Pavano's Adjustments Pay Off in KC

Based on how poorly Carl Pavano pitched in his last start (1.0 IP, 6 H, 3 BB, 9 ER), I said it would take a miracle for him to come up with a quality start against Kansas City. Well, Pavano must have gotten whatever was ailing him out of his system because he looked like a completely different pitcher in his second start with Cleveland. Pavano finished with 4 ER, 8 H, 0 BB, and 8 K over six strong innings. That may not seem like much, but if he can provide those kinds of starts on a regular basis I think Cleveland will be very pleased with him. Plus, Pavano proved that he's not just damaged goods and actually has some gas left in the tank.

The guy's arm seemingly came back from the dead in the span of a week, although I'm still not sure where on this spectrum of extremes Pavano will eventually settle. Maybe all he needed was a wake-up call after a gruelingly long spring, but whatever the case Pavano bought himself a "get out of jail free card" by bouncing back against Kansas City.

I was curious as to what type of adjustments Pavano made between his first and second starts. To do this required more than just the typical box score stats. I wanted to track any changes in velocity, movement, and location on Pavano's pitches between starts. Fortunately, MLB.com archives the results from its Gameday application over the course of a season, allowing fans to access Pitch f/x data for every batter a pitcher faces.

I've ventured into more detailed Pitch f/x analysis on occasion, but I'd like to take a new approach today and try to glean some insight from only a small sampling of individual at-bats. I decided to keep the number of batters from each start down to three this time in order to keep the presentation and workload simple. I did review the data from the entire start before drawing any conclusions, so any analysis tends to be within a broader context. The idea is to key-in on specific pitches and see how their velocity and break may have changed, in addition to how effective they were (like if a certain pitch was thrown for a strike more consistently than before). If you haven't seen MLB's Gameday application before, this link offers a basic explanation of what's displayed in the images below.

4/9/2009 at Texas: TEX 12 - CLE 8

Michael Young - 1st Inning

Batter	Inning	No.	SPD (MPH)	BRK	PFX	PITCH	RESULT
M. Young	1st	1	87	3"	17"	FB	Called K
		2	88	5"	14"	FB	Foul
		3	88	3"	16"	FB	2B

Pavano's approach to Michael Young is pretty typical of his pitch selection to the 12 batters he faced. With the exception of an occasional changeup or slider, Pavano rarely went to anything other than an 86-88 MPH fastball. This allowed Texas to get comfortable early and just sit on the same type of pitch the entire time Pavano was on the mound. Five of Pavano's six hits came off that bland 86-88 MPH fastball. Pavano also did a poor job of locating his pitches. He took an overly cautious approach throughout the game, nibbling around the edges of the strike zone (this trend appears in all three featured at-bats). It didn't take long for Texas to realize that Pavano wasn't going to challenge them, so they started hacking away at anything that crossed into the zone.

Marlon Byrd - 1st Inning

Batter	Inning	No.	SPD (MPH)	BRK	PFX	PITCH	RESULT
M. Byrd	1st	1	79	8"	8"	CHG	Ball
		2	87	3"	13"	FB	Ball
		3	87	4"	13"	FB	HR

Pavano continues to skirt the strike zone against Marlon Byrd due to either a lack of control or a lack of confidence in his pitches. Here, Pavano fails to establish himself in the count by getting behind 2-0. When he does manage to find the zone, it's with a nearly identical, flat, 87 MPH fastball which Byrd launches into the stands for a two-run homer.

Ian Kinsler - 2nd Inning

Batter	Inning	No.	SPD (MPH)	BRK	PFX	PITCH	RESULT
I. Kinsler	2nd	1	86	3"	17"	FB	Ball
		2	88	5"	14"	FB	Ball
		3	86	5"	16"	FB	HR

At this point, Pavano had already given up five runs and just walked Omar Vizquel to start the second inning. He's really got nothing to lose in being more aggressive and trying to throw something other than a fastball (although honestly, I'm not sure how big a role Shoppach played in the terrible pitch selection). Instead, he goes to the exact same type of pitch that got him hammered in the first and gives up another two-run homer after falling behind 2-0.

There's really no excuse for going to the exact same type of pitch and watching it get hit every time. I knew Pavano had to realize what was going on, yet he made little to no adjustment in his approach even after heading to the dugout for half an inning.

4/14/2009 at Kansas City: KC 9 - CLE 3

Coco Crisp - 1st Inning

Batter	Inning	No.	SPD (MPH)	BRK	PFX	PITCH	RESULT
C. Crisp	1st	1	88	5"	12"	FB	Ball
		2	88	7"	13"	SNK	Foul
		3	82	9"	6"	CHG	Ball
		4	82	6"	13"	CHG	Ball
		5	79	14"	10"	CRV	Pop-up (out)

You can see a difference in Pavano's approach against Kansas City from the very first batter. Note the variety of pitches that he uses against Coco Crisp here. Pavano starts out with a nice combination of fastballs and off-speed pitches. The subtle difference between his sinker and standard fastball (both are thrown at similar speeds, but the sinker has more bite to it) fools Crisp and causes him to foul off a pitch that probably looked like another outside fastball before breaking.

Pavano also does a good job of using the entire plate, giving Crisp multiple locations to worry about in the at-bat rather than clustering all his pitches in one quadrant like he did against Texas. I really like the way Pavano started out Crisp with the fastball before gradually decreasing his velocity and ramping up the break of each consecutive pitch. He eventually gets Crisp to pop-out on a 79 MPH curveball.

He may not be pounding the strike zone in this particular example (there was plenty of that later on), but he gives the batter a reason to protect the plate and actually swing at a few pitches outside the zone as opposed to just giving them the exact same look every time.

Alex Gordon - 2nd Inning

Batter	Inning	No.	SPD (MPH)	BRK	PFX	PITCH	RESULT
A. Gordon	2nd	1	92	5"	17"	FB	Swinging K
		2	81	8"	11"	CHG	Ball
		3	91	6"	15"	SNK	Swinging K
		4	83	8"	13"	CHG	Swinging K

Pavano is very aggressive against Alex Gordon, striking him out on four pitches. This approach resulted in Pavano throwing 70% of his pitches for strikes against the Royals, which is excellent given the stuff he had on display that night. Check out the two 91-92 MPH fastballs Pavano serves to Gordon (he hit 90 MPH on a regular basis throughout the start). I didn't even know Pavano could still throw that fast, let alone nail the inside corner with it. Pavano gives Gordon a tempting look inside with a legit heater, inducing a swinging strike. He then moves down and away with another heater before cutting 8 MPH off the previous pitch with a breaking changeup for the strikeout (this time he would have struck out Gordon whether he swung or not).

It was clear early on that Pavano wasn't going to give the Royals batters any room to breathe.

Mike Jacobs - 6th Inning

Batter	Inning	No.	SPD (MPH)	BRK	PFX	PITCH	RESULT
M. Jacobs	6th	1	90	6"	15"	FB	Called K
		2	89	5"	13"	FB	Foul
		3	82	8"	5"	CHG	Called K

Jumping ahead to the sixth inning, Pavano is faced with home run-threat, Mike Jacobs. The set-up for this at-bat is similar to what Pavano faced in Texas. He knows Jacobs can easily take him out if he makes a mistake and there are already two runners on base ahead of him. Unlike with Texas, Pavano chooses to go right after Jacobs with a called strike up and in on the hands. Pavano doesn't allow Jacobs to get his arms extended on the swing and tries to jam him up and in for a second time. This does a great job of setting up the changeup which completely catches Jacobs off-guard after fighting off the last pitch inside. The changeup is located right along the border of the strikezone, but away from the barrel of the bat as Jacobs takes the bait for the strikeout.

This is another example of Pavano setting up his pitches and effectively changing speeds, something sorely lacking in Texas.

Conclusion

Normally, I wouldn't be so interested in a 6.0 inning, 8 H, 4 ER day by the team's third starter. However, Pavano's outing against Kansas City is significant because it answers the question of what Pavano is still capable of in a Major League game. The fact that he racked up 8 strikeouts while walking none leads me to believe that he has the potential to be even better in the future. He displayed the ability to command multiple pitches for strikes, change speeds, set-up his pitches, and get out of jams. Pavano could have easily gone seven innings in this game (he only threw 80 pitches), but Wedge probably decided to sit Pavano in case trouble found him again (a similar approach was taken with Paul Byrd, who was notorious for his late inning breakdowns, although this isn't necessarily the case with Pavano).

Granted, this was all against a struggling KC offense, but a start like this could really boost Pavano's confidence and encourage him to continue this aggressive, varied approach. He may have even validated some mechanical adjustments between starts, as evidenced by the additional 2-4 MPH on his fastball the second time out.

With the pitching staff still sorting itself out, any quality starts from Pavano will give Cleveland some much needed stability in the starting rotation as they struggle to get their season back on track following a 2-7 start. I still have no idea which Pavano will show up in New York on Sunday, but the flash of potential in KC is reason enough to be at least a little more optimistic about his future with the Tribe.

All Pitch f/x images were generated by MLB's Gameday application.

Is Betancourt Broken?

This is the first time I've used any Pitch f/x tools in depth, so this article is a bit of an experiment. If you happen to notice any flaws in my use of these stats or if something is not explained well enough, feel free to leave a comment. I'd like to use more of this type of analysis in the future, but for now there might be a few speed bumps.

Rafael Betancourt took the phrase “lights out” to a new level last season. The Tribe’s bullpen ace put together a career year setting up closer Joe Borowski through the regular season and into the playoffs. Cleveland rewarded Betancourt with a two year extension (plus a club option) worth $5.4 million guaranteed over the winter.

Betancourt has struggled for much of 2008 though, underperforming in almost every category compared to his career numbers. Betancourt may have greatly raised expectations for 2008 given his outstanding 2007 campaign, but to expect those types of numbers a second consecutive year is unrealistic. Betancourt’s 1.47 ERA, .76 WHIP, and ridiculous 8.89 K/BB ratio was the pitching equivalent of Hafner’s 179 OPS+ season and would be extremely difficult to match.

What Cleveland did expect from Betancourt was a productive, reliable reliever for the 8^th inning, but unfortunately Raffy has been unable to deliver. Betancourt has yet to throw at least nine innings in a month with an ERA under 5.00. He has yet to go on an extended scoreless streak and has followed a rough pattern of no more than three or four scoreless innings with a couple flawed outings for most of the season.

Despite a 6.00 ERA and 1.48 WHIP that has shown little sign of improving, Manager Eric Wedge has consistently gone with Betancourt in the late innings of the game. Betancourt has made the majority of his appearances in the 8^th and 9^th innings and in high leverage situations. Opponents have posted a 1.071, .922, and .716 OPS against Betancourt in the 7^th through 9^th innings, respectively.

Has Betancourt fallen so far that he can no longer make a positive contribution to the bullpen? There’s not much to counter Raffy’s inconsistency this season, but what has caused such a steep decline? To find out, I wanted to look at Betancourt’s previous performance, pitch selection, velocity, location, and delivery.

Stat Comparison

Below are Betancourt’s stats for the last three seasons:

Year	IP	ERA	WHIP	K/BB	K/9	BB/9	HR/9	LD%	LOB%
2006	56.2	3.81	1.11	4.36	7.62	1.75	1.11	25.1	71.4
2007	79.1	1.47	0.76	8.89	9.08	1.02	0.45	19.7	86.4
2008	42.0	6.00	1.48	3.00	9.00	3.00	1.71	20.8	65.0
Career	350.1	3.19	1.11	4.33	8.10	2.11	0.92	21.1	75.6

The first items that jump out to me are Betancourt’s BB/9 rate and his LOB%. These two items seem to be driving each other up, since the more batters he issues free passes to, the longer the inning and the more scoring opportunities the other team will get. Even in an average 2006 season, Betancourt stranded over 70% of all baserunners with that number spiking at 86.4% in 2007.

A lack of control seems to be driving up the walk rate, but Betancourt is still missing bats at the same rate as last season, which is kind of odd. I would have expected Betancourt to issue fewer strikeouts if his control was suffering. Betancourt’s combination of peripherals basically comes from a surplus of opportunities. An increased number of walks (3.00/9 IP) and hits (.288 AVG against compared to .188 last year) issued makes the average appearance longer, thus providing more opportunities to strike people out, albeit in a less efficient manner.

Betancourt’s strikeout rate is somewhat misleading, but the other peripherals are more accurate in terms of his actual performance.

Betancourt is also giving up about the same number of line drives (balls in play more likely to fall for a hit) as in 2007, but this could be deflated by a lack of opportunities, i.e. the batter is walked rather than hitting a ball into play. The balls that do get hit seem to be leaving the yard at a much higher rate than normal though (1.71 HR/9 compared to 0.92 career).

Pitch Selection and Velocity

Here is a table showing the percentage of time a pitch was used with the average velocity in parentheses. The pitch selection data was taken from Betancourt’s Fan Graphs page.

Year	Fastball	Slider	Change	Total Pitches
2006	85.9% (92.3)	4.0% (82.2)	4.4% (84.1)	903
2007	85.8% (91.4 mph)	10.1% (81.8)	4.1% (82.9)	1161
2008	76.1% (91.4)	13.1% (80.9)	10.8% (83.1)	747
Career	83.6% (91.8)	7.0% (81.6)	5.2% (83.4)	3845

I included data for the past three seasons, but Betancourt only used a curveball in 2006 5.7% of the time, so I excluded it from the table. The percentages will not add up to 100% because there was a small percentage from each year that could not be classified as any one pitch.

Betancourt’s average velocity has changed very little between 2007 and 2008. I was surprised to see no change in his fastball velocity, but he did lose about one MPH from his slider. Another surprise was the degree to which he lessened the use of his fastball. In terms of velocity, the fastball doesn’t seem to be any less effective, yet it has given way to a 6.7% increase in changeups thrown and a 3.0% increase in sliders.

According to a recent brief from Indians.com,

Betancourt has battled some back issues and tweaked his mechanical delivery a bit. The change in command has included Betancourt not relying completely on his fastball, which has historically been both his predominant and dominant pitch.

This is the first I’ve heard of Raffy’s back problems and mechanical changes, so I’m not sure if they occurred mid-season or in Spring Training. It seems that this reduction in the use of his fastball and reliance on secondary pitches were prompted by these injuries and on advice from the coaching staff.

One thing that I’m not sure of is how much last season’s work load is affecting Betancourt this season. Betancourt threw a career high 89.1 innings, including the playoffs, in 2007 and may be suffering from some wear and tear. This may be related to the back issues he has experienced, thus the need to alter his delivery.

Delivery and Location

I used a set of Pitch f/x data from 2007 and 2008 to try and identify any changes in Betancourt’s pitch location and delivery. All the Pitch f/x data and graphs below were compiled by Josh Kalk and can be found online at The Hardball Times and his blog.

Plot of Release Points for 2007

Plot of Release Points for 2008

The two release point tables basically show what their name implies; the point at which the ball leaves Betancourt’s hand during his delivery. This shows Betancourt’s arm slot, or the range in which his arm passes during a pitch, and any inconsistencies or trends that appear between the two seasons. Since the data shown is only a sample of Betancourt’s total pitches thrown it’s best to think of the data as an average representation.

Using the graphs, I made a rough measurement of Betancourt’s arm slot range and broke it down by pitch type (some pitchers vary their arm angle depending on the pitch). I focused on the difference in the arm slot between the two seasons to see how much Betancourt may have altered his mechanics. Note that the scale used is in feet, so even a small change could have a noticeable impact on how a pitch crosses the plate.

Arm Slot Ranges for 2007

Pitch	Avg. Horizontal Release Point Range	Horizontal Range Difference	Avg. Vertical Release Point Range	Vertical Range Difference
Fastball	-3 : -1.25 ft.	1.75 ft.	5.8 : 6.5 ft.	0.7 ft.
Slider	-2.75 : -1.5 ft.	1.25 ft.	5.8 : 6.6 ft.	0.8 ft.

Arm Slot Ranges for 2008

Pitch	Avg. Horizontal Release Point Range	Horizontal Range Difference	Avg. Vertical Release Point Range	Vertical Range Difference
Fastball	-2.7 : -1.5 ft.	1.2 ft.	5.8 : 6.8 ft.	1.0 ft.
Slider	-2.6 : -1.6 ft.	1.0 ft.	6.0 : 6.8 ft.	0.8 ft.
Changeup	-2.7 : -1.7 ft.	1.0 ft.	6.1 : 6.8 ft.	0.7 ft.

In 2007, the overall arm slot range for the horizontal axis was 1.75 ft (21 in) and a vertical axis of .7 ft (8.4 in) for his fastball. Similar numbers appear for the slider with a 1.25 ft (15 in) vertical axis and a 0.8 ft (9.6 in) horizontal axis. I keep referring back to 2007 because whatever Betancourt was doing back then was extremely effective.

In 2008, Betancourt is using a wider horizontal range to deliver his slider, but has not altered the vertical range. The biggest change is in his fastball delivery, showing a .55 ft (6.6 in) difference in the horizontal range and a .3 ft (3.6 in) difference in the vertical range between 2007 and 2008. This is actually a very small difference in deliver considering the precision of the data and the amount of times the pitcher has to repeat the motion.

Given how large Betancourt’s normal arm slot is (21 in by 8.4 in for the 2007 fastball), a small change like this may have pushed him just over the edge and affected his pitch location. This is just a theory though, since a difference of 6.6 in and 3.6 in just seems so small when compared to how big the overall range is. Another theory is that the small change in his release point is simply result of intentional mechanical changes and is not a detrimental effect from a tired arm or other ailment. As far as I can tell, it does not appear that Betancourt is suffering from any major inconsistencies due to a change in his release point.

Vertical versus Horizontal Pitch Movement for 2007

Vertical versus Horizontal Pitch Movement for 2008

The two graphs above show the final location of Betancourt’s pitches as they cross home plate. Home plate is represented by point zero on the horizontal axis. A pitch located on the negative horizontal axis would be inside to a right-handed batter and off the plate to a left-handed batter.

Difference in Average Movement for 2007

Pitch	Difference in Avg. Horizontal Movement	Difference in Avg. Vertical Movement
Fastball	10 in.	14.5 in.
Slider	12.5 in.	12.5 in.

Difference in Average Movement for 2008

Pitch	Difference in Avg. Horizontal Movement	Difference in Avg. Vertical Movement
Fastball	11 in.	10 in.
Slider	8.5 in.	8.8 in.

First, Betancourt’s fastball has lost a full 4.5 in of vertical movement between 2007 and 2008. Considering how small the difference between a homerun and a groundout can be for a batter, I think this is a significant factor. While the fastball still runs in to rightys, it tends to run flatter on the vertical axis than it used to. This lack of vertical movement could make it easier to make contact with the ball and send it farther.

Second, Betancourt’s slider has seen a drastic drop-off in overall movement. The slider has a difference of 4 in on the horizontal and 3.7 in on the vertical axes between 2007 and 2008.

Just looking at the graphs shows a lot less movement on the slider. This could result in more hanging sliders that don't drop off or sliders that are less effective in fooling the batter and setting up the fastball properly. Considering Betancourt has actually increased the use of his slider by 3% and reduced the use of his fastball by 9.7%, this seems like a really bad combination. Unfortunately, there isn’t a comparison for the changeup because it wasn’t recorded by Pitch f/x in 2007 for some reason.

I think Betancourt is struggling to integrate the new mechanics and secondary pitches into his routine. The significant reduction in using his fastball combined with a seemingly steep learning curve in mastering the slider appear to be the biggest culprits in Raffy’s poor season. There is a good chance that Betancourt will eventually become more comfortable with these changes and revert back to form in the near future.