+------------------+------+---------------------+
| Name             | Team | Current Closer      |
+------------------+------+---------------------+
| Scot Shields     | LAA  | Francisco Rodriguez |
| Jose Arredondo   | LAA  | Francisco Rodriguez |
| Jon Rauch        | ARZ  | Brandon Lyon        |
| Chris Ray        | BAL  | George Sherrill     |
| Fernando Cabrera | BAL  | George Sherrill     |
| Carlos Marmol    | CHC  | Kerry Wood          |
| Rafael Perez     | CLE  | Jensen Lewis        |
| Manny Corpas     | COL  | Brian Fuentes       |
| Joel Zumaya      | DET  | Fernando Rodney     |
| Kyle Farnsworth  | DET  | Fernando Rodney     |
| Matt Lindstrom   | FLA  | Kevin Gregg         |
| Eric Gagne       | MIL  | Salomon Torres      |
| Huston Street    | OAK  | Brad Zeigler        |
| Joey Devine      | OAK  | Brad Zeigler        |
| Santiago Casilla | OAK  | Brad Zeigler        |
| Heath Bell       | SD   | Trevor Hoffman      |
| Chris Perez      | STL  | Chris Perez         |
| Dan Wheeler      | TB   | Troy Percival       |
| Grant Balfour    | TB   | Troy Percival       |
| C.J. Wilson      | TEX  | Eddie Guardado      |
| Frank Francisco  | TEX  | Eddie Guardado      |
+------------------+------+---------------------+

---

Posted by Derek Carty at 7:38pm

Wednesday, August 20, 2008

What’s wrong with Johan Santana? (PITCHf/x)

What's wrong with Johan Santana? This question seems to be on the minds of many Mets fan and likely quite a few fantasy owners.

Note: There is a fair amount of things we'll examine, especially when we get to Santana's PITCHf/x data. If you're interested in cutting to the chase and just seeing the conclusions, I'm going to begin including a new section at the end that summarizes what we find.

Numbers

+------+---------+-------+----+------+----------+-----------+-----+------+------+-------+
| YEAR | LAST    | FIRST | W  | ERA  | LIPS ERA | DIPS WHIP | K/9 | BB/9 | xGB% | BABIP |
+------+---------+-------+----+------+----------+-----------+-----+------+------+-------+
| 2004 | Santana | Johan | 20 | 2.61 |     3.06 |      1.00 | 10.5|  2.1 |   39 | 0.252 |
| 2005 | Santana | Johan | 16 | 2.87 |     3.02 |      1.04 | 9.3 |  1.8 |   38 | 0.265 |
| 2006 | Santana | Johan | 19 | 2.77 |     3.25 |      1.04 | 9.4 |  1.8 |   41 | 0.273 |
| 2007 | Santana | Johan | 15 | 3.33 |     3.30 |      1.09 | 9.7 |  2.1 |   37 | 0.275 |
| 2008 | Santana | Johan | 10 | 2.89 |     3.66 |      1.21 | 7.6 |  2.5 |   43 | 0.276 |
+------+---------+-------+----+------+----------+-----------+-----+------+------+-------+

On the surface, things look alright for Santana. The 2.89 ERA is great, although just 10 wins is causing some grief for Mets fans. Some of the slightly more rational ones realize that the low win total isn't Santana's fault, rationalizing that he's doing his job by posting the 2.89 ERA.

These fans would be partially right in that the bullpen has blown a number of leads. Metsblog notes that "[The bullpen has] also blown leads in six, including four of the last five, of Santana’s 24 starts this season, which is one more than the Twins bullpen accumulated in his last 100 starts with them."

The team's offense has, however, provided 5.14 runs of support to Santana but just 4.82 to the team as a whole. Metsblog also makes note, though, that the team "has scored less than three runs or less in 33 percent of his starts this season." So while the average run support number is good, there is huge variation in it. In the starts where the variation is on the low end, Santana is nearly assured of not getting a win.

Overall, the win total seems like a little bit of bad luck combined with overblown expectations from Mets fans. For a pitcher to reach 20 wins—likely what fans were expecting—a pitcher not only needs to pitch well and play for a good team but get quite lucky on top.

The real problem, though, for those who are really playing attention, is that Santana's 2.89 ERA is very lucky. His LIPS ERA shows that Santana should actually have a 3.66 ERA. Imagine what people would be saying if that's where his actual ERA was right now! The thing is, though, it would be a huge mistake to expect an ERA under 3.00 going forward.

So why is the LIPS ERA so high? The culprit is the 7.6 K/9. Santana is striking out two fewer batters per game than he was last year, and that is a major problem. He's also walking a few more batters, but his BB/9 is still at a decent level.

The reason his actual ERA is so low is because, despite the lower K/9, Santana has managed to keep his BABIP at his usual, exceptional level. Very few pitchers can keep a BABIP below .290, but Santana is one of them. It's possible it should be closer to league average and he's just getting lucky since his skills have fallen off, but I wouldn't say it is a certainty it will regress that far. As long as Santana can keep the BABIP better than league average, his ERA should always be projected to be better than his LIPS ERA.

Still, his career high 80 percent Left On Base Percentage (LOB%) means that his ERA is too low despite the BABIP and will rise going forward. Not as high as his 3.66 LIPS ERA, but it will rise (holding all else constant). Further cause for concern is his 6.1 K/9—ewww—since the beginning of July, though we are looking at a sample size of just 47.3 innings.

PITCHf/x

It's clear that Santana isn't himself this year, however you want to look at it. His K/9 is down, but what is causing it? Back at the end of May, the following quote appeared on ESPN (hat tip Hot Foot)

Said an AL scout who has seen Santana this month: “His stuff isn’t even close to what it was [with the Twins].”

Let's use PITCHf/x to check the validity of this statement and to see if we can figure out what's going on with Johan. First, his movement.

Movement

2007

+-------+-----+----------+----------+-------+------------+------------+---------+
| PITCH | %   | % vs LHB | % vs RHB | SPEED | MOVEMENT_X | MOVEMENT_Z | runs100 |
+-------+-----+----------+----------+-------+------------+------------+---------+
|    FB |  51 |       59 |       48 |  92.7 |       7.11 |      11.70 |    -0.8 |
|    SK |  10 |        6 |       11 |  92.4 |       8.47 |       8.84 |     1.0 |
|    SL |  14 |       29 |        9 |  84.7 |       2.74 |       5.01 |    -2.3 |
|    CH |  25 |        6 |       32 |  82.5 |       8.54 |       7.71 |    -1.1 |
+-------+-----+----------+----------+-------+------------+------------+---------+

2008

+-------+-----+----------+----------+-------+------------+------------+---------+
| PITCH | %   | % vs LHB | % vs RHB | SPEED | MOVEMENT_X | MOVEMENT_Z | runs100 |
+-------+-----+----------+----------+-------+------------+------------+---------+
|    FB |  38 |       44 |       36 |  92.9 |       5.07 |      10.76 |     0.1 |
|    SK |  23 |       19 |       25 |  92.7 |       7.31 |       7.75 |    -1.3 |
|    SL |  14 |       28 |        8 |  84.6 |      -0.54 |       4.50 |     0.8 |
|    CH |  24 |        9 |       31 |  81.2 |       6.33 |       7.96 |    -2.6 |
+-------+-----+----------+----------+-------+------------+------------+---------+

Very distinct, noticeable differences between his 2007 movement and his 2008 movement—some good, some bad.

He's getting less rise and less horizontal movement on his four-seam fastball, several of which aren't moving at all horizontally. The same could be said about the change-up except that it's getting a bit less sink. On a positive note, he has managed to add an additional 1.5 MPH difference between the two.

Furthermore, the slider is clearly improved. Last year it actually broke towards a left-handed batter (read Josh Kalk's recent piece on Gavin Floyd for some talk about sliders like this). This year it's breaking away from them, as most good sliders do. The sinker looks to be improved as well, getting more sink (albeit at the expense of less horizontal movement).

Santana's 2008 chart almost looks as if the 2007 chart was duplicated and an extra clump of each pitch amended to the left side of each cluster. Seems like the AL scout quoted above was onto something as far as the four-seamer and change-up go. As far as the two-seamer and slider go, not so much, although most scouting reports I've read don't seem to acknowledge the two-seamer (instead just saying he throws a fastball).

This is understandable because last year the two weren't easy to distinguish between. I only realized Santana threw two fastballs after reading a Mike Fast piece over the off-season and an interview with Santana he linked to. They were difficult to pick apart, though, even going start by start, so the classifications of the two aren't perfect.

It's interesting to note that, despite the improved slider, he's getting worse results with it in terms of runs100—0.8 in 2008 and -2.3 in 2007 (lower is better). I thought this might be because runs100 doesn't account for BABIP and he simply got unlucky, but when I looked at runs100 on balls not put in play, it was still worse. This could mean—ignoring some of the other possible issues, like runs100 being somewhat context driven or it simply being random fluctuation or bad luck—that the increased movement is unintentional and that Santana hasn't figured out how to use this different slider.

Location

Not only is Santana's slider getting better movement, it looks as though he's locating it better too. I haven't conducted any studies on slider location yet, but I did do one on curveballs recently. Against same-handed batters, curveballs are best thrown away and kept away from the middle-third of the zone. Down and in is the worst place to throw a curveball.

While sliders aren't curveballs, they are similar, and I suspect we'll find at least somewhat similar results when we look at sliders. For now, let's tentatively say that Santana is locating his slider better than last year, potentially because of that extra movement. He's throwing to the middle 38 percent of the time now (down from 58 percent) and to the outside 57 percent (up from 33 percent).

Judging by this, I'd say it's very possible that Santana has just received bad luck with the slider, and this is not a case of not having a good feel for it.


If we look at John Walsh's article on fastball location, we find that for a pitcher with Santana's fastball speed (roughly 92 MPH), there isn't much of a difference in effectiveness for most parts of the strike zone against same-handed batters. In the graph above, the zones with a deeper orange color are the places where there is a difference and where it is worst to throw a fastball. As we see, though, the percentage of pitches Johan threw there last year to this year hasn't changed much at all (a 2 percent increase).

Unfortunately, I haven't seen or conducted any studies on fastball location versus opposite handed batters, but this is something I should be doing soon. For now, you can click here to see the charts, but I won't try to draw any information from them. The same, unfortunately, goes for change-ups, but this is something I'll be looking into in the coming weeks.

Usage

Overall, the usage of his pitches is relatively unchanged from 2007 to 2008. He's shifted usage from the four-seam fastball to the two-seamer a bit (understandable given the improvements made to it, which shows up in runs100), but the fastball/slider/change-up usage remains pretty much unchanged overall and to either batter hand. Santana is the prototypical pitcher relying on a fastball to both batters, the slider to lefties and the change-up to righties.

If we look deeper into his usage, though, we can see that he isn't using his pitches exactly the same. Here are links to Santana's relative usage charts by situation (i.e. hitter's count, pitcher's counts, etc.). I know some readers are intimidated by these, so only click if you're interested:

2007 Relative Usage vs. LHB
2008 Relative Usage vs. LHB
2007 Relative Usage vs. RHB
2008 Relative Usage vs. RHB

Note: All percentages I mention in this section, unless otherwise noted, are relative to how the pitcher uses the pitch overall versus the particular batter type. A 50 percent increase does not mean that a 20 percent used pitch is now being used 70 percent of the time; it means it's now being used 30 percent of the time (50 percent of 20 is 10; 20 plus 10 is 30). Also, counts were classified using the numbers from Tangotiger at The Book Blog. Please remember that we are ignoring some context here, so none of the conclusions we draw are absolutely definitive.


Ask almost any scout what Santana's best pitch is, and they'll say it's his change-up. True, it has gotten a little worse this year, but it's still a very good pitch (runs100 has actually improved this year). While his overall usage has remained the same against right-handers, he isn't leveraging it's use very well.

He is using it relatively less frequently than he was last year in two-strike counts (60 percent to 31 percent). In these counts, it would make the most sense to use the change-up more often to get the strikeout. He's also using it relatively less than last year in pitcher's counts (40 percent to 15 percent). It would make sense to use the change-up more here as the batter is closer to getting out and is protecting. I think this could have a lot to do with his struggles.

Furthermore, if we look at his improved sinker (which Santana actually believes is his best pitch), we see the same trends we do with the change-up against righties: less in two-strike and pitcher's counts.

Against lefties, he's using the change an astounding 132 percent more in two-strike counts (putting the actual usage at 21 percent) and 71 percent more in pitcher's counts. He doesn't use it much against lefties overall, but it is a good pitch and runs more than six inches in on them, so if he's set on using it, using it at the most critical juncture is probably as good a time as any. He's also using his sinker more frequently in these counts.

This (in addition to using the four-seamer more in pitchers counts), however, has come at the expense of his slider usage. Last year, in two-strike and pitcher's counts versus lefties, it was the pitch that saw the highest relative increase of any pitch in any count, and this year he's barely using it more at all despite the increased movement and better location.

Against both lefties and righties, he seems to be using his four-seam fastballs a little bit less in hitter's count. Since hitter's are generally expecting a fastball here, this probably has a positive effect, keeping them off balance.

Ballpark

One more potential reason for the decreased strikeout rate is ballpark. David Gassko found in this article that ballparks do have a noticeable affect on strikeouts, likely due to humidity or related factors. Minnesota has one of the highest park factors, inflating strikeouts by 7 percent. Shea deflates strikeouts by 1 percent.

Applying these factors directly to Santana's 2007 K/9, we'd expect it to fall from 9.7 to 9.0. This could explain some of the drop this year, although the move to the National League would have a positive effect and might cancel a good portion of this out.

Release Point

As you can see, Santana's release point is definitely different in 2008. He is releasing a lot of balls from a lower vertical release point and a few higher. That inconsistent release point could be a driving factor behind the increased walk rate. His 2.5 BB/9 is the highest it's been since 2003, back when the Twins only used to him start 18 out of his 45 appearances and he was just beginning to blossom.

Santana, at age 29, isn't an old man yet, but he's passing his prime and has thrown nearly 1,500 major league innings, and it's very possible that this kind of workload is taking a toll on Santana's arm. Hopefully Chris Neault will be able to shed some more light on Santana from a physical standpoint in the coming days, but for now I think, at the very least, this inconsistency is causing him to be just a little wilder than he was last year.

Summarizing Santana

Summarizing, here are the changes that Santana has made this year (voluntarily or otherwise):

(-) His strikeout rate is down this year and his ERA is somewhat lucky
(+) He's managed to keep his BABIP at his usual low despite the decreased K/9
(+) His slider has improved movement, and he is (most likely) locating it better
(-) He is using the slider less frequently in two-strike and pitcher's counts against LHB
(+) His sinker has improved and he is using it more frequently (in place of four-seamers)
(+) He's using the sinker more against lefties in two-strike and pitcher's counts
(-) He's using the sinker less against righties in two-strike and pitcher's counts
(-) Some of his four-seam fastballs aren't getting much horizontal movement
(-) Some of his change-ups aren't getting much horizontal movement (but it has a better runs100 value)
(+) He's added 1.5 MPH in speed difference between the fastball and change-up
(-) He is using the change-up less frequently in two-strike and pitcher's counts against RHB
(-) Shea should deflate his Ks some, but the move to the NL should prop it back up some
(-) Inconsistent release point could be causing increased BB/9

Overall, things aren't looking as bad for Santana as one might think by looking at his K/9, which is down two points from last year. Two of his pitches are improved, and the other two aren't that much worse.

Some of his trouble seems to stem from usage by situation and location (this might not actually be a problem, but it's possible he isn't locating the fastball or change-up as well as he could be; we just don't know what to look for yet). These types of things are easier to correct than pitch speed or movement. We need to be a little wary of Santana going forward, but the outlook isn't as bad as I'd feared.