T.J. McFarland: A Rich Man's Casey Fossum?

T.J. McFarland was a Rule 5 pick by the Baltimore Orioles from the Cleveland Indians in the off-season. The left-hander stuck with the club as a reliever through spring training and has made two appearances so far in the regular season. I didn't know much about him when he was in AAA last year, and didn't see him in Spring Training, so I wanted to get a look at his Pitch F/X data (we are only looking at his MLB outings, and not the Arizona Fall League outing from 2011) and see if he is a guy worth carrying on the roster all season.

First, it is always helpful to look at where the pitcher is releasing the ball. It seems like a good start anyway. Here is McFarland's release point in graph form:

Because he doesn't come out very much but is still releasing the ball lower than average, McFarland is sort of unique when it comes to release point. The closest comparison may be a slightly more in and more upright Casey Fossum. While it is not always predictive of control, but he has repeated his delivery and arm slot pretty well with a .104 horizontal standard deviation and a .11 vertical standard deviation.

As far as the fastball goes, there really isn't much of one for McFarland. He has maxed out just over 89 MPH. The MLBAM tags also confused his changeup and fastball some and had some real weird classifications overall. I count 4 different pitches when looking at his spin and speed graph (with some labels that will prove to be a little hasty):

His average 4-seam fastball (the 42 pitches above 85 MPH) is 87.57 MPH. I think here we see the reason he was left unprotected by the Cleveland Indians. He is averaging less than 88 MPH out of the bullpen. This would rank 180th out of 189 left-handed bullpen pitchers in the Pitch F/X era. This is where the Casey Fossum comparison makes sense again, as he averaged 87.2 MPH on his fastball according to the traditional MLBAM tags and distance (and he averaged 86.1 MPH in Japan in 2010).

The pitches I am calling the moving fastballs in the graph average 81.1 MPH. Obviously this is significantly slower than his 4-seam fastball. Based on vertical movement (1.73 inches on average), it doesn't move much for either a cutter or sinker, or really any pitch. On the other hand, the horizontal movement is really dramatic. The Orioles park does not appear to be one of the parks that dramatically change horizontal movement, so this appears to be very real, and makes placing an actual classification on the pitch a little easier, as it clearly rules out a cutter, and puts him with a lot of sidearmers or lefty specialists. So while his arm may not come out as much as a "sidearm" type pitcher's might, his pitches can move somewhat like those of a sidearmer, and he is clearly not throwing very hard.

If we look at the pitches I am calling a changeup, the pitches that spin heavily, they average 78.7 MPH. This is a slow changeup, but frankly, it is slow for really any pitch except a curveball or special pitch. It's movement actually makes more sense as a breaking pitch than an off-speed pitch, but we will see the problem of classifications with his next pitch.

The average slider, what I am calling a slider with the speed and spin graph, averaged 79.8 MPH, with .857 horizontal movement and -2.091 vertical movement. This negative movement suggests that the pitch is actually a curveball (this is what the manual tags at Brooks Baseball call it). However, no other left-handed bullpen curveball in 2012 got that little vertical movement. Only 8 lefty bullpen curveballs in the Pitch F/X era have that little vertical movement, and they are all slower than McFarland's. These eight seem to be decent whiff curves, but horrible groundball curves. When you look at pitches that are still classified as sliders despite the "negative" movement, we see that Nick Hagadone provides a somewhat similar comparison based on movement. Hagadone's "slider" is thrown nearly 4 MPH harder than McFarland's, but Hagadone's average fastball is also about 5 or 6 MPH harder than McFarland's as well. However, we see that Hagadone's "slider" (looking at the sliders in the outing he has thrown the most pitches in for his MLB career, just 34 pitches) has some pretty inconsistent spin

McFarland's spin is most similar to a traditional curveball. Whatever McFarland is actually throwing is a strange breaking pitch without a ton of movement (it is hard to judge velocity because it is hard curve yet a soft slider). It is also his least frequent pitch so far (according to my classifications, as the MLB AM classifications say he throws it significantly more, I am basing my classifications on the spin and speed graph) Since we are considering the "moving fastball" above as a considerably separate pitch, this pitch is not all that important, because he is a reliever and it is his 4th pitch. It is also where the Casey Fossum comparisons end or take a significant turn, as Fossum relied heavily on a slider that looked different than McFarland's.

Here is McFarland's strike zone so far:

This is a small sample size, so we have to be careful, but it seems he would like to work low and inside/down the middle. I think this supports the idea that he isn't throwing a lot of sliders (though he could just be missing, which would be hard to prove without looking at actual video and catcher targets, especially since he has a repeatable delivery and arm slot), as we would expect them to sweep away from lefties. Instead, he is working with a lot of armside low fastballs, something that makes sense for a soft tossing, but not real sidearming, lefty.

So far, the Orioles have not used him as a LOOGY, instead, using him sparingly as a long reliever. How to actually classify him and his pitches is sort of difficult, as we are seeing that he is a pretty unique pitcher. One could change the labels I used and make an argument that he is a fastball, changeup (instead of moving fastballs), slider, and curveball pitcher (or two different breaking balls). This is a decent assortment of pitches, especially out of the bullpen, but other than the horizontal movement, none of these are really plus. There just isn't a lot of upside here, especially since, if you wanted him to be a starter, his fastball would lose even more velocity. So if the Orioles are going to continue to carry him, it shouldn't be because they think there is some future value he can provide. It has to be because he is a decent control lefty out of then bullpen. He may have some fringe value, but I wouldn't do backflips to keep him on the roster, as he is not a long term solution in either the bullpen or the rotation for a MLB team, especially in the AL East.

Colby Rasmus and Velocity

Ben Nicholson-Smith had an excellent piece that showed that the Blue Jays, specifically Colby Rasmus, have been facing top notch fastball velocity in the early going of the MLB season. In the article, Rasmus claims he hits faster pitches, specifically fastballs, better. I have always been interested in Rasmus from a distance (I haven't been as big of a fan of him as many, especially the Blue Jays evidently), and wanted to test the claim Rasmus made and look at some of his other tendencies as a hitter to see if we could figure out the puzzle that often is Colby Rasmus.

So using Pitch F/X (specifically Brooks Baseball's Tabular Data by games) I looked at all the pitches Rasmus has seen against starters in 2012 and early in 2013. If for some reason someone wants to look at those pitches and make their own observations or charts, you should be able to download my spreadsheet by clicking here.

Out of those 1495 pitches, Rasmus has seen 80 pitches over 95 MPH. He swung and missed at 9 of those (11.25%). He has put 15 of them in play, 12 of them for outs, and 3 of them for no outs. BABIP plays a role obviously, especially in a sample size this small, but it doesn't seem that he is having a lot of success on top notch fastballs (not that we would expect anyone to have a lot of success with 95-99.5 MPH fastballs).

He has seen 581 pitches at 91 MPH (about MLB starter average) or harder (regardless of classification). I count 40 whiffs in the 501 pitches between 91 MPH to 95 MPH (7.98 %).

I count 122 four-seam fastballs thrown below 91 MPH and count 7 whiffs out of that group (5.7 %). So it would seem that Rasmus gets better the slower the fastball is, which again, is what we would expect normally, but shows that Rasmus' confidence in squaring up 95 MPH fastballs is probably overstated (of course, no hitter is going to say "I just give up whenever I see 95 MPH fastballs because I don't hit them as well").

Of course, much of the context of Rasmus' confidence in hitting harder pitches was versus breaking pitches. When you again sort pitches into bins just by velocity and ignore classifications:

85-90.99 MPH: 466 pitches in total, 52 whiffs (11.1 %), 14 run scoring plays (3%)

80-84.99 MPH: 257 pitches, 44 whiffs (17.1 %), 2 run scoring plays (.7 %)

75-79.99 MPH: 151 pitches, 20 whiffs (13.2 %), 5 run scoring plays (3.3 %)

Slower than 75 MPH: 39 pitches, 1 whiff (2.6 %), 0 run scoring plays

So we see that the 80-85 MPH range is really his trouble spot. This is where many of the sliders and changes are going to be. However, when you slow it down a little more, where most curves are going to be, he has more success. Overall, his whiffs came on an average velocity of 86.67 MPH, while his run scoring plays came on an average velocity of 87.84 MPH. So it seems that, while he is better against softer fastballs than harder fastballs, he is somewhat right about liking harder pitches. However, if you look at the balls he has put into play, the ones that have turned into outs average 88.5 MPH, but the ones that were not turned into outs were 87.28 MPH. So the relationship is definitely complicated.

Because it may just be breaking pitches versus fastballs, without having to deal with classifications, spin may give us a little bit more of a look. High spin pitches are usually going to be fastballs and changeups, while low spin pitches are usually curveballs and sliders:

Whiffs: 184.63 degrees
Runs Scoring plays: 215.844
Outs: 199.35
No Outs: 184.88

Again, we see the somewhat inverse relationship that may just be affected by BABIP.

I find it interesting to look at the release points of opposing pitchers for hitters to see if we can find what hitters end up having problems seeing. So here are the release points by opposing pitchers on Rasmus' run scoring plays:

Here are the release points on Rasmus' whiffs

 

It is tough for me to see much of a difference (other than the obvious one, he is struggling against more sidearm type left-handers, which we would expect all lefties to do), especially since the sample sizes are different sizes. With that said, I find it fascinating how much he is missing on right-handers that have a very "out" or 3rd base side of the rubber release point. Usually, you would expect him to see the ball better from those pitchers.

It still may help us to look at his strike zone:

 

As we would expect from looking at his velocity data, he is really good on pitches up and in. The problem doesn't really seem to be bat speed or any kind of hitch in his swing (or the timing that comes with his leg kick). Even though he predictably struggles with harder fastballs, he gets around on average ones. The problems come on balls low, whether away, inside, or down the middle (or just middle to outside in the strike zone). He doesn't respond to hard/medium breaking balls very well. This may help us see why he is missing so much from "out" right-handers, as they may just be feeding him braking balls outside, which he can't really hit anyway, even if he sees it.

When we break down his stance and his swing (screenshots below), we see nothing that tells us he should have such problem with low breaking balls. He doesn't have a weird open stance:

He keeps it straight as the pitch is delivered, and even leans in to give him some reach:

And he has good plate coverage:

There doesn't seem to be anything obviously mechanically wrong with his swing. As hinted to by himself, he may just have problems recognizing pitches.

Thad Weber's Curveball

Thad Weber made his Padres' debut on Saturday, when he had to eat up innings after a poor Edinson Volquez start. Weber was optioned after the game.

While I was watching him, I was really impressed with the vertical depth he got on his curveball. It was also strange to me how hard he threw it considering his fastball velocity, which was usually around 90-91 MPH.

He really struggled to locate the curveball though, and several of them were hit because they were left up in the zone:

This may be because he had some problems repeating his delivery:

As you can see, this is also quite a bit far out horizontally as well, making it very easy for lefties to see.

Weber threw his curveball 20 times out of 65 pitches in his game against the Rockies, which is 30.8 % of the time. Since the Pitch F/X era began, there have been just 6 qualified relief pitchers that have thrown a curve ball that much. Mike Lincoln, Jason Bulger, Taylor Buchholz, and Jose Veras are the lefties.

If we don't look at the spring training outings because a) we don't have park factors for those parks and the movement data is kinda weird there sometimes and b) because I have heard pitchers complain that their curveballs don't always break right there, and we add the park effects of the each curveball thrown in MLB parks, we get a vertical movement of -9.655 and horizontal movement of 2.09. Not surprisingly, this vertical movement is elite. However, the horizontal movement is no where near the horizontal movement of the other elite vertical right-handed curves:

It is clearly a straight drop down curve, not moving or sweeping much horizontally. However, it is a hard curveball as I mentioned above.

Like I did with the movement data, I also added the velocity to Weber's pitches for his Padres' outing that Petco Park seems to take away according to the park effects (the Tigers' park evidently doesn't have an extreme effect on right-handed pitcher velocity). This changes his fastball velocity to 90.78 MPH and his curveball to 81.54 MPH. Only 11 right-handed relief pitchers in the Pitch F/X era (out of the 45 that have thrown at least 500) have had a faster curve than that, and they have an average whiff/swing percentage on their curves of 39.05 % (the median is about 33 %). These pitchers all, without exception, throw much harder than Weber, and averaged 94.51 MPH on their fastballs. This makes Weber's curve even weirder. It is a curveball, both the visual test and the Pitch F/X movement data assert this, so this isn't a classification issue.

This velocity data, along with the vertical movement we saw above, makes Weber's an elite curveball. He knows it is his best pitch, at least that is what his usage indicates, but how rare is his combination of velocity and movement, I think this stacked chart of the 46 right-handed curves (putting Weber with the 45) gives us a view (I changed the vertical movement, which is the red color at the top, to positive just so it would stack).

The only right-handed relief curveball that has more velocity plus vertical movement is Craig Kimbrel, who is the best reliever in baseball. Clearly, Weber doesn't have the fastball Kimbrel has, and will never be in the same league as Kimbrel. He also is already 28, so not a prospect by any means. We don't know how his command will be, and the release point data suggests not good, and he probably won't be very good against lefties, but we have seen late developing strange one pitch relievers have success before in the Majors, and Weber's curve may just be good enough to get MLB hitters out and give him a MLB career.

Cody Beam, Joseph Shaw, and Other DBU Notes

Cody Beam is a junior right-handed pitcher for Dallas Baptist University. Here is some video of his delivery:

The fastball looked good, comes in hard, and he seemed to work it to all sides of the plate. His off-speed pitch is soft and wasn't impressive in movement. This pitch was hit hard. He also had a tendency to let go of the ball too soon and have a lot of balls stay way up. This was particularly a problem with the breaking ball. He occasionally got some decent tow way tilt low and way to righties with it, and the usage suggests a soft slider. His best pitch was clearly the high heat when it was close enough to the zone. Here is a video of Beam pitching, his pickoff move, and Dustin McAlpine throwing to 2nd trying to get a runner (though I accidentally ended the video too early)

Late in his short outing, he was hit on the hand by a hard groundball and seemed like he couldn't get the breaking ball over after that (even though the fastball command stayed okay).

Joseph Shaw is a freshman right-handed pitcher that was selected in the 40th round by the Astros in 2012 but obviously did not sign. He is a pretty big human being (listed at 6-5), but doesn't have a big stride in his delivery:

He obviously gets on top of the ball well, and he threw a lot of low fastballs. He wasn't as overpowering as I thought he would be. He located his breaking ball well but didn't have much plus movement or velocity.

Speaking of Dustin McAlpine, I got four of his pop times to 2nd base in warmups: 2.22, 2.16, 2.19, 2.22. These are not good times, but it might be important to note that they were thrown with a lot of effort or urgency. Most of the throws were strong, but opponents ran on him quite a bit, and his in game throws were not as strong (I have been fascinated by McAlpine's inconsistency behind the plate over the last couple of seasons, and I may be just spitballing here, but I did notice that he takes warmup tosses normally, and then, when the batter comes up to the plate, he puts a sleeve on his arm that has a sheet with what is probably a list of pitches, that has a part that goes over his thumb. I wonder if this is affecting his throwing. Probably not, but I found it interesting he threw his warmup thrown down without it on). He did get a better time of 2.03 in the game, and picked off a runner at first. I didn't get that time, but when he tried to pickoff a runner at first later in the game, this time unsuccessfully, I got a time of 1.65. At the plate, he hit a 6.37 second fly-ball to centerfield.

Justin Wall hit a 4.25 second fly-ball to center. Nash Knight hit a 5.29 fly-ball to right, a 6.03 fly-ball to shallow right, and a 5.25 fly-ball to left. Ronnie Mitchell ran a 4.37 time to first from the left side. Shane Lynch hit a ball off the center field wall for a triple, and then later hit a home run at 5.44 seconds to center.

A Jair Jurrjens Update

Jair Jurrjens had a two year run, in 2008-2009, where he was a very good starter in the big leagues (worth over three wins above a replacement pitcher according to FanGraphs' WAR). However, things unravelled in 2012 and he threw more AAA innings than MLB innings. In 2013, Jurrjens signed a deal with the Orioles, which was changed to a minor league deal when it was revealed he had a knee injury (he then suffered another injury in spring training). Jurrjens just made his 2nd AAA start of the season, both against the Durham Bulls.

In his first start, Jurrjens threw 56 strikes out of 95 pitches, with just as many groundouts as flyouts (just like his second start). He had 7 strikeouts, 2 walks, only 1 hit allowed, a single to Wil Myers on a groundball.

Since he has now started against the Bulls' twice, here are the Steamer Projections (if they played in the Majors) of the batters he faced (2 in parenthesis means he faced them in both games):

Rich Thompson (2): .657 OPS

Hak-Ju Lee (2): .625

Jason Bourgeois (2): .677

Wil Myers (2): .754

Leslie Anderson (2): .684

Chris Gimenez (2): .659

Vince Belnome (2): .695

Cole Figueroa (2): .682

Tim Beckham: .627

Mike Fontenot: .686

These batters have an average OPS of .677, so Jurrgens should be dominating them. Unfortunately, he hasn't. He has struck out just 9 of the 45 hitters he has faced and walked 6 of them with a below average groundball rate.

First, before we look at his 2nd outing in detail,let's take a look at his delivery:

For comparison, here is Jurrjens in 2012:

2009 was when Jurrjens was really dominating (at least according to some of the more basic metrics). When you look at his 2009 highlights, two straight starts have highlights called "Jurrjens' Strong Start", and then the next one is named "Jurrjens Impressive start". Here is what he looked like then:

It is hard to see much of a difference if any in his final release point. His right leg is at the same angle, the back is twisted in the same way, and his head is pretty similar as well. I wanted to look at Pitch F/X data, to see if their was a change in his release, so I picked his start with the most pitches from 2009 versus his outing with the most pitches in 2012.

In graph form, here is Jurrjens in 2009:

Here is Jurrjens in 2012:

Jurrjens has clearly never been extremely consistent in his delivery. With that said it also looks like it was better in 2012 than in 2009. The numbers bear this out as well. In his 2012 start, his horizontal and vertical standard deviations were .135 and .081 respectably, solid numbers actually. In his 2009 start, it was .145 and .113 respectably, which isn't a bad number, but not as good as his 2012 start. Other than slight drop in 2012, his release point is basically the same, meaning that the release point doesn't seem to be why Jurrjens had so many problems in 2012 and regressed so badly.

So finally let's look at the pitch data that I collected from his AAA outing:


The first thing you notice is that his fastball averaged just 88.55 MPH. Out of the 215 right-handed starters that have thrown at least 500 pitches as a starter in the Majors, only 27 of them have fastballs under 89 MPH on average. It is not exactly a real successful group. His fastball acts as a sinker most of the time, and he threw it 52.7 percent of the time.

He threw a curveball 18 times in the game that didn't get a lot of vertical tilt at an average 74.67 percent of the time, but he relied on his off-speed pitch, the changeup, more often (25 times), as the curveball made its appearance late in the game, in the 2nd and 3rd time through the order. The change moved a lot like his fastball, and was thrown at an average of 81.96 MPH, again well below league average.

Jurrjens got 7 groundballs and 11 flyballs, again weird because he was a decent groundball pitcher in the big leagues, especially when he was going well. He also got just 6 whiffs (6.45 % of the time), and while missing bats was never his speciality, we see that he isn't missing bats, getting grounders, or even throwing a lot of strikes.

He clearly knows that he has below average stuff now, and is focusing on locating the ball low in the zone:

 While we are always taught that pitchers should "keep the ball low", we know that pitchers in the Majors that throw low in the zone constantly are not very good pitchers.

At this point, Jurrjens looks like a AAA pitcher. The stuff has really regressed, and he isn't throwing a lot of strikes. Without an uptick in velocity on the pitches, I can't really see him being a successful MLB starter again.

C.J. Wilson's Release Point

C.J. Wilson is probably the most successful reliever to starter conversion in at least recent memory. His move to the rotation helped the Rangers' reach two World Series and helped him get a 77.5 million dollar deal over 5 years with the Angels. So far, Wilson has not been the same pitcher in Anaheim, with a -1.3 WAA over 214.1 innings (he had a 5.2 WAA in 427.1 innings in his two years as a starter with the Rangers). When Wilson has had problems, it is usually because of walks. When he struggled against the A's on his most recent start (April 9th), he threw over 40 pitches in the first inning. Since becoming a starter, Wilson has thrown 3.97 pitches per plate appearance. This can be a good thing, in fact, the elite pitchers throw more pitches per plate appearance than pitchers that don't (strikeouts), so I wondered if we could use Pitch F/X data to see if we could reasonably conclude whether or not the high walk totals are a choice by trying to pitch around batters (and not being successful from a lack of stuff) or whether he just legitimately is having problems repeating his delivery.

If you look at all of Wilson's fastballs combined (cutters, 4-seamers, etc. all added together, skirting around most pitch classification problems), he is averaging 91.06 MPH so far in 2013 with 1.51 horizontal movement. In 2012, they averaged 91.81 MPH with a 1.59 horizontal movement. In 2011, his fastballs averaged 1.6 horizontal movement and 91.17 MPH. If you are looking for a dramatic loss of stuff for Wilson, then you need to look at something other than fastballs. We could break down the breaking pitches pretty easily, but instead, let's look at Wilson's release point to see if this gives us any insight into his command.

When looking at his release point data for his career, I noticed that he has come more "out", going from having his arm when it releases the ball about 1 feet from the center of the rubber/plate to about 2 feet in 2013. This started happening in August of 2012 so I decided to look at pictures (screenshots I took of highlights) and release point Pitch F/X data.

Here is C.J. Wilson in the World Series in 2011:

Here is C.J. in August of 2012:

The shots are at a little bit different times in the delivery and the 2012 photo is not zoomed in as much, but (especially if you zoom in) it seems that he has moved on the rubber closer to the 1st base side. Other than that, the body angle, posture, stride, and arm angle look the same to me. So it seems that he has simply moved on the rubber.

Let's look at the numbers to see if this changed how consistently he repeated his delivery.

2011: Horizontal SD: .214. Vertical SD: .192

2012 before he changed his stance on the rubber:
Horizontal SD: .211. Vertical SD: .1711

After he changed: Horizontal SD: .15578. Vertical SD: .1686

We wouldn't expect the vertical to change much, but it seems he has gotten more consistent at how high he is releasing the ball. He also improved horizontally, going from a horrible number to one that is a little more manageable.

In graph form, here is 2011:

Here is 2012 on a whole:

Here is C.J. after the change in 2012:

Obviously the further out part of his overall 2012 graph was simply his move on the rubber, which makes his year on a whole look worse. So the sample size was smaller, so we should be careful, but Wilson did repeat his delivery better after the change. Let's look at 2013 so far.

In picture form:

In Number form:

Horizontal: .207. Vertical: .1688

He is basically went back to his old horizontal inconsistencies, while his vertical release has remained better than it was. That is sort of a weird find, as we would expect the horizontal to be the one that is affected the most.

In Graph Form, here is the zoomed in version:

The larger view (like the other graphs above):

From a numbers perspective Wilson (small sample size alerts apply for early 2013 obviously) has not gotten better. His stuff seems to be more or less the same from 2010-2011, and he seems to be repeating his delivery at least a little better. So we are left with three choices in my view, 1) This data suggests he should get a little better, 2) He may be a little more consistent at repeating his delivery, but the move on the rubber is a negative thing and cancelling out any benefit or 3) It is some other variable that we are not considering. He is pitching in the same division, against similar opponents, and while yes he is older, the stuff seems to be there. I could be missing something, but I can't quite figure this one out.

Sortable Reliever Release Points

In a previous post, I collected the release point averages for every starter in the Pitch F/X era and created a chart. In this post, I will look at the relievers of the Pitch F/X era in the same way. I used the same methodology I did with the starters.

The best view of the chart is actually by following the link and seeing it full screen (and you should be able to download it). However,this isn't very convenient, so here is the chart and the spreadsheet below:


If you just want to see or save the picture: