Brett Anderson’s Elbow Eaten by Sliders?
Back in April, we took a look at every pitcher who had thrown the slider more than 40% of the time during the past three years. The results weren’t pretty. The “Slider Group” was more likely to have major surgery on their arm or on their shoulder than the general pitching population. They also hit the DL more than average. To say the least, the pitch seems stressful.
Since the slider cutoff was so high, and the innings minimum non-existent, our 25-pitcher sample was made up entirely of relievers. Now with Brett Anderson reporting some general elbow soreness and a hefty slider percentage (41.1%), it seems like a natural time to ask the follow-up question: Do we see the same propensity for injury, both major and minor, among starters who favor the slidepiece? Did that pitch gobble up Anderson’s elbow ligaments like they were so many miniature hamburgers?
Though some commenters favored using raw pitch numbers to compare pitchers, that approach has pitfalls of its’ own. If we counted pitches, we might run into a sample that eliminated the unhealthiest pitchers. And yet their use of the slider might still have had something to do with that failing health. So instead, we’ll continue using percentages but we’ll throw an innings limit in there in order to try to hone in on starters. Since we’re mostly trying to compare Anderson to his peers, we’ll use 200 innings (since the beginning of the 2009 season) as a cutoff. It’s random, but it helps us find comparable starters who have pitched over the same time frame and used the slider about as much.
Our Starter Slider Group:
| Name | FB% | FBv | SL% | SLv | DL Trips |
|---|---|---|---|---|---|
| Ervin Santana | 58.50% | 92.4 | 36.70% | 82.3 | 2 |
| Armando Galarraga | 52.70% | 90.5 | 36.60% | 85.8 | 0 |
| Bud Norris | 55.00% | 93.4 | 34.70% | 87 | 1 |
| Ryan Dempster | 53.90% | 90.7 | 34.50% | 84.8 | 1 |
| Brett Anderson | 50.90% | 92.1 | 33.80% | 82.9 | 2 |
| Francisco Liriano | 52.10% | 92.7 | 30.40% | 85.9 | 1 |
| Edwin Jackson | 61.00% | 94.4 | 29.70% | 86.6 | 0 |
| Chad Gaudin | 61.70% | 90.6 | 28.80% | 79.9 | 1 |
| Felipe Paulino | 57.80% | 95.4 | 28.20% | 86.9 | 2 |
| Hiroki Kuroda | 59.20% | 92.2 | 28.00% | 84.1 | 2 |
| Johnny Cueto | 58.40% | 93 | 27.80% | 85.6 | 2 |
| Randy Wells | 55.00% | 89.5 | 27.50% | 84.1 | 1 |
| Freddy Garcia | 34.00% | 87.8 | 26.80% | 80.4 | 0 |
| Colby Lewis | 57.10% | 89.7 | 26.70% | 83.3 | 0 |
| Tommy Hanson | 56.70% | 92.3 | 26.10% | 83.4 | 0 |
| Josh Johnson | 64.40% | 94.8 | 26.00% | 87.2 | 1 |
| Gavin Floyd | 50.50% | 91.9 | 25.90% | 85.1 | 0 |
| Ross Ohlendorf | 63.80% | 91.3 | 25.60% | 80.5 | 3 |
| Brett Myers | 46.20% | 89.1 | 25.30% | 83.2 | 1 |
| Ricky Nolasco | 49.90% | 91.2 | 24.80% | 83.8 | 1 |
| Joba Chamberlain | 63.90% | 93.3 | 24.10% | 85.6 | 0 |
| Anibal Sanchez | 52.20% | 91.3 | 23.80% | 84.5 | 2 |
| Chris Carpenter | 48.90% | 92.1 | 23.40% | 87.3 | 1 |
| Brad Bergesen | 65.20% | 89.8 | 23.30% | 82.3 | 1 |
| Derek Lowe | 63.20% | 88.4 | 22.40% | 81.1 | 0 |
| Starter Slider Group | 55.69% | 91.6 | 28.04% | 84.1 | 1 |
| Control Group | 58.50% | 91.7 | 14.00% | 83.8 | 0.92 |
That the starters here throw faster (91.6 MPH fastballs to 89.9 MPH) and throw more fastballs (55.7% fastballs to 46.3%) than our reliever group isn’t extremely interesting for this discussion. They probably only throw harder because of the size of this sample, most likely. Also somewhat likely is the possibility that they are starters because their fastballs are better than the fastballs in the reliever-only group. Starters are picked to be starters, generally, because their pitching mix is more diverse and they have more effective pitches. So, since they’re more likely to have a larger arsenal, it makes sense that they would throw fewer sliders than the relievers (28% sliders to 47.3%).
But the juicy findings are in the last column. These 25 pitchers averaged one stint on the DL since the beginning of the 2009 season. Jeff Zimmerman found that the typical veteran starter was 39% likely to hit the DL in a given year. Since these pitchers have played about 2.35 seasons on average since then, they were 42.5% likely to hit the DL during a season in the sample. Not quite as an extreme a difference from the general pitching population as our relievers showed us.
Over their careers, this group has had Tommy John surgery four times. That’s 16% of the sample. Again, that number is larger than the national rate of occurrence (10.7%), but not as far gone as the 20% in the reliever sample. Lastly, 17 of these 25 pitchers have either hit the DL for an extended bout of pain in their elbow or in their shoulder. Pitchers get hurt easily, but these pitchers seem to get hurt more often.
Brett Anderson has thrown his slider 33.7% of the time. That’s the fourth-most among starters with at least 200 innings since 2009. During that period, he’s also hit the DL twice for elbow problems. These days, his velocity is down a bit and he’s complaining of soreness in the joint. The general health of the other slider-heavy pitchers on this list suggests that Anderson’s upcoming appointment with the doctor is a significant one.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
I feel like I need to know how often the remaining qualifying sample of pitchers hit the DL to make any real inference.
Pitchers at large hit the DL at 39%. You want to know how often pitchers who have pitched 200 innings since 2009 have hit the DL? I’m pretty sure it’s close to 39%. It’s not a huge IP cutoff.
How much would available data allow this to be expanded?
It’s nice and all to use Anderson as the basis of comparison, but I think it would be more useful to go further, include more years.
This is tip of the iceberg material (not in a negative way, but in a not enough way.)
Yes, we could go back as far as P f/x lets us, and then compare the high-slider, high-curveball and high-fastball groups. May still happen.
One thing about the slider percentage: I’m not sure what’s going on with the pitch classifications, but it’s one of two things:
1. He doesn’t have a slider AND a curveball, just one hybrid holy crap pitch that he can throw anywhere from the low 70’s to the mid 80’s, and it breaks anywhere from 0 to -10 (or more) vertically.
2. He has both, and the classifications are calling a lot of curveballs as sliders.
Any PitchFX gurus want to lend a hand?
I called it a slider, sure it’s something crazy. Our own P F/x guru handled it pretty well:
http://www.fangraphs.com/blogs/index.php/a-rare-bright-spot-for-the-as/
Thanks Dave Allen!
I don’t know if it means anything but 3 guys who have pitched at least a little for the White Sox (Jackson, Floyd, Garcia) all have avoided the DL. I only bring it up because of other articles showing the White Sox overall good health.
I have nothing to add,other than calling for a moratorium on “slidepiece”…
Awesome.
Slidepiece … that is classic baseball vernacular.
I like it, but DL trips seems a bit cumbersome as a measurement of injury. I wonder if the data’s there to use missed starts (since we are talking about starters). Sm (starts missed) would increment by 1 every time a start was missed for any reason, assuming the pitcher was in the regular starting rotation of his team or teams in that time frame. This would measure with a bit more precision how much time a pitcher has missed time during the season – not all injuries requiring DL time are equal, and in many cases a starter who’s hurting would have a start skipped to give him extra rest.
Thinking about this again, we could just measure starts and compare to league average (or your control group) # of starts. Bottom line – how much value is lost through extensive use of the slider when compared to other secondary pitches.
As a Twins fan, my only complaint with this study involves Francisco Liriano’s TJ surgery and how this data ignores it. He threw the slider 37.6% of the time before getting hurt in 2006, and hasn’t thrown it that often since.
Well, Liriano’s TJ is counted in the TJ sentence, and his slider % still put him on this list. I wouldn’t say we ignored it, it just didn’t happen within the two-plus years we were focusing on.
Missed that. Thanks for pointing it out.
As an extension on Bryz’s point…Logically, I’m wondering if the result is higher based on the guys throwing the huge % of sliders not being previously damaged arms. By this I mean that the only guys who would throw 25% sliders since 2009 are guys who didn’t already have their arms torn up by sliders. That puts a group of guys with a serious history of arm trouble all out of our “sliders group” and into our control group. All things being equal, I think we can agree that more guys with previous arm trouble = more DL trips are likely from 2009 to now. The fact that the sliders group overcame that and still got injured more probably means more than just a .08 difference in the raw numbers.
So solved by upping the sample and reaching back further, right?
If you push the sample itself back further, you get a different set of “slider throwers” all together so I’m not sure if thats an option.
Or possibly try to quantify how many DL trips (or however you want to quantify it) the slider group had PRE-2009 compared to the control group. I’m thinking it will be LESS than the control group which makes the results to the orginal study more interesting (because the guys with less injury history are getting hurt more than the guys with injury history). The hypothesis being that these guys aren’t slider throwers who keep getting hurt and keep throwing sliders, they’re guys who just haven’t been hurt yet so they throw whatever is getting outs. I’m not sure though. I suspect you’d have to somehow figure out the change in slider percentage before and after significant injury to know what is really happening on that part of it.
Correct me if I’m wrong, but I believe JoJo should be listed with 2 DL trips in that timeframe.
Makes sense. I’ve heard that the Atlanta Braves don’t allow their pitchers to throw sliders until they’re at least in AA. Something that I think would be interesting to add to this. Maybe multiple regression study or something, would be body type. You always hear of the 6’5″ 220 pound kid out of college as a future horse. I’m guessing evidence suggests a bigge frame means more innings and it makes sense. However, I’d like to see how much.
Maybe look at all pitchers who throw a lot of sliders and then look at size and maybe even age.
The difference may not be significant based on the relatively small sample. Can you do a chi-square p-value on this? How many pitchers are in your control sample?
At first glance I read the subject line as “Brett Anderson’s Elbow Eaten by Spiders” and freaked out a little.
Did anybody else misread the title of the article as “Brett Anderson’s Elbow Eaten by Spiders”? I was really curious there for a second…
That a trick question??
Does it not matter what injury forces a starter to the DL? I fail to see how an leg injury or a back injury that results in a DL stint provides any answers to a question about whether throwing a lot of sliders is bad for your elbow.
With such a small sample size, if even a few of the DL trips among the Slider Group were caused by an injury other than to the pitcher’s arm then it would have an impact on the results, wouldn’t it?
It would… but, in fairness, the DL trips measured in our control group aren’t limited to arm injuries either. Maybe Zimm and I can bang this one out together.
This was my thought as well… it adds a little subjectiveness but if “non-arm injuries” make up a significant portion of the DL trips it may make it difiicult to see if a correlation exists
Just out of curiosity–and not to be contrary, for sure–how much did Randy Johnson rely on his slider vs. DL stints, early on, then later in his career?
You should find out and tell us.
“One area where there is a big difference is in pitch selection. Overall, Johnson threw 51.4 percent fastballs last year, compared to 48.8 percent this season. With his slider it is 40.4 percent this year compared to 35.2 percent in 2008. But if we look at what Johnson was doing at the end of last season, we get an even bigger difference. These numbers come from Dan Brooks’ wonderful site.”
http://www.fangraphs.com/fantasy/index.php/comparing-livan-hernandez-and-randy-johnson/
Randy threw 40% sliders late in his career. Perhaps he has a high slider frequency because he realizes that there is no need to preserve his arm because he realizes that he is likely to retire soon (after 2009).
The 3 things, regarding slider usage, I would be interested in:
[1] Total # of sliders thrown, not SL %.
[2] Slider usage as game goes on? (2nd, 3rd, 4th time through lineup).
[3] Slider usage in consecutive years.
———————————–
[1] Be interesting to see if there is sort of a “do not cross this line” in regards to numbers of sliders thrown.
[2] I can easily imagine that as the game goes on, slider pitchers are more reliant on it. This is only important in that you’re throwing a pitch that is more stressful on the elbow, at a point when you are more fatigued.
[3] It seems to me that once guys get good at throwing the slider, they want to throw it more often (Wainwright for example). It’s a highly effective pitch, and one that is likely very easy to become reliant on.
————————————
I see someone mentioned Randy Johnson and throwing more sliders late in his career. I would say that it’s no coincidence that his fastball velocity was significantly decreased, and his slider was all he had. RJ was throwing the slider that much because, for the first time in his established career, he was in “survival mode”. It’s also possible that Johnson has “optimal leverages” for the pitch.
Just one question here.
Were there really TEN other CircleChange’s already on Fangraphs?
While I understand and agree that throwing a slider puts additional stress on the elbow (anyone who has ever pitched even at high school level can attest to that), this article doesn’t delve into the reasons behind each DL stint. Does the DL data in this article ONLY pertain to shoulder and/or elbow problems? Are there guys on this list who hit the DL because they rolled their ankle, strained their oblique (like Brandon Beachy), or some other non-throwing related injury? Further more, is there corollary evidence to suggest that a shoulder injury can be the result of throwing too many sliders compared to other pitches? Annecdotally, I don’t see how there could be. All in all, I think this post just scratches the surface of research, and all-in-all, there just isn’t enough of a sample size to unequivocally answer the question: Do pitchers hit the DL more frequently if they throw more sliders?
Just a few days too early to catch another big DL trip for a name at the bottom of your list. Joba increases the TJ pitchers to 20%, although I’m not sure that he really fits into your starter group.
Would it be possible to change DL trips to number of days on DL?
I’m late to this discussion, but after reading all the replies, I have these questions/statements:
1) Do the DL trips in this sample count leg injuries? If so, then the data is flawed.
2) I have heard/read that it’s scientifically proven that the slider puts more torque on the UCL than other pitches based on the exaggerated twisting action of the forearm. Can we show a slider to TJ surgery comparison?
3) DL trips happen to everyone, serious injury (surgery) is what we should be concerned about.
4) DL trips and time missed for arm “issues” can vary from team to team and by the age of the player. Younger players are going to be treated more cautiously than veterans for soreness. The only true measure if sliders are more damaging is by using surgeries instead of DL stints or time missed.