2015 Starting Pitcher DL Projections
As our own Brad Johnson stated today, one of the reasons pitchers are difficult to evaluate in their inconsistencies. One of the biggest reasons for the inconsistencies is pitcher health. While a few hitters can lose an entire season because of injury, it is just a fraction compared to pitchers who have to sit out. To help understand each pitcher’s injury risk, I will release my 2015 starting pitcher rankings.
I have been releasing the values for a few seasons with accurate results. Last season, I estimated 50 of the 128 pitchers who threw 120 innings in 2013 would end up on the DL. I was a little low since 53 starters made a least one trip. It work out to 41% or 2 out of every 5 starters. The percentage always seems to hover around 40%.
The get the values, I used the same trustworthy formula I created a few years back which looks at three items:
- Age: The older the pitcher, the more the injury risk (+1% point increase each year older)
- Injury history: Nothing predicts future injury like past injuries (+10% points for each season of the past three on the DL).
- Games Started: A pitcher needs to show they can throw for an entire season without breaking down (-3% points for each full season up to three).
Additionally besides the overall injury percentage, I have found high breaking ball usage and ability to throw strikes pointing to higher injury risks. Here are the categories noted on the spreadsheet.
- Slider% > 30%
- Curveball usuage > 25%
- Throw strikes < 60%
- Pitchf/x Zone% < 47%
All the information can be found in this Google doc. For a reference, here are the top ten most and least likely players to go on the DL
| Name | Age | DL Stints (last 3 seasons) | GS (last 3 seasons) | DL% |
| Bartolo Colon | 41 | 2 | 85 | 60.3% |
| Tim Hudson | 38 | 2 | 80 | 57.1% |
| Colby Lewis | 34 | 2 | 45 | 55.6% |
| John Lackey | 35 | 2 | 60 | 55.3% |
| Charlie Morton | 30 | 3 | 55 | 55.0% |
| J.A. Happ | 31 | 3 | 68 | 55.0% |
| Ryan Vogelsong | 36 | 2 | 82 | 54.4% |
| A.J. Burnett | 37 | 2 | 95 | 54.4% |
| Matt Garza | 30 | 3 | 69 | 53.6% |
| Clay Buchholz | 29 | 3 | 73 | 51.9% |
| Travis Wood | 27 | 0 | 89 | 32.1% |
| Wily Peralta | 25 | 0 | 69 | 31.6% |
| Lance Lynn | 27 | 0 | 95 | 31.5% |
| Wade Miley | 27 | 0 | 95 | 31.5% |
| Mike Leake | 26 | 0 | 94 | 30.5% |
| Jose Quintana | 25 | 0 | 87 | 30.1% |
| Shelby Miller | 23 | 0 | 63 | 29.9% |
| Julio Teheran | 23 | 0 | 64 | 29.8% |
| Rick Porcello | 25 | 0 | 91 | 29.7% |
| Madison Bumgarner | 24 | 0 | 96 | 28.3% |
Old, injured guys on the top … healthy established kids on the bottom.
Predicting exactly which starting pitcher will get injured in nearly impossible, but we can understand the pitcher’s injury chance. An owner just has to understand the risks involved with pitcher and make their own evaluations.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
this is fuckin awesome
Should postseason GS not be included? Would the extra innings on the arms sway your projections? (Yes, MadBum I’m looking in your direction…)
And nope, that would have nothing to do with your calculation. Keep up the great work.
Here is my look at MadBum and the extra work load: http://www.fangraphs.com/blogs/did-bumgarner-and-shields-throw-too-many-pitches/
one of your rules should be that Buchholz has an automatic inclusion on the DL list.
looks great except Shelby hard to believe
Do you think adding in length of DL stay to the formula and/or type of injury would be helpful?
I did add days in the initial testing and they didn’t matter.
I have thought of adding a few, but I still want to stay simple.
Rather than using length of DL stint in your formula, do you think you could turn it into a dependent variable? I wonder if these same variables correlate well with predicted length of DL stint.
As soon as I saw Slider usage greater than 30%, I immediately though Tyson Ross gotta be around this post somewhere.
Wow I didn’t realize Porcello is YOUNGER than MadBum
Porcello is 6 months older…
Wainwright would probably be #1 on my list. Surprised he’s only 37% due to hit the DL
Shocked to see no Scott Kazmir. But looking at the numbers, it seems like it’s due to him missing 2012 and somehow not hitting the DL in 2013/14, and him all of a sudden learning how to throw strikes. I would say he’s a solid bet to hit the DL next year though. Higher than 50% to me.
Did you do any validation from years past? What percent of the variation in DL% is explained by your model?
In particular, how has it done in the years outside of the model development years of 2006-2009?
I used the past seasons (2002 to 2009) to set it up. It has just held up since then, so I haven’t felt I need to change it.
The problem is results is a % on an either or conclusion, but have grouped the results in the past:
The 25 players most likely to end up on the DL values ranged from 43.5% (Roy Halladay) to 55.1% (Daisuke Matsuzaka). Of the 25 players, 12 went on the DL in 2011, or 48%. The average percentage chance predicted that 12.2 players or 49% would make the DL. The model held pretty good. This group of players had an average age of 31 years old, pitched in only 65 games over the past 3 season and went on the DL 1.75 times over that time frame.
The pitchers with less of chance to end up the DL ranged in value from 27.3% (Clayton Kershaw) to 32.7% (Paul Maholm). Of the 25 pitchers, 9 went on the DL in 2011, or 36%. The average percentage chance predicted that 7.75 pitchers or 31% would end up on the DL. The final prediction for these pitchers was not as good as the DL prone pitchers, but close. The pitcher’s average age was 25 years old, pitched in 79 games in the previous 3 season and has never been on the DL.
http://www.fangraphs.com/blogs/revisiting-2011-sp-dl-projections/
First, I removed pitchers who were not on a team in 2013 like Kevin Millwood and Randy Wolf. Next, I added the DL percentage chances for the remaining pitchers. The formula predicted 44.8 pitchers were DL bound. In reality, 44 went on the DL. I will take it. Additionally, I looked at the 20 most and least likely players to go on the DL. Of the twenty least likely, I predicted 6.2 to go on the DL and the actual number was five. Looking at the most likely candidates, I predicted 9.9 to go on the DL and nine actually went. Overall, the results were outstanding and I will put out the 2014 prediction at a later date.
http://www.fangraphs.com/fantasy/mash-report-111913-starting-pitcher-dl-projections-reviewed/
Awesome work; thanks for the followup. That’s exactly what I was curious about!
Looks like Cliff Lee is not on your google doc. Curious about him.
He didn’t hit the min. 120 IP in 2014 from which I based the formula on. You can plug his numbers into the linked equation above and get a decent idea of his injury risk.
Jeff, does percentage IP increase from previous year factor into your evaluation?
i.e., if a pitcher threw 100 IP in 2013 and 200 IP in 2014, is he a slam dunk higher risk than a pitcher whose thrown 150IP –> 200 IP over the same time?
(Apologies if this has already been answered)
No, I don’t and from some recent research, inning jumps don’t point to future injuries.
I thought younger pitchers, under 24 anyways, had as high an injury risk as older pitchers
“Games Started: A pitcher needs to show they can throw for an entire season without breaking down (-3% points for each full season up to three).”
Might be a bit of survivor bias there that excludes younger pitchers
Also, the innings limit of 120 IP seems a bit odd if the low number is due to injury. Example, Pineda, CC, and Nova all did not meet 120 IP limit due to injury
Is there some debate over Jordan Zimmermann’s injury potential this season, or am I overreacting?