex Cobb is in for a very big year.  He’s not likely to fill all 220+ innings that Shields would have likely thrown, but it’s likely he hits 200 or more and he has every chance in the world to be just as good as Shields using a similar mix and an outstanding change.


About Jason Hanselman

Rays fan.
This entry was posted in preview and tagged , , . Bookmark the permalink.

2 Responses to Meet the New Boss, Same as the Old Boss

  1. Awesome post, and I totally agree that Cobb has been operating under the national radar.

  2. Pingback: Reviewing the Starters in 2013 at the Team Level |

Leave a Reply

Please log in using one of these methods to post your comment: Logo

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

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s