The Mets Starting Staff and Injuries

by Eno Sarris on Mar 2, 2011 3:00 PM EST

After a year of "prevention and recovery" that resulted in fewer games and dollars lost to the DL and yet ended with the ace on the mend, it may seem like it's impossible to avoid the doctor these days. And recent research by Jeff Zimmerman at FanGraphs suggests that it really is, at least when it comes to starting pitchers.

Just take a look at his introductory piece for background and an introduction to his DL-predicting equation. Then check out his second piece, and one number practically leaps off the page. 39% of all starting pitchers that started last year will hit the DL this year. And once they get there the, mean stay lasts over two months. Pitching is a dangerous thing, it seems. Your shoulder and elbow are perhaps not uniquely created for the purpose of hurling a small ball over 90 MPH.

What does this mean for the Mets? Well, not only has Zimmerman done the work to show us the likelihood of starting-pitcher-related disaster in the coming year, but he's also provided us the tools we need to estimate young pitchers without a track record.

Here are the Mets pitchers that Zimmerman has already projected.

30.4% Mike Pelfrey

43.6% Jon Niese

47.5% R.A. Dickey

Ouch, we haven't even gotten to the question marks and we've already got a pitcher that is almost as likely to hit the DL as not. Then again, any projection system that uses pitchers as a whole may miss the mark on knuckleballers. Just as they've outshone their peripherals in the past, knuckleballers may be able to stay healthier than this system projects. We know that some knuckleballers have joked about being to pitch every day.

Niese, on the other hand, may seem high for a young pitcher, but he's already made two DL trips in his short career, and he did not go to college. Those are two marks against him at his age that help move the needle north from 39.1%.

For the rest of the rotation, we'll have to do our own work. Here's the formula.

1/(1+e^(-z))

where:
z = (.2209)(Years with Trips to DL)+(-0.0040)(GS in last 3 year)+(0.0509)(Age in previous season)-1.7692

That produces some ugly numbers, some of the highest in the game:

62% Chris Capuano

63% Chris Young

Yeah, so maybe we get one full pitcher-year from these two pitchers.

For the rookies, we have to use his third piece, where he linked college status, age and BMI. That puts our rookie, and let's call him a rookie because he's still eligible for ROY, in the following bracket. (I used this link for BMI.)

30.3% Dillon Gee

Thanks to stat guru Geoffrey Young, I can also tell you that there is only a 2.02% chance that the starting five stay healthy. There's only a 11.9% chance that only one hits the DL! There's an 86.1% chance that two (or more!) are injured this year. It's a virtual lock that we'll have a patchwork rotation this season.

So this is why starting pitching depth matters. This is why we may even see Jenrry Mejia starting games in Citi Field this year - and not because we necessarily think he's ready. And if you're sorry to see Dillon Gee fail to make the major league team, fret not. He will make some starts in Citi Field this year, and the Gee Unit will be in effect.

Tweet Comment 21 comments | 0 recs |

Do you like this story?

Comments

so?

Halladay 43.5, Lee 42.8, Blanton 40.4, Hamels 31.3

Don’t see Oswalt on their but I doubt he’s lower than 40.

And our offense is better. Maybe things aren’t really that bad.

-Ceetar, the Optimistic Mets Fan

by Ceetar on Mar 2, 2011 3:10 PM EST reply actions 1 recs

What he's saying is that we can make the post-season

if we pull a Tonya Harding on the Phillies?

by brooklynlou on Mar 2, 2011 3:30 PM EST up reply actions

we don't have to

age/luck will likely take care of it without having to pull a Harding. The question would be, if you could take out one Phillies pitcher, who should you take out? Do you aim for Hamels because he’s the most likely to be healthy or do you just do Halladay because Hamels could very well be average and you’d best take out the best pitcher, even if he might get hurt anyway?

-Ceetar, the Optimistic Mets Fan

by Ceetar on Mar 2, 2011 3:35 PM EST up reply actions

Like in Boondock Saints, when they bust into the basement poker game:

All of them.

by psuwxman on Mar 2, 2011 5:22 PM EST up reply actions

So this is why starting pitching depth matters.

Where is Figgy when we (eventually) need him?
When our pitchers break down, are they really calling up Mejia before September? Even if he starts the year dominating AAA, he still won’t have many IP logged in the minors. It would be better to use our “depth” (Gee, Misch, Carrasco) and let Mejia develop until he is ready.

and not because we necessarily think he’s ready

In lobby for: Jaime Cevallos, Zack Lutz, orange unis
The Unwritten Rules of AA

by Michkin on Mar 2, 2011 3:22 PM EST reply actions

Though Santana will fill the shoes of one of the disabled

by nerfan on Mar 2, 2011 3:29 PM EST reply actions

40%+

Santana was up there at 40something too. I’m not real sure what that’s measuring though. Is it 60% off, or is he that likely to get hurt again in 10- starts?

-Ceetar, the Optimistic Mets Fan

by Ceetar on Mar 2, 2011 3:33 PM EST up reply actions

You really just wrote this whole thing

so that you could use “Gee Unit will be in effect” didn’t you?

Much less scientifically then you’ve just done I was already resigning myself to seeing Young, Capuano and Niese miss some time this season anyway so I guess it’s nice to see some science applied to those depressing thoughts.

by Jay_What on Mar 2, 2011 4:00 PM EST reply actions

actually

I think it was so I could write that caption.

www.RotoHardball.com
www.AmazinAvenue.com

by Eno Sarris on Mar 3, 2011 7:52 AM EST up reply actions

Just thought you'd like to know

that Jeff Francouer, with 1 out and a runner (Hosmer) on 2B in the top of the 4th, just pulled a grounder to deep short and was thrown out failing to advance the runner. They’re replaying this game tonight on MLB Network for those who want to bask in the glory.

by Jay_What on Mar 2, 2011 4:07 PM EST reply actions

One thing that really annoys me about baseball statistical analysis

is that it’s hard to take any analysis like this seriously without knowing the std errors on those coefficients or the goodness of fit. I’m willing to believe that p-values will be low, but I’d really like to know at least the likelihood ratio over the mean.

Not that I have huge problems with this analysis, just a peeve.

by robotoverlord on Mar 2, 2011 4:42 PM EST reply actions

I know that's a legitimate concern.

Really I do. But all those words make my brain freeze and I think of things like this.

by bjk3047 on Mar 2, 2011 5:04 PM EST up reply actions 3 recs

I know it's just sports

but relying on stats without knowing what they mean is worse than not using stats at all.

Example:

(-0.0040)(GS in last 3 year)

That’s close to 0. It’s probably because it’s a conflicting signal. It’s reasonable to posit that a history of health indicates general healthy behavior, but we could also say that 3 30 game start years in a row is too many and is an indicator for an overuse injury. A one-way test of direction would be interesting. Does it mean we should try to give as many starts as we can to Niese so that he’ll be healthy next year?

by robotoverlord on Mar 2, 2011 6:02 PM EST up reply actions

I plan on poring through the data more thoroughly later

But my first guess is that it doesn’t really take mechanics in mind. I don’t see how Dickey is an injury risk comparable to the other players. The knuckleball is a relatively stress free pitch.

He’s an extreme case, naturally, but there are some people who just don’t get hurt and others that can’t help but get hurt. I’m also surprised to see Halladay’s risk listed so high.

by TheBigStapler on Mar 2, 2011 5:02 PM EST reply actions

Outliers.

There’s very little accounting for them. Dickey’s health is like Ichiro’s success despite a >55 GB%.

Save Jenrry Mejia!

by Ogre39666 on Mar 2, 2011 5:12 PM EST up reply actions

it's just a projection

combined with the caution you need to show regarding health of pitchers.

I think it’s best used as a tool that reminds you how volatile pitchers’ arms are. The numbers aren’t based on the individual, just the trend of injuries observed over the past 9 seasons.

You can’t account for injuries. This is a reminder of how important depth is.

Kicking knowledge in the face.

by BlackOps on Mar 2, 2011 5:29 PM EST up reply actions

I agree completely, and I would take it farther

I think mechanical loads give you a better prediction than a statistical analysis ever could. With pitchers in general, you deal with really small, noisy sample sizes, that can be full of bad data. Pitchers on their way to the DL can lead to very fickle data. There are years when pitchers just skip a start before the all star break instead of a DL stint, and there are years when pitchers pitch terribly through injuries. There are also pitchers who go to the DL because they are “so bad they must be injured” (see Perez, Oliver).

This is the reason that general predictive rules don’t withstand statistical rigor (think the Verducci Rule that is only true until you look at the data).

Once you account for what is sure to be awful, awful input data, you also have to account for the fact that biomechanics matters. Everyone agrees on that essentially, they just disagree on how much. I would be interested in a statistical analysis of injuries on a team that we know uses biomechanical load analysis (Boston is the only one I know for sure of) and a team that doesn’t (I’m pretty sure we didn’t last year).

Finally, I know that the injury guy at Baseball Prospectus put together systems like this a few years ago, and they were often pretty hit or miss. His guesses mid-season, based on visible mechanical flaws, however, were significantly stronger. I imagine his statistical analysis influenced those guesses. Some combination is the best system, I’m sure, but what balance?

In short, I think the projection for Dickey is close to useless (as I do for performance projections for him).

by djg2111 on Mar 2, 2011 6:43 PM EST up reply actions

Based on those projections

It won’t only be Gee and Meijia. We could also see Cohoon and Misch in the mix at certain points. Yikes. I think Dickey will be fine though, knock on wood…

by David G on Mar 2, 2011 6:44 PM EST reply actions

The representation is skewed in that

its too shallow in the given field of data because it gives absolutely no indication of DL duration applied per given percentile. Now we know there’s given increments of time but you can’t discern from top to bottom the significance of the duration regardless of how you look at it. I can understand the relevance with a classical probability approach but it makes no sense to try to apply the data with an expectation to somehow live up to an empirical set of raw data.

"Intelligence is not a genetic predisposition. Think stupid!!"

by Wright of passage on Mar 2, 2011 8:26 PM EST reply actions

D'oh!!

I didn’t see the link for the 2nd piece. The calculations are correct and the odds are significantly stacked against our rotation’s collective health. Hot dog anyone?

"Intelligence is not a genetic predisposition. Think stupid!!"

by Wright of passage on Mar 2, 2011 8:54 PM EST up reply actions

“That produces some ugly numbers, some of the highest in the game:
62% Chris Capuano
63% Chris Young
Yeah, so maybe we get one full pitcher-year from these two pitchers.”

With the history of these two guys, it can’t be that there’s, respectively, a 38 and 37% shot that they’ll be productive over 30+ starts. I wonder what the percentage is of starters with their histories who get sent down or to the pen or are released due to ineffectiveness.

by Brian Singer on Mar 4, 2011 12:54 AM EST reply actions