Clay Buchholz PFX

I have a new PFX article up at RotoSavants. I hope you check it out. I think that the switch in arm slot is his main problem, not any injury.

Billy Wagner breakdown

With the news coming out of the Mets organization that Wagner is experiencing elbow pain, fans are understandably worried about their championship chances. Most fans are worried that he will need Tommy John surgery, ending his chances of coming back this year and almost surely next year as well. Looking at the PFX data, I think the Mets fans and those in the organization might not want to hold their breath. Nothing is 100% conclusive but I am worried (he's on my fantasy team).

Injuries to the elbow are commonly said to effect control more than velocity so let's take a look at that first. Initially looking at the chart of % of pitches that were thrown, there doesn't seem to be any sort of general trend. I think this is somewhat deceiving though. The pitches that were recorded as balls were either 1) so far out of the zone that the batter was never going to swing in the first place or 2) the batter guessed wrong but the pitcher didn't hit his spot. Another result of the pitcher not hitting his spots (and therefore not having as good command) is the balls that were put in play.


Some pitchers do pitch to contact, however non of them want the ball to be hit hard. They either want weak pop ups or weak ground balls. That's where the 3rd chart come in. In the 3rd one the cumulative percentage of pitches that were balls or hit into play trended towards increasing throughout the year. There are two trend lines. The first flat one is a straight shot linear trend line. It's crude and ineffective for huge variences in numbers like these but it helps to show the general trend. The other trend line is a 15 period moving average. Common in financial analysis, it helps to show more of the trend over time while taking into account each individual case rather than the group as a whole. Clearly both of these lines trend more towards increasing amounts of pitches that are balls or are hit in play.

Is that part of Wagner's plan? Is he pitching to contact and letting his defense take care of things? The answer is no (sort-of). He started off well, giving up no hits but he has given up more hits over the last few months and that worries me. He does have period where he doesn't give up hits but he's given up hits more often. This data, when combined with the other chart, gives me the best indication that something is wrong. He is consistently missing his spots and he is getting hit hard when he misses those spots.



Is his speed any different? The answer is yes, but not in the manner that you would think with an arm injury. His average fastball speed has actually trended upwards throughout the year. This could be a compensation for the elbow/forearm woes. Commonly when a pitcher has elbow/wrist pain, they end up throwing the pitch harder in an attempt to get more movement on it. Most of the time this is not the proper way as movement is commonly tied to hand/wrist/forearm motion. Therefore you could see a general trend upwards in end speed with an elbow injury.

Which leads me to vertical movement, one of the key things scouts look for. Here is movement is also depressed relative to where it was earlier in the year. Each of his pitches are losing vertical movement over time.



Last but not least, his release point has consistently changed throughout the season. It varies widely throughout the range between a little over 4 feet to close to 5.5 feet.



So after a long post, what can we see. We see a pitcher who is losing his command and control while also losing movement throughout the year. He's throwing his fastball harder, possibly in an attempt to compensate for the drop-off in movement. Finally we see a pitcher who's release point is only consistent in the fact that it is inconsistent. To me that spells trouble, especially with the news that his elbow is now hurting.

What's going on with Brad Penny using PFX?

Well we know what's going on with him as he was put on the DL with shoulder inflammation again. But what effect was that having on his pitching? Was the inflammation truly causing him to lose control? Was it making him more hittable? Looking at the data, if this is a true reflection of how shoulder inflammation manifests itself, we might have to look at certain statistics differently. I took a look at all of the pitches thrown within the last 90 days.

Penny has been on the DL twice now this season alone for inflammation in his shoulder. When he returned from the DL last time, his end speed just was not at the same level as it was before he went on the DL the last time averaging about 3-5 MPH less than what it was before the DL the first time. It's important to compare pitch speeds within each group and the drop-off is pretty dramatic. To me this says that he clearly wasn't fully healthy when he returned even if he reported no pain.

Release points might be able to help show if there were any problems recently. First though, as a group it appears that he clearly releases the fastball from a different slot than he does the other pitches. Scouts and GMs everywhere want the same arm action for every pitch so that


The charts clearly show how those are more skewed towards the right (towards 0) on the x-axis. What I found interesting though was that there wasn't as much of a vertical difference between the starts as I expected. With that being said, a range of a foot in terms of release point is quite a large distance. Ideally it should be equal across everything the board, but realistically you want to get the difference closer to a range of 2 baseball widths or 6". The same goes for the horizontal axis and here you can really notice a large difference. This horizontal measurement is much more effected by factors such as where the pitcher is standing on the rubber which can lead to a difference of several inches as well. In these two charts, he appears to be in these parameters. He definitely tends to release his fastball differently from his other pitches.

But we're mainly focusing on the vertical release point. The last chart is very telling. As the game goes on his vertical release point lowers dramatically especially after the 5th inning. While I expect that a decrease in release point height is normal for all pitchers, seeing that dramatic a change is telling. To me the inflammation is due mostly caused by fatigue and not necessarily a new mechanical disruption of the muscle fibers (strains). This can also go hand in hand with the finding that his average end speed is lower now than before. Note: The problem is that if he had a minor tear in his shoulder prior to this year, the difference would not be seen as much within this year.

Looking at his movement next, I don't see anything that really sticks out to me initially.

Looking at the second chart though we start to see a general trend showing fatigue setting in fairly early. His pitches start to lose movement as early as the 2nd inning. Usually if a pitcher is healthy fatigue doesn't set in this early, which makes me worried and curious. There are many injuries which can cause early fatigue including over-training (overworking), labrum pathology, rotator cuff pathology, or general laxity in the shoulder.

The next two charts show the result of the pitches:

These results don't show anything that really jump out to me. The % of pitches that are thrown for balls, strikes and those that are hit in play stay roughly constant throughout the times, except for immediately. It seems though that the rates stay roughly constant except for 3 main events. Balls in play with outs recorded are very far down in each of the starts leading up to being put on the DL. Also the amount of pitches that were balls increased while the pitches for strikes decreased. This is what I expected.

I think that while we are too far away from being able to diagnose what is wrong, but I think that we will be able to tell if something is wrong. From these I think we can be fairly certain that his shoulder never recovered during his initial DL stay. Until we can get these measurements back to normal, I would worry about a recurrence.

Would the real Huston Street please stand up.

Over the last couple of months, Huston Street has completely disappeared. He's been extremely inconsistent and many times horrendous. Since June 20th, he's been in 17 games with an ERA of 7.11 giving up 15 ER, 12BB, 14 K, and 26 Hits in only 19 IP. He has blown the lead (either blowing the save or losing the lead) 6 times since that time.

In the previous 29 appearances, he only gave up the lead twice, with giving up 11ER and 8BB, 35K, in 31.1 IP. He gave up 22Hits in these 29 appearances.

K/9 of 10.12 and a BB/9 of 2.32 before
k/9 of 6.63 and a BB/9 of 5.68 after

That's quite a difference in control.

What's causing it though? I personally believe he's injured or in a "pre-injury" period where the body isn't functioning properly but there aren't any symptoms. Looking at the PFX data only confirms what's I'm fearing.

To me I want to show the most damning pieces of evidence first. I want to look at the vertical movement over his season to see if there has been any drop off.






You can clearly see in all 3 graphs there is a general downward trend in the vertical movement of his pitches, especially his change-up and slider. This is a very bad thing. The pitches are basically coming in on a straight plane and there is a much less of an adjustment that the hitter needs to make. Eventually they'll start coming in like a batting practice pitch and of course it's going to leave like it too.



His pitches are also slowing down but not nearly as dramatically as the downward trend of the vertical movement. His fastball pitch is slowly decreasing which is not a good sign.

Next, I want to look at how many pitches he's throwing in each outing.

Ever since the middle of the season his pitch counts are basically trending upwards. He's less efficient, getting hit harder, and not being as effective as he should be. Going along with this you can see the general upward trend of the percentage of pitches that were thrown for balls. This means he's losing his control of his pitches, which is often a sign of injury.

Throw all of these together and to me it's troubling. I think there is something going on. It may be fatigue, but it may also not be fatigue. It could be an injury that we don't know something about. It could be a trunk injury, it could be a shoulder injury, or it may be an elbow injury.

Examining what separates great pitchers from above average or average ones using Pitch FX

So this was planned for yesterday afternoon, but of course my computer started having problems so I couldn't upload it.

The question comes up often, at least in my mind, what makes an elite pitcher so good. Why is it that year in and year out, he can consistently be that much better than the other pitchers in the league? So of course, that’s what I set out to prove and it’s why I haven’t had a full update in a couple of days.

To start off, I think I need to explain how I broke down the pitchers into groups. I chose 3 groups (Elite, Above Average, and Average) and focused only on the #1 pitcher on each team according to CBS Sportsline. From here on out when I say an elite, above average, or average anything, it is not to be compared pitch to pitch. Those descriptions are for the groups that they belong to. I chose only right handed pitchers in order to simplify the graphs and also because some people have noted that many left handed pitchers have different mechanics than right handed pitchers. Thus, in order to make this as close to a uniform sample as possible, I chose only the right handed pitchers. I chose the groups based on the research by Derek Jacques using ERA, IP, W, L, BB, K as a basis for what category that pitcher is in.

Pitch Type	Category	End speed
CH	Above Average	76.826
	Average	76.362
	Elite	77.378
CU	Above Average	67.116
	Average	70.121
	Elite	72.487
FA	Above Average	82.761
	Average	83.652
	Elite	85.349
FC	Above Average	80.897
	Average	81.827
	Elite	81.627
FF	Above Average	80.343
	Average	82.5
	Elite	83.626
FS	Above Average	81.067
	Average	74.5
	Elite	78.182
SI	Above Average	84.797
	Average	79.417
	Elite	83.08
SL	Above Average	76.086
	Average	78.027
	Elite	77.772

You can see above that an elite fastball is at least 2 MPH faster than the other groups giving them a definitive advantage when compared to others. Interestingly, the average has a higher end fastball speed than the above average ones. The key though is comparing this speed to the end speed of the off-speed/breaking pitches. The difference between an average changeup and his fastball is about 6 MPH while an elite pitcher’s is almost 8 MPH. For a major league hitter this is a huge difference, especially when located well. The same differences show up again with the slider vs fastball comparison. An elite difference is about 8 MPH, an above average difference is 6 MPH, and an average pitcher difference is only 5 MPH. To give you an example of how much of a difference this is to MLB hitters, an elite fastball reaches the plate in about .48 seconds while an average one reaches it in .49 seconds. For every hundredth of a second, it equals 6 inches less distance the elite pitcher gives to a hitter. This hundredth of a second is the equivalent of 6 extra inches to recognize the pitch.

The difference from changeup to fastball is greater. An elite pitcher has about a 3.2 foot advantage over the hitter with the speed difference. An average pitcher only has a 2.7 foot advantage. That is quite a difference. Imagine the difference of seeing the ball all the way to the bat compared to blinking your eyes at the end while trying to hit the ball.

There are two other things that really make an elite pitcher as good as he is, controlling his pitches and getting good movement on them. The next 3 graphs give different views, one from the side, one from top and one from the catcher’s point of view. They are broken down by pitch with each category of pitcher color coded. We’ll review each of the perspectives separately. We will also at times be referring to the speed chart above.

Catcher Perspective

Starting off with the catcher’s perspective, the movement on the fastballs of the elite group and the above average group are basically the same. Remember the difference in speed though, close to 2.6 MPH. The average pitching group appears to be slightly more vertical than the other 2 groups even though it is skewed over to the right. The curveball of the elite and above average group again, match up almost exactly. The average curveball approximates this as well. The curveball speeds are also somewhat telling but slightly intriguing as well. The average group curveball and the elite group curveball have basically the same speed difference (13 MPH) from the fastball, while the above average group is slightly slower (15 MPH).

Most of the other pitches also approximate each other but there are two that really stand out. The changeup, specifically, differs greatly from each other. The elite group starts out matching the above average trajectory, but it continues to move towards the outside part of the plate. The above average and average group though seems to focus their changeups on the inside part of the plate. The average trajectory of the changeup is much more vertical than the other ones. The other quite difference pitch is the FF, which you can see has quite a bit more movement on it than the other groups.

Top View

This view is much better at viewing the side to side motion of the groups. All of the initial release points are at the top of the charts, so a right handed batter would be standing on the left hand side of the chart.

The elite group has much more movement horizontally than others especially on his fastball and changeup. The movement of an elite FF has really late break in the last 20 or so feet. All of the average pitches exhibits less horizontal movement from this view.

Side View

This will help show the amount of drop in a pitch from each group. Many scouts are much more concerned with the vertical movement of a pitch than horizontal movement. There are several pitches here which really show the differences between the groups. The change-up, sinker, and slider all exhibit more drop than their counterparts, especially the slider and the sinker.

Pitcher perspective

In this 3-D perspective, we can see the other perspectives coming together. We can see just how much more movement the elite pitchers have over the other 2 groups. Those pitches mentioned above really get brought to our attention by how far they cross over.

Summary

I think this helps to explain some of the differences between elite pitchers and others. Speed to a large degree cannot be taught. Movement to a certain degree cannot be taught either. To me, when evaluating players, if a pitcher exhibits these characteristics noted above consistently, they are much more likely to be an elite pitcher than those who don’t.

Boston Rotation PFX Evaluation

Instead of focusing on one player, I wanted to take a different look at things. Let’s take a look at the Red Sox rotation as a whole. Looking at it I think it breaks down fairly well.

Beckett – Elite

Lester – Above Average

Matsuzaka – Average

Wakefield – Slightly Below Average

Buchholz – Below average

So I want to see what makes these pitchers different. Why is one successful, while the other are not. First thing I look at is the release point. Instead of the typical view, I decided to move it around some to really help gain some perspective on the difference between the pitchers.

Obviously Lester is the outlier on our right hand side since he is the lone left handed pitcher. What is interesting though is the differences between all of the other right handed pitchers. I expect a difference from Wakefield due to the knuckleball, but everyone else has a fairly distinct release point from each other.

Release Point

1. Matsuzaka is the furthest out along the horizontal axis and also the widest of the groups which leads me to believe he has a lower arm slot and also a less repeatable delivery than the other pitchers. After seeing the intricacies in his windup this makes perfect sense to me.
   
2. Beckett is next but one major difference I see is how tight his release point is. It's not nearly as wide as Matsuzaka's and also isn't as tall. The tighter a release point, the less the hitter has to recognize which pitch is coming. It's also the lowest among all of the starts which may mean several things. First, it could be due to a player being shorter than others but this is not the case because Beckett is 6'4". Another possibility is he strides longer than most pitchers because as the farther you stride, the lower your body has to be. That's what I think is going on with his release point.
3. Lester's release point is the next tightest, especially horizontally
4. Buchholz's release point looks just like Matsuzaka's except slightly higher.
5. Wakefield's is all over the place but that is somewhat to be expected with a knuckleballer.

Note: The interesting thing I saw in the chart to the right is that most of the balls that were hit into play were pitches that were centered in the release point group. I personally thought we'd see more along the fringes but instead they're centered. This is likely due to the frequency of pitches in the center and also maybe a slight over-estimation of the importance of the cluster size in relation to performance.

Movement

This is the one aspect that is always important when evaluating pitchers. First, let's look at it by pitch.


A couple of interesting things that I find here. In the first chart, clearly has a 2-seamer and a 4-seamer. Second it looks likes there is a slight second grouping to Beckett's fastball but not nearly as distinct as Lester's grouping. Lastly in that first chart, both Matsuzaka and Buchholz have very little control of their sliders. They don't know where the pitch will end up and sometimes it just rolls in there for the hitter to hit.

In the second chart, I broke it down by pitcher to see if there is overlap between the vertical movement on the speed of the pitch. Horizontal movement really doesn't have nearly as much of an effect on ineffectiveness as the vertical break does according to several scouts. Also, Buchholz's slider and change-up are basically coming in the same way and thus are likely slightly easier to hit. With the professional game, that's all that is needed. Looking at the pitchers who have been relatively successful this year (Lester and Beckett) and compare them to those who haven't (Buchholz and Matsuzaka), it's clear that the successful ones have very little overlap while the unsuccessful ones have at least a decent amount of overlap.

Pitch Result breakdown
last			pitch_type
				CH	CU	FA	FC	KN	SL	Total
B e c k	B	Count	5	1	93	216	2			317
		%	25.0%	50.0%	43.9%	31.5%	40.0%			34.1%
	S	Count	13	0	90	323	1			429
		%	65.0%	0.0%	42.5%	47.1%	20.0%			46.2%
	X	Count	2	1	29	147	2			183
		%	10.0%	50.0%	13.7%	21.4%	40.0%			19.7%
	Total	Count	20	2	212	686	5			929
		%	100%	100%	100%	100%	100%			100%
B u c c h	B	Count	6	21	42	86			29	186
		%	66.7%	38.2%	36.2%	36.8%			37.2%	37.7%
	S	Count	2	24	58	104			41	229
		%	22.2%	43.6%	50.0%	44.4%			52.6%	46.4%
	X	Count	1	10	16	44			8	79
		%	11.1%	18.2%	13.8%	18.8%			10.3%	16.0%
	Total	Count	9	55	116	234			78	494
		%	100%	100%	100%	100%			100%	100%
L e s t e r	B	Count	18	25	66	261	39		9	418
		%	43.9%	53.2%	34.9%	33.9%	33.6%		37.5%	35.2%
	S	Count	15	16	84	342	60		10	527
		%	36.6%	34.0%	44.4%	44.5%	51.7%		41.7%	44.4%
	X	Count	8	6	39	166	17		5	241
		%	19.5%	12.8%	20.6%	21.6%	14.7%		20.8%	20.3%
	Total	Count	41	47	189	769	116		24	1186
		%	100%	100%	100%	100%	100%		100%	100%
M a t s u z a	B	Count	9	30	27	229	44		56	409
		%	29.0%	51.7%	55.1%	43.5%	35.2%		28.6%	40.1%
	S	Count	13	14	17	231	46		107	443
		%	41.9%	24.1%	34.7%	43.8%	36.8%		54.6%	43.5%
	X	Count	9	14	5	67	35		33	167
		%	29.0%	24.1%	10.2%	12.7%	28.0%		16.8%	16.4%
	Total	Count	31	58	49	527	125		196	1019
		%	100%	100%	100%	100%	100%		100%	100%
W a k e	B	Count	56		0	3		260		321
		%	37.1%		0.0%	27.3%		34.8%		35.2%
	S	Count	70		0	6		339		416
		%	46.4%		0.0%	54.5%		45.4%		45.6%
	X	Count	25		1	2		148		176
		%	16.6%		100%	18.2%		19.8%		19.3%
	Total	Count	151		1	11		747		913
		%	100%		100%	100%		100%		100%

The chart above breakdown whether or not the type of pitch resulted in a ball/strike/or hit into play. As mentioned above Buchholz's slider and change were hit above average compared to others off speed stuff. Matsuzaka throws his CH, CU, and FA for balls for often then strikes so it is no wonder he has as many troubles as he does. I think this really does show that you have to have command of your fastball in order to be successful.


Showing the release point of the balls/strikes/in play shows, confirms what we saw with the 2 dimensional view earlier. More balls and strikes were seen on the periphery of the release point cluster and balls were hit in play more often in the middle of the cluster. Could it be that the change in release point throws off the hitter some? I think a little more investigation is needed at another time.


Velocity


These two aren't as telling as the others but it does show that the elite pitcher of this group has a higher velocity than the others.

Summary: I think more investigation needs to get done but I think I have shown some key things. First, release points do make a difference in terms of size, especially horizontal size. Second, overlapping of vertical movement at the same speed on the different pitches lead to problems. When a slider comes in like a changeup, only bad things will happen.

I'm going to try and focus on more and see if this a trend or if this is merely a coincidence with the Red Sox.