Tales of the changeup: an analysis of Johan Santana

Who is the best pitcher in baseball right now? Some might answer that question with Jake Peavy or Josh Beckett, but I’d guess that at least 7 out of 10 times, the answer you would get is Minnesota Twins left-hander Johan Santana. Santana is a 28-year-old from Tovar, Venezuela, and after his fourth full year in the starting rotation, he already owns two Cy Young Award trophies.
Now, as Santana approaches the final season of the 4-year, $39.75 million contract he signed three years ago, the Twins appear eager to trade him, and the reported suitors include such teams as the New York Yankees, Boston Red Sox, and New York Mets, subject to Santana’s approval. I’ll leave the predictions of where he’ll land to those who are better qualified or more eager to comment than I am. However, I’d like to take a look at the pitching repertoire and strategy of possibly the best pitcher in baseball.
If you look at the scouting reports, they all talk about Johan Santana’s devastating changeup and how he works to make his throwing motion identical for all pitches. Most scouting reports list three pitches for Santana–fastball, changeup, and slider–and mention that his changeup comes in 15-20 mph slower than his fastball. Were this true, it would be highly unusual. Most major league changeups are 7-10 mph slower than the pitcher’s fastball. A few scouting reports speak of five pitches–two fastballs, a slider, a circle change, and a straight change. The most useful and interesting scouting information I found was an interview from 2006 that Pat Borzi conducted for the Sporting News with Johan Santana and his catcher Joe Mauer.

Santana throws four pitches for strikes-four- and two-seam fastballs between 92 and 95 mph, a slider/curve in the 84- to 87-mph range and a changeup that’s about 15 to 20 mph slower than the fastball. The changeup is his strikeout pitch; when Santana is on, he throws it from the same arm angle and release point as his fastball, and hitters can’t tell the difference until it’s too late.

I also found this quote from Santana interesting given that most people acknowledge his changeup as his best pitch:

“I want to make sure my two-seam fastball is working,” Santana says. “That’s my best pitch, and it’s going to make my other pitches look even better. That’s what I try to do all the time.”

We have detailed data from the PITCHf/x system for 1032 of Santana’s 3345 pitches during the 2007 season. Let’s dive in and see what we can learn about Santana’s repertoire and effectiveness with his various pitches.
Santana Pitch Speed vs. Spin Force Angle
Santana has at least three obvious pitch groupings: fastball, changeup, and breaking ball. Here I’ve shown two graphs that I use for pitch classification. The first graph shows the speed of his pitches versus the direction they break, in polar graph format. The second graph shows the movement on his pitches in the last quarter-second before they cross the plate, due to the forces of spin deflection and gravity.
Santana Late Break
The fastballs run 89-95 mph, and it’s hard to tell from these graphs alone whether Santana really does throw two different fastballs or just one. Through additional analysis, which I will explain shortly, as well as Santana’s own comments, I concluded that he did in fact throw a four-seam and a two-seam fastball and have coded them separately in these graphs.
We can also see that Santana throws two different offspeed pitches. One has a movement very similar to the fastball but is thrown slower at 80-84 mph. This is his changeup. It’s interesting to note that we see a 10 mph difference in speeds between his fastball and his changeup, typical of other major league changeups and nothing like the 15-20 mph difference that was reported by other sources. I don’t know if that was just the stuff of legend or whether Santana has changed his approach in recent years. More likely, people were comparing Santana’s very slowest changeup with his very fastest fastball and writing as if that represented a typical pitching pattern.
I could not find any sign of two different changeups in Santana’s repertoire, at least not two changeups that consistently have different movement or speed.
Santana’s other offspeed pitch is an 83-88 mph breaking ball, described in various scouting reports as either a slider or a curveball. Based on the spin direction, the speed, and the direction of break, it’s very clearly a slider. In the first graph of pitch speed vs. spin deflection angle, the calculation of the spin deflection angle for some of the sliders contains a good deal of error since the spin of those sliders is nearly aligned around the direction of travel of the pitch, resulting in spin deflection of only a couple inches or less. This is one of the classic indicators of a slider.
The sliders and changeups look difficult to separate at the margins in the two graphs I presented above, but including the (x-z component of the) spin rate in the discussion makes that task much easier.
Santana Pitch Speed vs. Spin Rate
Returning to the topic I mentioned earlier, how did I determine whether Santana threw both a four-seam and a two-seam fastball? Looking at the data in aggregate, it was impossible to see a dividing line, but when I examined the spin and break on a start-by-start basis, a little bit of order appeared out of the murkiness. In some starts, two separate groupings were obvious. In most starts, the dividing line was subtle. In a few cases, it was hard to find a dividing line at all. I did notice that the fastballs with the most sink and the slowest speed were thrown almost exclusively to right-handed hitters, and this, in addition to Santana’s own comments about throwing a two-seamer, gave me confidence in making a distinction between the two fastballs.
If you look at the comments from John Walsh and John Beamer on my Erik Bedard analysis, you’ll see that having to examine the data on a start-by-start basis in order to make an accurate pitch classification diagnosis is a recurring problem. We’d like to be able to look at a pitcher’s season data as a whole. This is an important area for further investigation.
Here are a couple more traditionally-used PITCHf/x graphs of pitch movement for those who are interested:
Santana Vertical vs. Horizontal Spin Deflection
Santana Vertical vs. Horizontal Pitch Deflection
How does Santana use his pitches to left-handed and right-handed hitters? As a left-handed pitcher, he naturally sees predominantly right-handed hitters, making up 75% of his opponents. To righties, he throws about 41% four-seam fastballs, 35% changeups, 18% two-seam fastballs, and 6% sliders. To lefties, he throws 60% fastballs, 29% sliders, 7% changeups, and 4% two-seam fastballs. Against righties he’s the stereotypical fastball-changeup Santana that I’ve heard about. Against lefties, he’s a totally different pitcher, eschewing the changeup and the two-seam fastball and relying on a fastball-slider combination.
Next, let’s look at how Santana mixes his pitches in different ball-strike counts. I’ve split this out by batter handedness as well.
Santana Pitch Mix to RHH by Count
Santana Pitch Mix to LHH by Count
Against righties, you can see that the changeup is his favorite pitch with two strikes (57% of the time), and he mixes in his two-seam fastball more if he falls behind in the count (28% when behind vs. 15% when ahead or even).
Against lefties, he’s relies on the four-seamer about 70% of the time in most situations. With two strikes he feels confident enough to occasionally (14%) introduce the changeup to lefties, and on an 0-2 count, you can count on getting a slider two thirds of the time.
What’s the bottom line–what results does Santana get with his pitches? I attempted for a while to cast the answer to that question in terms of run values for each pitch determined by linear weights, but I’ve postponed that endeavor for the moment. There are too many pieces that I haven’t figured out how to put together yet. So here are the results in the same format I used in the Bedard article.

Fastball 0.32 0.20 0.25 0.10 0.13 0.316 0.188 0.842 0.158
Sinker 0.70 0.10 0.00 0.10 0.10 0.000 0.000 0.000 0.000
Slider 0.34 0.13 0.17 0.17 0.19 0.308 0.308 0.462 0.000
Changeup 0.24 0.06 0.18 0.24 0.29 0.400 0.400 0.400 0.000
Fastball 0.32 0.20 0.26 0.12 0.11 0.235 0.188 0.500 0.059
Sinker 0.35 0.17 0.24 0.06 0.19 0.333 0.250 0.741 0.111
Slider 0.38 0.12 0.24 0.08 0.18 0.111 0.000 0.444 0.111
Changeup 0.32 0.08 0.15 0.31 0.15 0.357 0.325 0.667 0.048
Lg. Avg. Ball CStrk Foul SStrk InPlay Avg BABIP SLG HR
Fastball 0.36 0.19 0.19 0.06 0.19 0.330 0.304 0.521 0.037
Slider 0.36 0.14 0.17 0.13 0.20 0.310 0.286 0.481 0.033
Changeup 0.40 0.11 0.14 0.13 0.21 0.319 0.295 0.502 0.035

The league average information comes from John Walsh’s article, and once again I’m using an adaptation of his format to present this information.
Santana 4-seam Fastball Strike Zone Chart
The four-seamer is Santana’s bread and butter, especially to lefties, and a good bit of creamy butter it has. He throws it for strikes and gets more swings and misses with it than most pitchers do. Hitters have a hard time putting the four-seamer into play, and when they do, Santana also gets really good results (a .188 BABIP compared to .304 league average BABIP on the fastball), although lefty batters–Hafner, Sizemore, and Thome–did hit three home runs off the four-seamer in our data set. He mostly pounds the zone with the pitch to both lefties and righties, although there appears to be some tendency toward pitching up and away from lefties and up and in to righties.
Santana 2-seam Fastball Strike Zone Chart
Santana doesn’t use the two-seamer much against lefties, and when he did, it was mostly for a ball. He works in the zone against righties and gets fairly average results with the two-seam fastball. One surprising thing to note is that he still gives up a lot of fly balls off the two-seamer; almost 70% of balls in play off the two-seamer were fly balls. The two-seamer seems like his weakest pitch based on the results we have from 2007, so I’m not sure I understand his statement from the Sporting News interview that it’s his best pitch.
Santana Changeup Strike Zone Chart
Just look at all the red bleeding over the graph from the swinging strikes, and you know all you need to know about Santana’s changeup. The hitters can’t hit it. Santana can throw it for strikes just as well as his fastball. He throws it down and away from righties, and he gets a lot of swings and misses when they chase the changeup down out of the strike zone. When he gets it too close to the heart of the zone, they do make decent contact. It would go without saying, but this is an outstanding pitch.
Santana Slider Strike Zone Chart
Against lefties, Santana uses the slider mostly down and away, and he gets pretty average results with it. Against righties, he features the slider less often. When he does throw it, he keeps it inside. When he gets it up, it gets put in play, but he had fairly good results on a limited sample of balls in play except for one slider that Alex Rios launched 414 feet into the left field seats at the stadium formerly known as SkyDome.
I also looked a bit at pitch sequencing. Here’s a table showing what pitch a hitter is most likely to see from Santana based on what the previous pitch was.

Previous Pitch Fastball Sinker Slider Changeup
Fastball 66% 4% 26% 4%
Sinker 67% 0% 33% 0%
Slider 60% 9% 27% 4%
Changeup 76% 0% 24% 0%
Previous Pitch Fastball Sinker Slider Changeup
Fastball 52% 16% 5% 27%
Sinker 46% 21% 3% 31%
Slider 42% 30% 9% 18%
Changeup 43% 15% 8% 34%

I don’t notice any particular patterns to lefties, but to righties he’s more likely to throw the two-seamer after a previous two-seamer, and he’s more likely to throw a changeup after another changeup.
Johan Santana had yet another great season in 2007. He allowed a few more walks and home runs than in previous years, but without PITCHf/x data from previous seasons, I don’t have any way to know whether that was simply luck or a change in his pitching abilities and strategies.
I looked at the 11 home-run balls off Santana for which we have PITCHf/x data, and I couldn’t detect any useful patterns. They were mostly hit off pitches up and over the plate, but that doesn’t come as much of a surprise. Looking at the HitTracker data, he wasn’t burned by many short home runs barely sneaking over the fence, so he wasn’t unlucky in that regard, at least. This may be a topic for further investigation or possibly just the result of Santana being a fly ball pitcher and getting a little unlucky with how hard the hitters hit 33 of those fly balls in 2007.
Santana obviously has an outstanding changeup and a strong fastball, but you probably knew that already. What I didn’t know was how infrequently he uses the changeup against lefties or most of the other nuances of his pitching strategy. Unless you’re Joe Mauer or Mike Redmond (in which case, Hi!), hopefully you feel like you know the best pitcher in baseball a little better than you did before.
If you’re an employee of a Mr. Steinbrenner or a Mr. Henry gathering information for a future trade, by all means feel free to contact to me regarding where to send that check for my services. 🙂


11 Responses to Tales of the changeup: an analysis of Johan Santana

  1. Anon says:

    You must have a lot of time on your hands….

  2. John Beamer says:

    Great work. This is probably the best pitcher breakdown I have seen to date.
    Keep it up

  3. Mike Fast says:

    I had a very good email comment from SirKodiak that I would like to share with the rest of you:
    “I looked around the internet for pics of Santana’s grips, and came up with a few that you might find useful:
    circle change:
    straight change?:
    The straight change grip seems to be much too deep into the hand to be a fastball. The circle change supposedly comes in two varieties, one that acts just like a straight change where the pitcher does not pronate his wrist, and one where he does pronate his wrist and it acts like a screwball.
    I saw a couple grips that might be his slider, but was never sure.”

  4. What a great breakdown of a player. Amazing job there, Mike.
    As for the photos: Those pictures are all legit. Those are the grips for each. Nice work.

  5. Anonymous says:

    […] also a groundbreaking R&D sub-goal which is to pioneer the use of new sources of data, such as PitchfX.) With an eye toward more truly understanding the game, completely discarding wins or RBI is not […]

  6. […] also a groundbreaking R&D sub-goal which is to pioneer the use of new sources of data, such as PitchfX.) With an eye toward more truly understanding the game, completely discarding wins or RBI is not […]

  7. […] Statistically Speaking has a mind spinning look at Johan Santana that includes Pitch Fx and all types of other colorful graphs. I try to learn as much as a I can about various stats these days and stay knowledgeable regarding how to evaluate talent using them, but this is hard stuff. – (Source) […]

  8. Jesse says:

    Incredible work, this is fantastic. I wish I had the time and resources to delve into things like this!

  9. Mike Fast says:

    In Buster Olney’s blog on ESPN.com Insider yesterday, he mentions that he spoke with a “talent evaluator” who believes the Twins are concerned that Santana hurt himself last year in his 17-strikeout performance against the Rangers and cites evidence that his velocity was down and he used his slider less frequently.
    (If you’re not an ESPN insider, you can check this entry at Matthew Cerrone’s Mets Blog: http://www.metsblog.com/2008/01/17/blog-santanas-physical-condition/)
    I checked the PITCHf/x data, and we have data for 4 of his last seven starts after the August 19 start. There is some evidence of reduced velocity, about 1 mph on average lower, and particularly the last two starts of the year, Santana was barely breaking 92, whereas he usually tops out 94-96. I don’t know whether this is due to injury or just fatigue at the end of the year. I suspect the latter, but I can’t prove it from the data one way or the other.
    On the other hand, the claim about the slider appears to be false, or at least not connected to an injury sustained in the 17-strikeout game. Santana threw the slider 11% of the time through August 19, and 12% of the time after August 19.

  10. Bob says:

    How did you make the circular polar graph?!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: