An effective velocity analysis of pre-injury Michael Wacha

Reports throughout Spring Training have been nothing but positive with regard to the health of Michael Wacha. After an excellent start to the season in 2014, Wacha dealt with a stress reaction in his shoulder which put him on the DL from mid-June through early September. When he returned, he wasn't the same pitcher, and one of the most notable differences was his decreased use of his devastating changeup.

As Joe wrote about last month, though, it appears that Wacha's changeup is back, and this could be extremely important to his success in 2015. Wacha has the small sample size spring stats to suggest that he could have a good year (20.1 IP, 1.77 ERA, 16 K, 4 BB), and by all indications, he looks like the pitcher he was before his injury. In the words of manager Mike Matheny,

"He looks like the guy we saw before, when he was healthy. That's how I'm going to continue to look at it."

Wacha himself seems to feel the same way, as he told Rick Hummel of the Post-Dispatch,

"I've been real happy with the way (the shoulder's) been bouncing back after each start. And the way it feels throughout the game as well."

If these reports are to believed and Wacha is truly healthy and pitching as well as he was pre-injury, then I thought it would be helpful to see just how good Wacha can be by analyzing his pre-injury PITCHf/x data (for 2013 and 2014) using effective velocity. If you recall, Joe and I introduced the concept of effective velocity a couple weeks ago in this post. Joe also wrote a piece that explained the concept of EV in more detail using video. I would strongly recommend that you refer back to them if you have any questions about the concept or terminology.

To start, here's a look at Wacha's PITCHf/x data pre-injury:

Pitch Type	Frequency	Velocity	Vert. Release Pt.	Dragless H. mov.	Dragless V. mov. + gravity
Fourseam	61.82%	94.08	6.93	-2.72	-10.76
Change	23.05%	86.59	6.85	-5.60	-24.97
Curve	9.18%	75.54	6.74	4.50	-52.31
Cutter	5.92%	90.17	6.68	1.44	-19.88

While Wacha has a four-pitch mix, he relies primarily on his fastball/changeup combination. This was especially true in 2013, when he threw just 110 curveballs (7.26%) and 27 cutters (1.78%) all year.

For this post, I decided to collect data on all of Wacha's non-pitcher strikeouts in 2013 and 2014, pre-shoulder injury. I focused especially on the last two pitches of those at-bats, collecting data on pitch type, location, and velocity. While effective velocity is certainly a concept that can be extended beyond the last two pitches of non-pitcher at-bats ending in strikeouts, I thought focusing on these instances would be a good place to start, especially since they, in theory, illustrate Wacha at his best.

Before I go into the data, here's the chart (which was included in the introductory EV post) which details how I adjust from real velocity to effective velocity based on pitch location. (Vertical location: U=Up, D=Down, M=Middle; Horizontal location: O=Out, I=In, M=Middle).

Pitch Location	Velocity Adjustment
UI	+5 MPH
DO	-5 MPH
UM	+3 MPH
MI	+3 MPH
DM	-3 MPH
MO	-3 MPH
DI	No change
MM	No change
UO	No change

With four different pitches in his repertoire, Wacha has the potential to throw 16 different two-pitch sequences to finish an at-bat. In the data collected, he struck out batters using 15 of the 16 different combinations, with cutter-curveball being the only one not used.

Here's a graph which illustrates the frequency with which he used each pitch combination.

As you can see, the pitch sequences involving only fastballs and changeups are far and away the most common, as those four sequences account for 80% of his strikeouts. Because of this, I will discuss each of these four sequences individually and then say a little bit about the rest of the pitch sequences. For each sequence, I will include each of the following: the frequency of the sequence, the average real velocity gap between the two pitches, the number of instances with a real velocity gap of 6 MPH or greater, the average EV gap between the two pitches, and the number of instances with an EV gap of 6 MPH or greater. (The significance of having a 6 MPH difference between pitches is that a hitter expecting a pitch at a certain velocity can still make solid contact with it if it is up to 6 MPH faster or slower than anticipated.) I will also sort the strikeouts by whether the final two pitches had a higher EV gap, a higher real velocity gap, or the "same" gap in EV and real velocity.

Fastball-Fastball

Frequency	Avg. Real Velo Gap	6+ MPH Real Velo Gap	Avg. EV Gap	6+ MPH EV Gap	Higher EV Gap	Higher Real Velo Gap	"Same" Gap
45	1.38 MPH	0	3.11 MPH	5	28	3	14

This was Wacha's most common strikeout sequence, and given the fact that it involves doubling up on the same pitch, it's not much of a surprise that there isn't much of a velocity gap between pitches. On average, there was a 1.38 MPH gap between the two pitches, and none of them had a gap of six MPH or more. With effective velocity, though, the average gap between the two pitches increased to 3.11 MPH, with five strikeouts having a EV gap of six MPH or more. Almost every sequence had a same or higher EV gap than real velocity gap. Since most of the fastballs are nearly the same velocity, this isn't surprising. Mathematically, the velocity gap between the two pitches can only widen using the EV adjustments, since the real velocity gap between the pitches is so small to begin with.

And even when Wacha doesn't widen the EV gap between his pitches, he can still make his mid 90s fastball harder to hit through the use of EV. In one notable example, Wacha struck out Garrett Jones with two fastballs to finish the at-bat, with pitch velocities of 97 and 98. Because both pitches were up and in to Jones, the effective velocities of these pitches were approximately 102 and 103. While the real velocity and effective velocity gaps between the pitches were the same, Wacha still made use of effective velocity by throwing fastballs in spots to maximize their effective velocity and make them that much harder to hit.

Fastball-Changeup

Frequency	Avg. Real Velo Gap	6+ MPH Real Velo Gap	Avg. EV Gap	6+ MPH EV Gap	Higher EV Gap	Higher Real Velo Gap	"Same" Gap
34	7.79 MPH	33	10.59 MPH	33	23	3	8

While we already knew that Wacha's fastball-changeup sequence can be difficult for hitters, his use of EV makes it a particularly devastating combination. He increases the gap between the two pitches from 7.79 MPH to 10.59 MPH using EV. In addition, 33 of the 34 fastball-changeup sequences had an EV gap of over 6 MPH. While this isn't too surprising, since there is already a sizable velocity gap between his fastball and changeup, it is possible to reduce the effective velocity gap between these two pitches by throwing them in certain locations  Wacha does not do this much (or if he does, it rarely results in strikeouts); what happens far more often is that Wacha uses EV to increase the velocity gap between his fastball and his changeup. In 31 of his 34 strikeouts, the EV gap is the same or higher than the real velocity gap.

One of the interesting things I noticed in looking at Wacha's fastball-changeup strikeouts was that 21 of his 34 fastballs were in the upper third of the strike zone or above and 30 of his 34 changeups were in the bottom third of the strike zone or below. It appears that Wacha has been able to effectively locate his pitches vertically to both increase the effective velocity of his fastball and decrease the effective velocity of his changeup. Given the fact that Wacha's changeup has 14 inches more vertical movement than his fastball, these results aren't surprising. In addition, because his fastball and changeup have a similar release point and horizontal movement, these pitches are especially effective when used together. They have the same pitch tunnel, meaning they look the same to the batter even as late as halfway to the plate. With this pitch tunneling effect, as well as the velocity difference between his fastball and changeup, it is much easier to understand how Wacha was able to dominate in the 2013 playoffs despite throwing these pitches 90% of the time.

Changeup-Fastball

Frequency	Avg. Real Velo Gap	6+ MPH Real Velo Gap	Avg. EV Gap	6+ MPH EV Gap	Higher EV Gap	Higher Real Velo Gap	"Same" Gap
27	8.30	25	9.89	24	15	6	6

Much of what I just said applies here as well, as this sequence is simply the reverse order of the previous sequence. Wacha maintains the large velocity gap between his fastball and changeup no matter what order they are used in. He also increases the velocity gap between the two pitches using EV in over half of his strikeouts, although there were six instances in which EV worked against him, based on his pitch locations. Vertical location once again appears to be an important factor in this sequence, as 21 of his changeups were in the bottom third of the strike zone or lower while 15 of his fastballs were in the upper third of the strike zone or higher.

Changeup-Changeup

Frequency	Avg. Real Velo Gap	6+ MPH Real Velo Gap	Avg. EV Gap	6+ MPH EV Gap	Higher EV Gap	Higher Real Velo Gap	"Same" Gap
25	1.00	0	2.92	2	16	1	8

Wacha's changeup-changeup combination, like his fastball-fastball combination, results in very little velocity gap. Wacha increases this gap to nearly 3 MPH using effective velocity, but he only had two strikeouts with a velocity gap of 6 MPH or higher. This sequence, in particular, highlights the importance of looking at more than the last two pitches of an at-bat. Given the fact that Wacha throws his fastball nearly 62% of the time, it is likely that he threw one or more fastballs earlier in the at-bat. Even though the last two pitches were changeups, it's possible that the batter was still prepared for a fastball and wasn't ready for consecutive changeups.

Everything Else

The other eleven pitch sequences make up the remaining 20% of Wacha's strikeouts. Since none of the individual sequences had a sample size of more than eight, I didn't think it would be worthwhile to make tables for each of these sequences. However, a couple things stood out.

First, Wacha actually seems to reduce the effective velocity gap between pitches when he uses his curveball. In the 13 times he used his curve as a setup pitch, he reduced the EV gap between pitches seven times, which is much higher than with any other pitch combination. Perhaps this didn't hurt him all that much, though, since his curveball is so much slower than the rest of his pitches, and he was able to maintain a large EV gap regardless of where his curveball was located.

Also, there doesn't seem to be much evidence of Wacha using EV with his cutter either. Only three of the thirteen sequences that involved his cutter had an EV gap of 6 MPH or higher. Given the fact that he barely even threw his cutter in 2013, it is likely that he is still developing this pitch and has yet to control it with any consistency.

That's all I've got for now. Since this is my first attempt at using PITCHf/x data to talk about EV, all questions, comments, and suggestions are appreciated!