Fluke watch: Jhoulys Chacin

Last year, at age 22 in his rookie season, Jhoulys Chacin pitched 137.1 pretty good innings to a line of a 3.28/3.54/3.62 ERA/FIP/xFIP. Those are all pretty solid numbers, though his pitching at Coors Field might give a fantasy player a little worry that he is unlikely to perform to his peripherals in the future—or, as in the case of 2010, to put up a better ERA than should be expected.

Now, this year, Chacin’s numbers have changed in some very interesting ways. His ERA is near identical to last year (3.33), but his xFIP has dropped from a fairly solid 3.62 to a really good 3.17 (his FIP has risen thanks to some bad home run luck). More interesting is how this change has occurred: Chacin’s strikeout rate has declined, but the decline has been compensated for by a decrease in his walk rate and a large increase in his ground ball rate. Whereas previously Chacin relied on his strikeout ability (along with an average ground ball rate) to get outs, he has increasingly been able to rely upon the ground ball this year to get outs.

And for a pitcher in Colorado, that’s a pretty good thing to be able to rely on.

But the question is: Is this sudden groundball ability REAL? Or is it just a mirage, a fluke of a small sample size? Lets look at his pitches in search of an answer.

Chacin’s pitches:

Chacin appears to have five pitches: a four-seam fastball that has a lot of cut, a sinking fastball (two-seam probably) without much tailing action, a change-up, a curveball, and a slider. The movement and velocity of these two pitches each of the last two years can be seen below*:

Pitch Type Year Pitch Velocity (MPH) Horizontal Movement Vertical Movement
4-Seam 2010 91.4 -0.64 +6.20
4-Seam 2011 91.0 -0.07 +7.20
Sinker 2010 90.2 -4.54 +3.79
Sinker 2011 89.6 -4.25 +5.19
Slider 2010 81.2 +6.16 -1.62
Slider 2011 81.3 +6.62 -0.36
Curve 2010 78.3 +7.27 -5.81
Curve 2011 78.3 +6.66 -5.16
Change-Up 2010 83.1 -3.18 +3.07
Change-Up 2011 83.1 -2.75 +4.70

*The two fastballs are not super-clearly distinguishable via PITCHf/x and the two breaking balls (the curve and the slider) aren’t either. However, I’m pretty sure that my classifications are for the most part correct.

As you can see, there have been barely any change in the movement or velocity of any of Chacin’s five pitches. Really, all of the changes in velocity/movement you see in the chart above are within the PITCHf/x system’s margin of error.

This suggests, of course, that the change in Chacin’s results is probably not caused by a change in the movement of his pitches.

The mysterious increase in ground balls

So the question is: Where has the increase in ground balls been coming from? The ground balls appear to be coming from Chacin’s two fastballs, especially against left-handed batters. Check out the GB rates on Chacin’s fastballs each of the last two years:
Against lefties:
Four-seamer in 2010: 43.6%
Four-seamer in 2011: 71.9%
Sinker in 2010: 44.4%
Sinker in 2011: 67.5%
Against righties:
Four-seamer in 2010: 40.7%
Four-seamer in 2011: 61.8%
Sinker in 2010: 57.14%
Sinker in 2011: 52.38%

As you can see, the groundball rates have gone up pretty dramatically on these pitches to lefties, and they’ve gone up on the four-seamer to right-handed batters as well. Now the fact that Chacin, a right-handed pitcher, might get a decent GB rate against left-handed batters (and a better rate than against right-handed batter), is not surprising. Chacin uses a fastball with a lot of cut, which tends to result in reverse groundball splits—the pitcher gets good GBs against opposite-handed batters (lefties in this case). In addition, Chacin’s pitches to left-handed batters (more than 50 percent of the batters he faces), are near always aimed on the outside part of the plate, the best part of the plate to get ground balls. So a decent GB rate on these pitches against left handers (even on the “sinker,” which doesn’t have the same cut) shouldn’t be totally surprising.

But at the same time there’s nothing in the way these pitches have been thrown against left-handed batters to explain why the GB rate should have increased so much this year. The pitches are being located in the same part of the strike zone, have basically the same movement, and the distribution of these pitches has basically remained unchanged. While perhaps last year’s GB numbers were lower than we might have expected, the increase we see this year is a bit much for us to simply explain it as regression from bad luck last year.

In fact, it appears that the groundball rate increase this year against left-handed batters is simply caused by good luck. Last year on the four-seamers, Chacin would give up non-ground-balls on pitches in the middle to inside part of the plate. This year, despite mostlyseeing his pitches in the same area, batters have not put these middle-of-the-plate to inside pitches into play at all. I would suspect this trend not to continue.

Now there are several possible explanations for the GB increase. First, Chacin is locating more of his pitches within the strike zone this year. Perhaps this is resulting in an overall increase in GB rates? I’m not sure why it would, and the GB rate change appears to be universal, on pitches both in and out of the strike zone. Second, Chacin has slightly changed his pitch usage from last year by throwing more curves and fewer change-ups. Perhaps this is affecting batters’ swinging habits so that they hit the fastballs more into the ground? Once again, I’m not sure why that would have this effect either.

Finally, the four-seamers do seem to have about a half inch more cut on them than they did last year. However, as noted above, this change is basically within the PITCHf/x system’s margin for error, and I hesitate to claim it as real and the cause for any change in results.

Against right-handed batters, it appears that there is one additional change: The four-seam fastballs are being aimed a little bit more outside than last year, which would make us expect a few more ground balls (we don’t see this change in the sinkers, whose gbroundball rates have remained basically constant this year from last year). So perhaps the GB rate increase is sustainable against these batters.

Improved control

There is one change to Chacin’s pitching that is briefly mentioned above: He is throwing more pitches in the strike zone than last year against both right and left-handed batters. This is an easy explanation for how Chacin has managed to reduce his walk rate this year. So this improvement by Chacin should be sustainable (as long as he can keep up the greater “accuracy.”)

Conclusion

Chacin has been very impressive this year. However, at least some of his improvements this year are bound to regress, particularly the improvement of his GB rate against left-handed batters. It wouldn’t be surprising for him to end up with a groundball rate higher than last year’s against such batters, but it should be closer to last year’s rate than this year’s really great rate. Against right-handed batters, his GB rate may also regress.

Chacin’s improved walk rate should remain the same, but signs also point to his decreased strikeout rate continuing as well (presumably because, with fewer pitches being out of the strike zone, batters are less likely to miss when they swing).

The result is that Chacin’s peripherals should get worse as the year goes on, and I doubt he’ll continue to put up better peripherals and overall numbers than last year for the rest of the season.


26 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Jeffrey Gross
12 years ago

Good article Josh. Plugging Chacin’s numbers into my xWHIP calculator, his normalized ROS numbers spit out a 3.56 eFIP, 1.28-1.32 WHIP. This assumes a 60% GB% heading forward, however.

Do you think Chacin can maintain a GB% north of 50%? Is he a sell high candidate? 10% SwStr% says better K days could be ahead….

garik16
12 years ago

Jeff, I think a GB rate north of 50% is certainly possible….like I said, his pitches are such that his GB numbers from last year were probably a little low.

His current career GB rate is 51.1%, which is probably a good estimate. 

Also, I’m not really sure a resurgence in Ks is ahead….his whiff rate is DOWN this year, despite him throwing more curves to right-handed batters.

Jeffrey Gross
12 years ago

Josh,

Do you have his SwStr% by pitch? My research shows that SwStr% is the majority of the battle with K% and pitchers with SwStr% around 10% tend to post K9 rates in the mid-upper 8s or higher

garik16
12 years ago

Jeff, not going per pitch since the distribution of pitches isn’t the same, but here’s my overall numbers:

SwStrike by Handedness and Year:
Lefties:  8.216% in 2010, 7.235% in 2011
Righties: 12.970% in 2010, 12.897% in 2011

So it’s only a small drop, but it does suggest a drop in K rates from last year.  Whether it suggests as big a drop as currently…that’s another question.  Keep in mind he faces roughly 50/50 Lefties vs Righties.

Jeffrey Gross
12 years ago

Perhaps thats a better way to phrase it. The K rate might be down in expectations from last year, but it should be above its current level.

In terms of elasticity, a +1.19 percent change in a pitcher’s swinging strike rate generally sees a corollary +1.0 percent increase in K/9. Accordingly, if we assume that last year’s K/9 rate was Chacin’s true talent line, then his expected K/9 rate should be ~8.3% lower this year, which would be approximately 8.29.

Jeff Free
12 years ago

LHB v RHB splits for any pitcher at Coors are far less relevant than home v road splits. Coors is death to breaking balls because of the thin air. OTOH – Coors is, relatively, a FB (and esp changeup) friendly environment – for the same “thin air” reason—if you can keep hits on the ground. Pitchers – smart ones – change their pitch distribution because of that.

Chacin has pitched 48 innings at Coors vs only 25 innings on the road. 2010 was 75 IP at Coors and 63 on the road. That is the stat that will regress as the season goes on. His rotation slot will likely get more road starts than home starts from here on out. And thus he will also begin changing his distribution of pitches.

I’d be interested in seeing data—but I suspect that Chacin has simply now figured out the vagaries of pitching at Coors. And no regression is in order.

garik16
12 years ago

Sorry Jeff, that’s not the explanation:

Home/Road Whiff Rate Splits:
2010:
Road: 28.6% Whiff Rate
Home: 24.3% Whiff Rate
2011
Road: 22.2% Whiff Rate
Home: 23.5% Whiff Rate.

The drop in whiffs has come ON THE ROAD, not at home.

JK
12 years ago

I’m sorry, but this whole article is based upon the assumption that his numbers last year were the norm, not the exception, and thus his gb% should regress back to them.  In his four minor league seasons he consistently posted gb rates of 60%+, and this was one of the reasons he was considered a good prospect.  It really seems that what has happened is his gb rate has regressed up to his career (both minor and major) norm, not the other way around.  Which is understandable given he’s 23 and transitioning into the majors.

LMV
12 years ago

JK, totally agree.

garik16
12 years ago

@JK- What we’re dealing with here are two very different sample sizes – a big one from last year and a small one from this year.  When there is little difference in the samples, aside from the results, in an area, it’s logical to trust the larger sample size as being more indicative of true talent. 

Pitchers can develop, and yes, last year’s GB rate seems perhaps low given his pitch location.  Which is why I’d say a GB rate over 50 is possible. 

But when there’s extremely little changes in his pitching from last year, it’s much more likely that his GB rate being above 60 is a fluke of luck than development. 

Development of a pitcher leaves signs, in movement, velocity, and location of pitches, for example.  There’s no sign here.

Donald Trump
12 years ago

So vertical movement going from 6.2 to 7.2 is withing the margin of error… Can you explain a little bit… is this a sample size problem, i.e. if you have two pitchers with long careers and lots of data who have vertical movement of 1.0 difference, can that be meaningful or is it too close?  What I am getting at is what if that difference in vertical movement is real (even if, statistically speaking, it cannot be verified)

garik16
12 years ago

Donald, changes of 1 inch or less, are not uncommon simply due to how the pitchf/x system is calibrated in parks where the pitcher has pitched.  In other words, it’s not a big enough change in a single season (nevertheless a third of a season), to conclude that it’s an actual change rather than a calibration ghost. 

Regardless, I doubt it’s relevant.  That increase in vertical movement is an increase in “RISE.”  If real, it represents less sink, rather than more. 

It would be very unlikely for that to be the cause in the increase of GBs (for more GBs, you need more sink).

Jeff Free
12 years ago

i thought the key issue was about GB% not about whiff%.

The home/road split for GB’s for Chacin is:
2011 – home – 48 IP – 60.7%
2010 – home – 75 IP – 47.8%

2011 – road – 25 IP – 56.7%
2010 – road – 63 IP – 45.0%

He has clearly increased his GB% in both home/road. But the 2011 road IP is so small as to still be completely meaningless. While his home GB% may regress a bit, he won’t be starting as many games from this point at home as he will on the road. And his road GB% doesn’t seem out of whack.

I don’t know what his minors GB% was – but it does seem to me that would probably be a more statistically valid number than a rookie year MLB number – esp a rookie year at Coors and its unique and serious pitching issues.

Also, his dropping K rate may be as simple as his first pitch strike %. It has declined from 55% last year to 52% this year. So he is falling behind in the count—and STILL walking fewer batters than last year. Does he change his pitch distribution when he falls behind in the count—to pitches that induce fewer strikeouts but which he has more control over? Possibly. But if that is the case; then his stats are likely to improve if his first strike % starts to regress towards either last year (55%) or the ML average (? – but prob higher than 52%).

IOW – it seems to me that Chacin could blow the doors off as the season goes on. Limited only by pitch count and innings cap. I think Chacin is an ace in the making.

Now maybe I’m just talking my fantasy roster. But I’ve got a keeper team with four young stud pitchers – Pineda, Chacin, Garcia, and D Hudson. Chacin is the one who is untouchable and not for sale.

garik16
12 years ago

Sorry Jeff, was dealing with Whiff rates above with another commenter. 

As it is, I’m still not understanding you.  There’s no reason why Coors Field should cause the home/road GB splits you mention.  There’s basically no difference in his pitch usage from home/road (There is a slight difference in 4-seam/2-seam breakdown, but I suspect some of that is my classifications, and it’s minor at best).  It’s not an explanation here.

Jeffrey Gross
12 years ago

@Josh,
Climates affect stuff like that

garik16
12 years ago

Jeff Gross – Yeah, but there’s no reason it would affect it like this way this year as opposed to last year.

In other words, yes climate can affect things, but there’s no reason why it should have caused this year the super high GB numbers (

Jeffrey Gross
12 years ago

Well last year, he pitched in the summer months. We’ve had an atypically cold April/May season.

garik16
12 years ago

He pitched in May last year as well, to lower GB numbers.  I’m just not seeing how climate or home/road issues would present a case for no regression.

Jeffrey Gross
12 years ago

Not saying no regression, just pointing out factors that may have influenced the numbers

Jeff Free
12 years ago

Yes. It has been very cold and wet in Denver this spring. The biggest issue though re his stats at Coors last year is that 2010 was his rookie year.

Coors is not like anywhere else. The thin air (20% less dense than sea level) messes with breaking balls. They don’t break. So either you exert 20% more force on your muscles to force the ball to break where you want it to go – risking injury. Or you throw fewer breaking balls at Coors and rely on other pitches. Or you pick a different aim point for your breaking pitches at Coors v road. That all requires a transition period where stats don’t mean as much because the pitcher is still basically learning something that they have never had to learn at any time before. Unless they’ve pitched at C Springs – and the Rockies try to keep prospects out of the Springs because it doesn’t have a humidor and every FB is a HR which messes with a young pitchers confidence.

If his pitch distribution is the same at coors v road; then the rockies are really doing something wrong. The changeup and fastball are easily the most effective pitches at Coors. Pure physics. but at any rate – if he has been pitching breaking stuff at Coors—and getting the results that he has been getting – then watch out when he starts getting the road starts at sea level.

Jeff Free
12 years ago

Just looked at Chacin’s AAA numbers. He only spent a total of 50 innings at the Springs in 2009/2010 combined. That’s what he’s pitched at Coors this year so far. So they did keep him out of there

Again I’m not sure this is relevant to anything except that I don’t think that 2010 data is meaningful simply because there is more of it. One wouldn’t simply average time trials which include a)learning how to ride a bike with b)already knowing how to ride a bike.

garik16
12 years ago

“Yes. It has been very cold and wet in Denver this spring. The biggest issue though re his stats at Coors last year is that 2010 was his rookie year.

Coors is not like anywhere else. The thin air (20% less dense than sea level) messes with breaking balls. They don’t break. So either you exert 20% more force on your muscles to force the ball to break where you want it to go – risking injury. Or you throw fewer breaking balls at Coors and rely on other pitches. Or you pick a different aim point for your breaking pitches at Coors v road”

Stop here. His Home pitch movement #s this year (in terms of movement/speed changes) are different in that he’s gotten LESS sink so far at home than he did last year on his fastballs, the big pitch in terms of his ground balls.  (Horizontal Movement is virtually identical)

In other words, once again, the home/road thing should have caused him to get LESS ground balls, not more. 

The fact that it was his “rookie year” and he should be developing doesn’t change the fact that we DO NOT SEE ANY DEVELOPMENT THAT SHOULD CAUSE THE INCREASE IN GROUND BALLS – even if we account for Coors. 
—————

“That all requires a transition period where stats don’t mean as much because the pitcher is still basically learning something that they have never had to learn at any time before. “

This is baloney.  If you want to say you don’t like my numbers, fine, just ignore them.  But the stats and movement numbers DO tell us something – that there’s nothing that should have caused the ground ball increase in terms of pitch changes. 
————————

“If his pitch distribution is the same at coors v road; then the rockies are really doing something wrong. The changeup and fastball are easily the most effective pitches at Coors. Pure physics. but at any rate – if he has been pitching breaking stuff at Coors—and getting the results that he has been getting – then watch out when he starts getting the road starts at sea level. “

You’re inventing reasons to love Chacin, here.  Chacin does get greater break on his curve (vertically) on the road: true.  But a key thing is really the velocity on the pitch being quite good (Averaging around 78+ MPH), which doesn’t change from home to road. Meanwhile, in his career so far (note small sample size), the Curve has been a better pitch for some reason at HOME than on the road.

garik16
12 years ago

“Again I’m not sure this is relevant to anything except that I don’t think that 2010 data is meaningful simply because there is more of it. One wouldn’t simply average time trials which include a)learning how to ride a bike with b)already knowing how to ride a bike. “

Jeff, Jeff, Jeff.  The point above is this:  There’s been basically NO changes that show such development from last year (and your bike learning analogy is terrible and extreme at best).  If he really was LEARNING on the job last year, we SHOULD see some change in the aim/movement/velocity!

Jeff Free
12 years ago

>>If he really was LEARNING on the job last year, we SHOULD see some change in the aim/movement/velocity!

Not necessarily. Coors is a field that extracts a high cost on mistakes. Not one where more mistakes are made. Or one where there is a particular “solution” pitch (a la most hitting parks that are quirky because of dimensions or wind).

The process of learning at Coors does not involve just pure pitching technique (aim/movement/velocity). It involves gaining the confidence in each of your pitches – in a new environment – so that you can still pitch from your entire arsenal even if there is a man on first with the game tied and the #4 at the plate.

I would argue that in those circumstances – at Coors – being able to mix up your pitches is gonna be far more likely to induce a groundball or mishit than is going to your known “groundball” pitch. Any pitch that is known/predictable to the batter can be lofted – and at Coors if it is lofted it is not a good thing (well unless of course it is the Rockies who are doing the lofting).

Is there a way that that can be analyzed? I’m sure there is. I don’t know how.

But I don’t see that a focus on particular horizontal/vertical movement of a pitch (without afaik home/road splits) is it. And, in any case, I certainly don’t think that an exclusive focus on that movement of a particular pitch in isolation is conclusive enough to assert – as if proven – that there are NO (your emphasis) changes that show such development.

Tom Wilson
12 years ago

Well the title is “Fluke Watch” so as someone who has watched Chacin’s development with interest, I was ready to consider last year a lot more of a fluke than this year.

His MiLB numbers would point to his last year being the fluke, in both k% and GB% so while 150 batters is usually the barrier for the GB% to become predictive, it seems that in this case maybe the numbers are normalizing later.