---

Posted by Derek Carty at 1:46pm

Monday, July 28, 2008

Player spotlight: Manny Corpas (PITCHf/x)

With the trade deadline looming, a deal that would have one of the biggest fantasy impacts would be one involving Brian Fuentes. If Fuentes, the closer for the Colorado Rockies, left, Manny Corpas.

Corpas started the year as closer but struggled mightily, and it wasn't all bad luck. His skills took a serious downturn from 2007 and he lost the job. This put Fuentes into the role and established Taylor Buchholz as next in line. On June 1, though, Corpas struck out two batters and walked none in two innings pitched. From there, he went on a serious run to leapfrog Buchholz on the depth chart. Last week, Rockies general manager Dan O'Dowd said that in the event of a Fuentes trade, Corpas would resume closing games.

Because Fuentes is one of the closers most likely to be traded over the next week, the quality of his potential replacement is of great interest to fantasy owners. Let's first look at Corpas' numbers and then check out his PITCHf/x data.

Numbers

Surface numbers

+--------------+-----+----+------+------+------+---+----+----+---------+
| YEAR         | AGE | G  | IP   | ERA  | WHIP | W | SV | BS | SV%     |
+--------------+-----+----+------+------+------+---+----+----+---------+
| 2006         |  23 | 35 | 32.3 | 3.62 | 1.36 | 1 |  0 |  2 |    0.00 |
| 2007         |  24 | 78 | 78.0 | 2.08 | 1.06 | 4 | 19 |  3 |   86.36 |
| 08 March-May |  25 | 27 | 27.3 | 6.59 | 1.83 | 0 |  4 |  5 |   44.44 |
| 08 June-July |  25 | 19 | 21.7 | 2.91 | 1.02 | 1 |  0 |  1 |    0.00 | 
+--------------+-----+----+------+------+------+---+----+----+---------+

Peripherals

+--------------+-----+----+------+------+------+-------+-----------+------------+
| YEAR         | AGE | G  | IP   | K/9  | BB/9 | xGB%* | LIPS ERA* | DIPS WHIP* |
+--------------+-----+----+------+------+------+-------+-----------+------------+
| 2006         |  23 | 35 | 32.3 | 7.52 | 2.23 | 45.45 |      4.31 |       1.27 |
| 2007         |  24 | 78 | 78.0 | 6.69 | 2.31 | 53.81 |      3.97 |       1.19 |
| 08 March-May |  25 | 27 | 27.3 | 4.61 | 5.27 | 56.38 |           |            |
| 08 June-July |  25 | 19 | 21.7 | 8.31 | 0.42 | 48.44 |           |            |
+--------------+-----+----+------+------+------+-------+-----------+------------+

Note 1: xGB% is used for 2006 and 2007, but actual GB% is used for both 2008 periods.
Note 2: LIPS ERA and DIPS WHIP aren't given for the 2008 periods because they would be a bit of a pain to calculate and it's pretty obvious looking at his K/9 and BB/9 that he has been 8,000 times better in June and July.

Luck indicators

+--------------+-----+----+------+-------+-------+-------+
| YEAR         | AGE | G  | IP   | BABIP | LOB%  | HR/FB |
+--------------+-----+----+------+-------+-------+-------+
| 2006         |  23 | 35 | 32.3 | 0.344 | 78.95 |  8.82 |
| 2007         |  24 | 78 | 78.0 | 0.263 | 84.86 |  9.52 |
| 08 March-May |  25 | 27 | 27.3 | 0.323 | 65.32 | 17.39 |
| 08 June-July |  25 | 19 | 21.7 | 0.303 | 72.82 |  5.56 |
+--------------+-----+-----------+-------+-------+-------+

We see a huge difference between Corpas from March to May and from June to now. He walked more batters than he struck out early in the year, yet he's walked just one since June 1. He's also struck out a ton more. If Fuentes is traded and he continues to pitch like this, he should have no problem holding down the job for the rest of the year and providing tons of fantasy value.

If we look at his 2007, though, his LIPS ERA was just below 4.00. His skills were decent, but they weren't outstanding for a closer. They were much closer to being borderline than elite. I would be comfortable owning Corpas for the remainder of the year given those skills, but there might not be much leeway for him to struggle or even get unlucky.

All this being said, let's examine Corpas's PITCHf/x data in all three time periods and see if we can figure out which Manny Corpas we'll see in the second half of 2008.

PITCHf/x

Note: I'll be using a few graphs I've never used before, so if you have suggestions for making them easier to understand or more visually appealing, please let me know.

Looking at Corpas' PITCHf/x data, I can't figure out why he's been so successful since June 1. When he was removed from the closer's role, everyone seemed to know it was because of his slider.

Rotoworld said that "Corpas will move into a setup role and work on improving his slider." Corpas himself said, "I am going to look at some video. Obviously, my slider isn't working as well. It's spinning." Yahoo!'s RotoArcade said on July 2 that "since June 1, he's rediscovered his slider, notching respectable 3.50 ERA and 9.0 K/9."

What's weird is, despite his excellent numbers of late, Corpas hasn't actually rediscovered his slider. I was fully expecting to see his 2007 slider and his June-July slider match up, but that's not the case at all. In fact, his June-July slider is actually a tiny bit worse than his March-May slider. Check out the movement on it:





2007 - Speed/movement

+------+-----+-------+------------+------------+
| TYPE | %   | SPEED | MOVEMENT X | MOVEMENT Z |
+------+-----+-------+------------+------------+
|   SK |  78 |  93.8 |      -7.16 |       5.44 |
|   SL |  22 |  80.5 |       3.13 |       6.83 |
+------+-----+-------+------------+------------+

2008 - March to May: Speed/movement

+------+-----+-------+------------+------------+
| TYPE | %   | SPEED | MOVEMENT X | MOVEMENT Z |
+------+-----+-------+------------+------------+
|   SK |  72 |  92.0 |      -7.00 |       4.78 |
|   SL |  28 |  80.8 |       0.71 |       4.19 |
+------+-----+-------+------------+------------+

2008 - June to July: Speed/movement

+------+-----+-------+------------+------------+
| TYPE | %   | SPEED | MOVEMENT X | MOVEMENT Z |
+------+-----+-------+------------+------------+
|   SK |  61 |  92.3 |      -7.91 |       3.90 |
|   SL |  39 |  80.5 |       0.56 |       6.07 |
+------+-----+-------+------------+------------+

As you can clearly see, Corpas's sinker seems to have improved a little over the last month and a half. It's getting nearly an inch more sink than the beginning of the year and an inch and a half more than last year. It's also getting close to an inch more horizontal movement. It has, however, lost 1.5 mph.

The slider, however, is perhaps the worst it's been since the PITCHf/x cameras were turned on. His horizontal movement is slightly worse than earlier in the year and nowhere near where it was in 2007. He's also getting less downward break on it, though he was succeeding in 2007 with some "rise" on the slider. I'm not sure if he does this intentionally, but sliders generally break down, not rise up. Let me know if you know anything about this.

This loss of slider movement is especially bad for Corpas. Josh Kalk found that the difference in horizontal movement between a pitcher's slider and fastball is one of the most important components of success. Because Corpas is strictly a fastball-slider pitcher, this loss of movement figures to hit him hard (though the added movement on the sinker helps).

Another thing I found strange: His sinker is what's improved here while his slider really hasn't gotten much better. Despite this, he is using the slider 11 percent more in June and July. It seems that it would make more sense to use the sinker more until he can figure out what's wrong with the slider. Yet the results, as we saw above, have been fantastic. What gives? Let's look at some more things to try and figure it out.

Pitch usage by count

So if his slider still isn't fixed and he's using it more often overall, maybe he's leveraging when he's using it. The following charts break down how much more frequently he uses each pitch by count type in comparison to how frequently he uses them overall. An explanation follows the charts.






Explanation of the charts: Think of these numbers as percentage over normal. Let's say, as an example, a pitcher uses a fastball 10 percent of the time overall but 20 percent of the time in hitter's counts. That means he is using the pitch twice as much in hitter's counts, a 100 percent increase. While the raw increase is 10 percentage points (20 minus 10), the chart will say 100 percent for comparative purposes. Let's say in the next period, the pitcher uses the fastball 20 percent of the time overall but 40 percent of the time in hitter's counts. He is still using it twice as often (100 percent), even though the percentage point increase is 20 percent. Comparing that raw 10 percent to 20 percent makes it seem as though he is using the fastball more often in hitter's counts in year two, when he is using them exactly the same percentage more, in relative terms. I hope this gives you an idea of how to read these charts.

As far as what is classified as a hitter's count, pitcher's count, and neutral count, things get a little tricky. Tango at The Book Blog ran some numbers and found how counts should be classified as, but this this article by Joe Sheehan shows that pitchers see things a little differently. I used Joe's classifications because I think it serves our purposes here better. One key to pitching is throwing off a hitter's timing and rhythm, his expectations. If Corpas is to succeed with a subpar slider, he will need to do what the batter isn't expecting. These expectations aren't necessarily grounded in reality; they have been formed by facing real life pitchers to get a general sense of how they pitch that likely matches pretty closely to Joe's findings.

If you feel differently, please let me know why you think I should do things differently in future analyses.

I've also included all two-strike and all three-ball counts, which should be pretty self-explanatory and should give us some information on how Corpas is using his pitches when a strikeout or walk is one pitch away.

Please also note that breaking pitch selection down by count can be useful, but there is also some context that we are ignoring, so take this section for what it's worth. The sample sizes we're looking at aren't exactly enormous, either.


So, if we know that Corpas' slider isn't as good as it used to be but he's using it more frequently, it figures that the least amount of damage would be done if he throws it more when the hitter is least expecting it and less when the hitter is expecting it. This would mean using it more in hitter's counts and less in pitcher's counts.

We see that he is using the slider a little more frequently in hitter's counts in June and July than he was in March through May but still much less than he does overall. If you're going to use a below-average pitch, this seems like the optimal time to be using it, maybe even more often than you usually do (although it is a hitter's count for a reason; they can relax and take more chances, and it's important to get the pitch over the plate).

On neutral counts, he isn't deviating much from his overall percentages.

On pitcher's counts, he is using the slider only slightly more than he does overall, less than he was in March through May, and much less than he was last year. Still, he's using it more than he does overall, which doesn't make a whole lot of sense to me. The hitter is probably expecting a breaking pitch, plus his sinker is so much better. The quality of the sinker combined with the hitter not necessarily looking for it seems like it would make it all the more effective. Also, on a pitcher's count, the batter is more likely to swing since he's on the ropes, and while it might only be a protective swing, Corpas should probably want fewer swings on the slider since it really isn't moving much.

Now let's examine the two-strike and three-ball counts.

Given the state of his slider and the filthiness of his sinker, we should expect Corpas to use the sinker as his primary strikeout pitch on two-strike counts. Looking at the charts, we see that he is using the slider a little bit less frequently on two-strike counts than he was previously. Still, he is using the slider relatively more often than the fastball on these counts, and because he throws the fastball just 61 percent to begin with, that 12 percent decrease means he's using it just 54 percent of the time on two-strike counts. It should be much higher than that; these pitches are not equal by any means.

On three-ball counts, we generally see pitchers use their fastball the most because it is easiest to control. Corpas has actually started using the fastball relatively less than he was from March to May and the slider relatively more, although we're looking at very small sample sizes (just 18 pitches in June and July), so we probably can't draw too many conclusions here.

Overall, these changes should probably have a net positive effect on Corpas' results, but they aren't drastic changes, could be better, and certainly don't explain the incredible swing he's experienced. Let's keep going and see what else we can find.

Location

Next, I thought he might be locating his pitches differently. Below you'll find the location charts for his sinker.

Note: The below charts break the rulebook strike zone down into nine sections. Balls out of the zone are ignored both in the graphic and in the percentages given. The green sections don't represent an actual zone, but rather simply show the combined percentage of pitches thrown on the inner third of the plate.







We saw in John Walsh's terrific article "How fast should a fastball be" that fastball speed doesn't matter a whole lot except when pitching inside. Faster fastballs are more effective than slower ones on the inner third of the plate and even more effective when thrown down and in. Corpas' sinker (92.3 mph) is pretty fast and would certainly benefit from proper location.

If Corpas' location was bettering his results, we'd expect to see more sinkers thrown inside and more thrown inside and low. Surprisingly, we see that (to right-handed batters), he threw the most inside pitches in the March to May period and the lowest percentage from June to July.

The same can be said when looking strictly at the low, inside corner. He threw an astounding 50 percent of pitches there from March through May. It's possible that he was actually throwing there too often and batters were expecting it. Of his pitches to righties, 67 percent were sinkers, and half of those were thrown to that one zone (when thrown in the strike zone). That's a full one-third of pitches going to that one zone. It's possible he was simply overdoing it, though I'm not entirely convinced.

When looking at the left-handed hitters, it's interesting to note that he very rarely goes inside on them with the sinker. I first found this incredibly strange, but when we examine the movement of his sinker, it makes a little more sense. His sinker has a lot of horizontal movement (more than seven inches in all three periods), breaking away from a left hander. He'd have to be aiming at the batter himself to get the pitch to fall in the inner corner, so it makes sense that he'd be afraid to do this all that often.

It seems, though, that he might have been aiming for the inside part of the plate and seeing it instead go down the middle more frequently from March to May. Fifty percent went somewhere in the middle three zones compared to 43 percent in 2007 and 41 percent from June to July. Keeping the ball away from the middle of the zone to lefties could be contributing to his success, though the changes aren't all that extreme. With the excellent horizontal movement and added sink though, he's probably getting a lot of swings and misses on the outside corner (the zone he's now throwing to the most) as the ball tails into the corner or out of the zone.

I don't have any data that say where it is best to locate a slider, so we can't analyze it too deeply, but it seems to me that a slider down the middle is bad news, especially when it isn't getting much horizontal movement. The second column of the following table shows the percentage of sliders that Corpas threw in the middle zone (see graph above), and the third column shows the percentage of sliders thrown within four inches of the middle zone (on any side).

+-----------+---------+------------+
| YEAR      | MIDDLE% | MIDDLE+4%  |
+-----------+---------+------------+
| 2007      |       7 |         34 |
| March-May |      12 |         41 |
| June-July |      15 |         55 |
+-----------+---------+------------+

The trend is clearly upward; Corpas is throwing more sliders into the heart of the zone. It's perplexing how he has enjoyed so much success with his strikeouts, although a lowered walk rate would make sense if he's just feeding the middle of the zone. If you're interested in seeing the full slider location graphs, you can see them here, here, and here.

As far as the strikeouts are concerned, I thought he might be throwing a lot on the edges of the plate to make up for the ones thrown in the middle. In the following table, I define edge as one-sixth the width (or height) of the rulebook strike zone on either side of the rulebook boundary. The "edge percentage" is defined as the number of edge pitches divided by the number of edge pitches plus all non-edge in-zone pitches.

+-----------+-------+
| YEAR      | EDGE% |
+-----------+-------+
| 2007      |    28 |
| March-May |    40 |
| June-July |    21 |
+-----------+-------+

No dice. Not only is he throwing more sliders into the middle of the zone, he's throwing fewer onto the edges (which I guess makes sense considering how many are thrown to the middle, but still).

Thoughts on pitching in general and the use of PITCHf/x analysis

When examining PITCHf/x data for a pitcher, it seems that there are two types of things we can analyze. The first deals with the pitcher's physical abilities, like the speed and movement of his pitches. These remain relatively static, and while they can improve or worsen, we generally don't see frequent, wide swings. The second deals with things like pitch selection, location, sequencing, etc. These are based more on mentality and judgment (and context) and are much easier for a pitcher to change, sometimes without him even realizing or without really having a choice.

It seems to me that unless you can talk to the pitcher himself and he speaks freely about his game plan and about his judgment skills, they would be difficult to predict. We can see what they have been in the past, but it seems that they are less likely to remain the same going forward as a pitcher's physical skills are.

Banking on physical skills is probably the better bet unless the pitcher clearly pitches a particular way (such as the willingness of a pitcher, maybe Greg Maddux, to throw any pitch in any count), although as we saw with Corpas, these physical skills aren't impervious to falloffs, even for a pitcher reaching his physical prime.

Final thoughts on Corpas

A portion of Corpas' physical skills (his slider's movement, the speed of his sinker) has declined this year. His overall pitch selection (a judgment skill) has also been made less efficient (i.e., using the slider more), though he has done a little bit with his selection by count (a judgment skill) to help compensate. It looks as though the location of his slider (a judgment skill) has worsened as well, although I did little more than a superficial analysis. The location of his sinker was a bit perplexing as it looked best from March to May, though it could have been too predictable for hitters.

Overall, it seems that Corpas could very possibly be in for a poor second half. Because a portion of his most stable skills—his physical skills—has worsened, and because his judgment skills don't seem to really qualify this turnaround, a falloff is a legitimate threat, especially if his judgment skills fluctuate for the worse.

Of course, he could certainly rediscover his slider at some point. I also didn't look at all factors, and I could have ignored other judgment skills. (I tried to look at a lot of things, but this article has run very long. If there's anything else you'd like me to look at, feel free to let me know.) Conversely, it is also possible that his strikeout and walk rates were lucky over the past couple of months, or were a little unlucky over the first couple. Maybe it was a batter quality issue or something like that; it's not like we're looking at very large samples.

My guess would be that it's a combination of both. As it stands, I couldn't find a whole lot to explain those excellent June-July stats, though it is hard to argue with them. However, he still has that excellent sinker working for him (which has actually improved), and his slider is still very well disguised in his sinker. Plus there's a 12-mph gap between the pitches (though that isn't as important as it is with curve balls or change-ups).

I'd still pick up Corpas if he is unowned, because the chance he'll get to close, combined with the chance he'll actually pitch well, combined with the chance he'll pitch poorly but either get lucky or not blow games is too great to ignore. Saves are saves, and Corpas might be the best speculative pickup in baseball. Just be aware that the team has another worthy candidate in Buchholz and that Corpas could very well struggle.