Which small sample breakouts are for real?

I’ve been working on a few things that are taking longer than anticipated, so I figured I’d try a fun little exercise this week. As you know, I’m a big believer in the combination of stats and scouting, having attended the MLB Scouting Bureau’s Scout Development Program last fall. Scouting can provide insight into things that numbers alone can’t.

Today, I thought I’d use more subjective means to try and evaluate pitchers who are performing well in 2010 but who have no previous track record of such performance. Essentially, I’ll look at 10 pitchers and try to decide if their 2010 strikeout rate success is sustainable using PITCHf/x and other relevant data. At the end of the year, I think it’ll be very interesting to go back and see how accurate these end up being.

Projection systems take a very formulaic approach to predicting the future, taking many years worth of data into account and regressing everyone to the league mean. This works very well on the whole, but some guys will get missed in the shuffle, and I don’t believe that everyone should be regressed to league average. Some players change in significant ways, and projection systems aren’t particularly adept at identifying these guys. The point of this exercise is to try and do just that.

For purposes of comparison, I’ll use Oliver’s rest-of-season projections and come up with my own rest-of-season predictions so that we have some way of testing this at year’s end. Please take my predictions with a grain of salt, though. I’m putting a precise number to it for the purposes of end-of-season testing, but they’re mostly just coming off the top of my head, a rough estimation of how I feel about a player. The general point is more important than the precise number.

Now, onto the players.

*Credit to TexasLeaguers.com for the PITCHf/x graphs.

Clay Hensley

IP: 27.2
K/9: 11.39
Previous career high K/9: 6.00

20082010

image

image

Hensley was a starter for most of his career and has never posted a K/9 above 6.00. This year, though, he’s dominating out of the Marlins pen and could be next in line for saves behind Leo Nunez. I’m definitely buying into Hensley based on a few factors, the biggest being his improved stuff. His breaking ball (which may or may not be two different pitches) is getting terrific movement with a wide range to keep hitters guessing. He’s also reversed the usage of his curveball-like breaking balls and slider-like breaking balls, using the curve more, which generates more swings-and-misses. His changeup is being used more often and is getting more sink.

Combine this with Hensley’s full-time move to the bullpen and pitching in Florida (which inflates Ks by 10.1 percent), I wouldn’t have a problem predicting Hensley to strike out at least a batter per inning the rest of the way.

Verdict: Largely sustain
Oliver ROS K/9 Projection: 6.3
Carty ROS K/9 Projection: 10

Sergio Santos

IP: 18.2
K/9: 11.09
Previous career high K/9: N/A

image

Santos is a converted pitcher, having been a prospect at shortstop up until this year. While we’re dealing with a very small sample (much smaller than Hensley), it does look promising. He’s throwing his fastball an average of over 95 MPH, getting pretty good rise and horizontal movement. He’s only thrown 45 or so sliders, but so far it appears like a perfectly suitable secondary offering as he keeps it down and gets a good amount of “slide” on it (3.4 inches). He’s thrown even fewer changeups, but it looks pretty good as well. I’m not as confident in Santos as I am on Hensley, but I’d have no problem owning Santos and expecting him to continue performing well in an AL-only league. A K/9 over 8.00 seems fine to me.

Verdict: Partially sustain
Oliver ROS K/9 Projection: 7.8
Carty ROS K/9 Projection: 8.5

Vicente Padilla

IP: 21.2
K/9: 9.55
Previous career high K/9: 7.02

20092010

image

image

A Hardball Times Update
Goodbye for now.

Padilla is currently on the DL (he figures to return in mid-June), but before going down, his strikeout rate was among the league’s best. If it looks sustainable, Padilla could be a great bargain right now considering his DL-stint and 6.65 ERA (compared to a 3.92 xFIP). Unfortunately, however, it doesn’t look like Padilla will sustain it. His stuff is almost identical to what it was in 2009 (ignore the pitch classifications, regroup them with your eyes and see how everything is moving). Considering that we have nearly 1,500 innings prior to 2010 saying Padilla is a 6.25 K/9 pitcher and no evidence to support a change in skills, I’m not buying into him at all.

Verdict: Full regression
Oliver ROS K/9 Projection: 6.3
Carty ROS K/9 Projection: 6.2

Carlos Silva

IP: 60.2
K/9: 6.23
Previous career high K/9: 4.05

20092010

image

image

Carlos Silva is having a career revival in Chicago, and it doesn’t look completely unsustainable (yay me! I own him in LABR). The move to the NL helps, and Wrigley boosts Ks about 0.5 percent over the Metrodome (not much, but I figured I’d point it out). His stuff also looks a bit different. His changeup has gained separation from his fastball, thrown 1 MPH slower and getting great tumbling action. His slider has also been slowed down over 3 MPH and is showing more downward bite. Don’t expect anymore games like his 11 K outing this past week, but Silva could definitely be a serviceable pitcher the rest of the way (especially in NL-only leagues) if he can keep his walk rate down (which is pretty likely considering it’s right in line with his career average).

Verdict: Partially sustain
Oliver ROS K/9 Projection: 4.4
Carty ROS K/9 Projection: 5.6

Hisanori Takahashi

IP: 38
K/9: 10.42
Previous career high K/9: 7.2 (Oliver MLE)

image

Takahashi is having a terrific first year in America at 35 years old. Unfortunately, we don’t have any PITCHf/x data from Japan to compare his 2010 to, so projecting what he’ll do is difficult. Still, he has a full repertoire (five different pitches) and pretty good stuff. None of his pitches stick out as being tremendous, but they all appear to be at least league average (except for maybe the fastball, which is only thrown at 88.5 MPH) and two or three are probably plus. I probably won’t differ much from Oliver here, as we don’t really have a point of comparison and nothing is eye-popping. Plus, most of Takahashi’s innings thus far have been in relief, which boosts Ks by around 15 percent on average. He’s in the rotation now, and if he sticks there, a K/9 in the mid-to-high-7’s seems about right.

Verdict: Partially regress
Oliver ROS K/9 Projection: 7.1
Carty ROS K/9 Projection (SP): 7.8
Carty ROS K/9 Projection (RP): 9.0

Matt Belisle

IP: 32.1
K/9: 9.46
Previous career high K/9: 6.39

20092010

image

image

Belisle spent most of his career as a starter, and I always thought he was underrated by the Reds, never really got a fair shake. He was moved to the bullpen in Colorado last year, but still only managed a 6.4 K/9. This year it’s above 9.0, although he doesn’t appear to be doing anything differently. His pitches are all moving the same and coming in at the same speed (his fastball is actually about a half MPH slower). I think his K/9 might have been a little low last year (it was only 31 innings), but I wouldn’t expect his 2010 production to continue.

Verdict: Regress
Oliver ROS K/9 Projection: 5.9
Carty ROS K/9 Projection: 6.8

Sean Marshall

IP: 24.2
K/9: 11.31
Previous career high K/9: 7.99

20092010

image

image

Like Belisle, Marshall was a one-time starter, but he never seemed as promising to me as Belisle. His K-rate is absurd this year, but also like Belisle, it doesn’t appear to be for real. If anything, his curveball (his most-used and nastiest pitch) looks a bit worse. The rest of the pitches all look similar, and he doesn’t appear to have the repertoire of a K-per-inning guy. In all honesty, it would be hard to do that with a fastball as mediocre as Marshall’s, even in the bullpen. Most of his secondary offerings are pretty good, but I’m expecting regression here.

Verdict: Regress
Oliver ROS K/9 Projection: 7.4
Carty ROS K/9 Projection: 7.7

Jesse Chavez

IP: 21
K/9: 9.00
Previous career high K/9: 9.60 (small sample)

20092010

image

image

Chavez has only been in the majors for three years, and he came into the 2010 season with a career 6.9 K/9 in 82.1 innings. He flashed this kind of dominance just once before (excluding the lower levels of the minors), over 15 innings in 2008 when he posted a 9.6 K/9. Still, there is some reason to be optimistic about Chavez. First, he’s now pitching with the Braves, and Turner Field inflates Ks 12 percent over PNC Park.

Second, his stuff has improved a bit. His changeup is running and tumbling more, and he’s getting a lot more swings-and-misses on it. He also seems to have added a very nice curveball a mere two weeks ago. If he increases his usage of it, hitters could be in serious trouble. His slider is a solid pitch, and he throws his fastball hard with good rise, so a lot of strikeouts wouldn’t look out of place with Chavez.

Verdict: Largely sustain
Oliver ROS K/9 Projection: 7.3
Carty ROS K/9 Projection: 8.6

Roy Oswalt

IP: 69
K/9: 9.00
Previous career high K/9: 8.03

20092010

image

image

Oswalt has always been considered an “ace,” but he’s always achieved it with merely above-average K rates combined with excellent control and very good ground ball rates. This year, he’s pushed the K rate into elite territory, but it doesn’t look like it will continue. His pitches are moving the same and have the same velocity, so after seven years of sub-8.0 K/9 rates and at age 32, it seems unlikely that Oswalt has suddenly figured out a way to strike out a batter per inning with the same stuff.

Verdict: Full regression
Oliver ROS K/9 Projection: 7.1
Carty ROS K/9 Projection: 6.8

Mike Leake

IP: 66
K/9: 6.14
Previous career high K/9: N/A

I’ve talked about Leake at length here, and he’s continued to pitch well. I like him a lot, so here’s my prediction for him.

Verdict: Improve
Oliver ROS K/9 Projection: 6.1
Carty ROS K/9 Projection: 6.7

Concluding thoughts

Feel free to let me know if you have any questions. As I said in the beginning, this is mostly just a fun exercise, but it will be interesting to see how things turn out at the end of the season. And, of course, all the usual PITCHf/x caveats apply.


9 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Mike Podhorzer
13 years ago

Great stuff Derek. I have another pitcher request- James Shields. 8.9 K/9 vs previous career high of 7.7 and just 6.8 last year. When I’ve watched him pitch, it appeared his fastball velocity was much higher as I didn’t ever recall him touching the mid-90’s in the past. Sure enough, his average velocity is up a full mile per hour. I would figure that would make his change-up that much more devastating.

Unfortunately, his opposition contact rate and swinging strike percentage are actually down and at career worsts. So though his stuff appears better to my eye, the results aside from his strikeout rate have been worse, which is confounding. Analyze away!

Huh?
13 years ago

You included Jesse Chavez in a column about pitchers who are performing well in 2010.  I am confused.

Josh
13 years ago

I’m very interested in this type of analysis. I’m very numbers-oriented and I tend to read much into small samples early in the year. Being able to scout players is something I’m trying to work on. Thanks again, this was fantastic. Bookmarked.

Like Mike P. up there, James Shields is the name that comes to mind after reading. I’ve noticed what Mike has, that the velocity is up but the swinging strike and contact %s are in line or even worse than his previous levels.

Nutlaw
13 years ago

Very interesting stuff. That made for a nice read.

One question, however. In at least a couple of places, you make mention of strikeout rates being affected by stadium. I’ve never heard of that. Why would that be?

Derek Carty
13 years ago

Thanks, Mike.  I don’t see a whole lot to support Shield’s strikeout-per-inning.  His fastball is up about a MPH, which should add some strikeouts, but not as many as he’s getting.  It should be noted that his change is also coming in about 1 MPH faster, though.  He is, however, getting a lot more swings-and-misses on the change this year.  The only real difference I see in it is the 1 MPH speed increase and that it appears to be getting about an inch more arm-side run.  It looks like he’s throwing more 2-seamers than 4-seamers, although it looks like his fastballs are being classified differently this year, so I don’t think that says much.  And 4-seamers generally are better for strikeouts anyway.  His other pitches seem to be moving about the same.  If anything, his slider/cutter seems to be getting a little less tilt, although that could be the sample size talking.

I’m only taking a cursory look here, but I don’t see anything that jumps out.

Derek Carty
13 years ago

Huh?,
I know Chavez has an ERA over 7, but a lot of that is bad luck (brought on by the small sample size).  His xFIP is 4.43 (and that would be much better if he were getting league average GBs like he did in 2008/2009).  Also, I’m mostly focusing on the K/9 here, using this exercise to see if this kind of analysis can predict players sustaining their K/9 rates (not to suggest that this will be in any way conclusive at the end of the year anyway, but it’ll be fun to look at).

Derek Carty
13 years ago

Nutlaw,
This is a question I get often, and people often don’t realize that parks can affect strikeout rates – and quite dramatically, at that.  The first person I know to talk about this was THT’s David Gassko a couple years ago here: http://www.hardballtimes.com/main/article/batted-balls-and-park-effects/

While I don’t think we have a definitive answer for the “why” (David suggests that “humidity is one of the biggest determinants of strikeout park effects”), the effects are absolutely observable, and the numbers I present use 4-years of data (2006-2009) and are properly regressed to the mean, so they should be pretty reliable.

Parks also affect things like walk rate, ground ball rate, pop-ups, and basically everything else that happens on a baseball field – just not always to the extent of HRs or Ks (or to any significant extent at all).

Brian Cartwright
13 years ago

Oliver uses a K park factor of 0.91 for PNC Park and 1.03 for Turner Field.

At higher elevations it’s harder to get a pitch to break resulting in higher contact rates – fewer SO and also fewer BB as there are not as many deep counts.

There are other stadium specific things such as hitting background, but they are hard to identify.

Derek Carty
13 years ago

Thanks for chiming in Brian.  For the record, I’m using K park factors of 0.924 and 1.041, respectively, so we’re close.