January 16, 2012

Prospectus Roundtable

Can Anyone Close?

by Baseball Prospectus

Printer-friendly

The following is an edited transcript of an in-house discussion that took place among the Baseball Prospectus staff when one of our number solicited examples of unsupportable baseball arguments for an upcoming article. After Kevin proposed "Anyone can close," the thread took off in a new direction.

Kevin Goldstein: "Anyone can close." While overstated, there is a human element to the game. These are not robots, nor are they strat cards you can just flop into any role or situation and expect the same results.

Gary Huckabay: Yes, and has anyone ever claimed otherwise?

Jonathan Bernhardt: If only they were; then I could just wait for the balance patch that would make Baltimore playable.

Rany Jazayerli: Wait, what?

I used to think that there were pitchers who were capable of being competent relievers, but who were incapable of handling the ninth inning. And then Kyle Farnsworth made me look like a fool this season. It's possible there are otherwise fine relievers who can't handle the stress of the closer's job, but I'll need some proof—or at least a list of their names—before I'll believe it again.

JB: The only pitcher I've ever heard of or seen who actually seemed to be affected by the Weight of the Closer's Mantle was one LaTroy Hawkins.

John Perrotto: When I was first covering the Pirates, I saw Jeff Robinson go from being perhaps the best set-up man in the National League in 1988 to completely melting down in 1989 when he was forced into the closer's role when Jim Gott blew out his elbow on Opening Day. There are some pitchers who can't handle closing. I've seen it happen with my own eyes and been around them. Not everyone can close, even though it seems like they should.

JB: Well, it makes sense that there's a feedback loop of some kind; if you're convinced closing is extremely important, and everyone around you is convinced it's extremely important, and you're praised or cursed as if your results in that role are extremely important, then psychologically speaking, it really doesn't matter if it's actually important at all.

KG: Also, Matt Thornton sends his warmest regards.

David Pease: Maybe there are people that can't handle the pressure, and maybe the average major-league reliever can handle the pressure.

Bradley Ankrom: I don't think we're anywhere close to being able to analyze players psychologically. I know some clubs have made advances in this area, but fans/analysts attempting to draw conclusions from limited circumstantial evidence seems foolish.

Rebecca Glass: The hullabaloo over Zack Greinke comes to mind. And yes, I'm replying just so I can use the word "hullabaloo."

Ben Lindbergh: The case for Thornton not having a closer mentality is based on, what, four consecutive bad outings from last April, when he blew a couple of saves and took a couple of losses? A total of four innings? His career OPS in the 9th inning is .666, vs. .660 overall. In save situations, it's .649, vs. .664 in non-save situations. Not saying there's no such thing as a closer mentality, or that Thornton is definitely cut out for the job, but can we really conclude anything about him either way from such a small sample? 

After those four bad games, his ERA was 7.71. Over the rest of the season, it was 2.95. It's possible that in some alternate-history version of the season in which he was allowed to keep closing all year, he would've been a complete disaster the whole way through. It's also possible that he would've made exactly the same turnaround had he been left alone. That's a lot easier for me to say with nothing riding on the answer than it would've been for the White Sox with the season on the line, so I don't blame them for not waiting around to find out when they had Sergio Santos available. (Apologies if I'm overlooking anything about this specific example--I haven't spent a lot of time studying Matt Thornton.)

Relievers are an unpredictable bunch, and a lot of them go from good to bad quickly without a change in role. Inevitably, some of them also go from good to bad quickly with a change in role. When sudden struggles and a role change coincide, it's easy to draw a connection between the two, but at least some of the time, it's bound to be noise. Short of a failed closer admitting that he couldn't perform because of the pressure, I'm not sure how we can ever know for sure when it's not.

Ken Funck: Totally agree with you about Thornton, Ben. I'll cop to believing that there are guys who are destined to struggle in the closer role because of some ninth-inning mental block, but it's a condition that's significantly over-diagnosed and usually treatable over time.

What Mitch Williams said on Clubhouse Confidential last week was fascinating, in that he believes it's easier to get guys out in the ninth than in the eighth once you realize the pressure is on the hitter, not on the pitcher—you don't even need to throw strikes, since batters are more likely to get themselves out. Not that Mitch could have thrown strikes in the eighth inning, either...

Jason Collette: Put Rafael Betancourt into that discussion as well, who has shown the skills for the role forever, but a handful of outings in Cleveland ruined that momentum.

As for Matt Thornton, he was struck by the Great Pumpkin on 4/8/11, and it was all downhill from there.

KG: Of course there is noise in the data, and it's quite possible that Thornton would have calmed down and figured out. It's possible that he still might at some point down the road. I personally think he deserves another shot, but I don't think April was just purely some random thing.

In some ways I'm playing devil's advocate. I was enjoying this thread quite a bit, and then I feared it was devolving into a trend I feel like I'm seeing that disturbs me. I don't know if it's the Moneyball movie or the HOF voting or general award season or whatever, but I feel like for many (and I'm talking generally, on the interwebs, not our team), the arguments have reverted to early-generation tomfoolery. Many have gone backwards to the argument that EVERY answer is in the data, and if you don't see that or agree with that, you are a caveman noob moron. The other side is just as guilty with Saber Boy accusations and get out of your basement and watch the game stupidity.

The data is mind-numbingly important. There are still plenty of breakthroughs to come from the data, and I'm convinced that many of them will come from people on this distribution list, but don't get trapped in the data. Don't just reject out of hand circumstantial evidence that cannot be proven with the data. That stuff is important too, and we can all learn from it. We can all get upset when someone like Jon Heyman shows a remarkable ability to break a story, yet at times shows a profound lack of understanding of the game. The easy reaction is to make fun and call him a moron and joke with all your friends in a corner. It's also an opportunity to educate, and more importantly to grow. The first person who wants to be a story breaker, who understands the value of picking up a telephone and creating a network of sources, AND understanding the value of sabermetrics and how to apply them... That person is going to be huge.

Sam Miller: The evolution of catcher defense measurements seems like a cautionary tale for all of us. Noise-to-signal ratio is generally massive in baseball, and each advance gets a little bit closer to the signal but not necessarily actually close, and sometimes leads to conclusions that are actually further away from being correct. Almost every argument I see seems a bit bad to me. Except "I dunno." "I dunno" is a pretty great argument.

JC: Kevin makes a great point. 2011 was my first season in clubhouses, and it gave me a new appreciation for the narratives I once mocked from my mom's basement while attempting to dehumanize the players. I still don't know how to qualify the human element with players, but I certainly have more respect for it now.

Jay Jaffe: Regarding Thornton, I thought it was pretty well-established—I'll have to dig for where I read it—that his early-season woes as closer were timed with an unsuccessful attempt to integrate a changeup into his reportoire.

JC: Data classifies 38 pitches as changeups in first 2 months and only 23 the rest of the season.

Adam Sobsey: I thought Jason Wojciechowski's ProGuestus piece did a great job in asking us to look for the balance between raw data and the inscrutable stuff that helps produce it.

JJ: Yeah, here's a blog entry on it. Can't find the news article I recalled.

KF: Which begs the question: Did Coop tell Thornton to throw more changeups, or did Thornton choose to do so himself? If it's the latter, then maybe it highlights a guy who suddenly didn't trust his stuff and felt he had to do something different to succeed in the ninth.

JJ: I disagree with the latter line of thinking and feel like it's a mistake to try to get inside a player's head, particularly in attributing "insecurity." Logic suggests that it was a mutual decision between Thornton and Cooper to see if they couldn't add a new weapon, because that stuff happens all the time; the spring is a time to try stuff out—you can always scrap it and go back to what works. Somebody should have asked him about it, though. 

BL: Yes, I saw the post Jay linked to also, but I didn't include it in my email, since I suppose you could say that changing your approach when your role changes might be a manifestation of not having a closer mentality (unless it was something that came from Cooper). If it was something that came from Cooper, at least in part, you'd think he might've advised Guillen to let him work out the kinks before losing patience. Then again, the kinks kind of have to be worked out before the season starts. That's what spring training is for. 

Like Kevin, I was uncomfortable with a few of the earlier suggestions in this thread. Obviously, if a team decides a player isn't cut out for closing and removes him from the role, that lends some credence to the notion that he might not have the right mentality, but I'm not sure exactly how heavily to weight it. Team employees have more information and more access, but they're subject to the same biases as anyone else. Unless Thornton was spotted crying and screaming "I wanna be a setup man!" in a dark corner of the clubhouse, I'll still reserve judgment, since I'm not sold on anyone's ability to get inside his head. After all, people with the White Sox watched Thornton up close for his first four years in Chicago and decided he had what it took to close.

Jason's guest piece was great, and his views about most of this stuff mirror mine. The hardest part, I think, is being able to write something compelling when your answer is always "I don't know." 

KG: Totally agree that it's a mistake to get into a player's head, as it can lead to dangerous assumptions. I'm just saying the assumption that nothing going on in a player's head affects his play can be equally dangerous.

Jeremy Greenhouse: Ben, I obviously agree with your sentiments here. I remember watching all of Thornton's appearances at the beginning of the season and thinking how unlucky he was. I'm going through the video again.

Thornton's first appearance is in a 7-1 blowout. 0.2 innings scoreless. Disregard.

Second appearance, CHW 7, KC 6. Overpowers LHB Alex Gordon for a K. Gets 0-2 on Billly Butler, just misses a K, and then Butler sharply singles through the right side. Pinch runner steals second as A.J. Pierzynski drops the pitch. Thornton gets 1-2 on Kila, throws a low-and-outside fastball that hits its spot, but Kila reaches and serves it the other way for a double. Tie game. Thornton Gets Escobar, Pena, Getz, Aviles, and Cabrera to ground out. 

Third appearance, CHW 7, TB 4. Thornton gets 1-2 on RHB Elliot Johnson, jams him up and in, but Johnson bloops it to right, and the RF plays it for a single. Gets 1-2 on RHB Felipe Lopez, makes several hittable pitches that Lopez fouls off but eventually gets the K on a 10-pitch AB. Got 0-1 on Sam Fuld, but do not have video of the remaining pitches  in the AB. Fuld reached on an error by Alexei Ramirez. Thornton gets 1-2 on LHB Johnny Damon, throws a slider low and away that Damon reaches for and bloops to left. Juan Pierre drops it. 1-1 on RHB B.J. Upton, breaks his bat, but Upton gets a clean single. First pitch to LHB Dan Johnson is over the heart of the plate, and Johnson homers. Sox lose 9-7.

Here is a summary:

"Of the two botched 9th innings recorded by Thornton in past few days, the amount that can be blamed on him is negligible.  In Kansas City on Wednesday, Thornton induced weak contact from Kila Ka'aihue reaching for a pitch off the outside corner, and was rewarded with a game-tying double.  It stunk, but the Sox won, so people get over those sort of things. 

On Friday night facing Rays with a 3-run lead, after allowing a leadoff single, Thornton induced a groundout from Sam Fuld that the second-best defensive shortstop in baseball short-armed for an error. For the next batter, Thornton got a lazy fly ball out of Johnny Damon, only to have the second-best defensive shortstop in baseball miscommunicate with the third-best defensive left fielder in baseball, resulting in said elite-fielding left fielder letting the ball bounce in and out of his glove.  Even after all this, Thornton recovered to break Justin Upton's bat with a hard fastball, and witnessed it drop over the infield for an RBI single."

Fourth appearance, CHW 1, OAK 0. Gets 0-2 on RHB Adam LaRoche. LaRoche hammers a slider for a double. 2-0 on RHB Coco Crisp, Crisp gets his hands inside on a fastball and hits a nice line drive that is caught by the 1B. 2-0 on LHB Daric Barton, another hittable fastball that Barton hits to deep left and Juan Pierre drops on the warning track. Tie game. Yanked.

Fifth appearance, comes on with bases loaded, one out, CHW 4, OAK 2. Overpowers LHB Ryan Sweeney for a K. First pitch to Cliff Pennington jams him up and in, but Pennington fists it over the 2B's head for a two-run single. Tie game. 1-2 on DeJesus, groundout. Comes out for the next inning, Ellis grounds out, and then Thornton loses it. Five-pitch walk to RHB Conor Jackson, five-pitch walk to RHB Josh Willingham, 3-1 on RHB Coco Crisp, Crisp lines a fastball over the middle up the middle for a single. 2-2 on Barton, Barton singles.

Personally, I think Thornton was pretty awesome for most of it. He pitched like a LHB who throws 95-MPH fastballs for strikes. Most of the time, that's going to get the job done. There were a couple of times I wanted to question his pitch mix, but you need such a large sample of pitches to be able to say that a pitcher is pitching sub-optimally rather than randomizing his distribution. The in-play pitches to LaRoche and Johnson deserve some analysis. There were three crucial errors made behind him, Pierzynski being his catcher, a couple jam shots, and one hard-hit ball that got caught. His two walks actually occurred after he had blown all four saves. That's all I got.

BL: I noticed that Thornton had a .417 BABIP over his first five appearances, but I didn't do the work to see what kind of high BABIP it was. Jeremy did.

Derek Carty: I think it's pretty obvious that the "human element" exists in baseball. The problem is, no one really knows what's going on inside a player's head. Sometimes, the player himself may not even be able to say for sure how external factors are impacting his performance. And even if we believe something is affecting him, what extent of his performance is attributable to this and what is mere random variation?

Yes, the human element exists, and it's fine if a writer or analysts wants to say that something like ninth-inning mentality exists (it likely does, to some extent), but don't try to pretend like you know that a certain player has it or doesn't have it unless you have a really good reason for doing so (an official for a team, for example, might have such a reason, whereas a MSM writer might not). Don't say that Matt Thornton doesn't have the mentality because you saw him throw four innings as closer in the beginning of 2011 and that's your entire basis for your opinion. It seems a lot of times MSM analysts who don't understand stats and don't care to learn about them use this "human element" stuff as a crutch. Because they don't have something truly meaningful to say, they'll spew some overstated crap about the human element, maybe attempting to support it by using some meaningless, small sample size data or split.

Even worse is when former players-turned-analysts talk about their own experiences playing the game—their memories of which can sometimes become warped over time—and generalize that experience to other players. Everyone has their own unique psychology, and just because you were a player yourself doesn't necessarily mean you have any idea what's going on inside the head of another player. 

I think the other thing that goes overlooked a lot is that MLB players, as a group, are inherently very good at dealing with pressure situations. They've been facing pressure their entire lives, and part of the reason why they are MLB players to begin with is because they dealt with that pressure effectively enough to succeed in high school and college and up through the minors to the point where they're even being considered for, say, a ninth-inning role. Writers like to say that this guy won't succeed in this role for X-Y-Z asinine reasons, but unless you have a truly legitimate reason for thinking that, odds are that the player will succeed because his track record of dealing with pressure, as a human being, is really, really good.

Tommy Bennett: I think the heart of this argument is about how much weight to give to various pieces of evidence. Everyone agrees that several years of success as a set-up man are good evidence that a pitcher is ready to take on closing duties. Everyone agrees that a terrible debut as a closer provides at least some evidence to suggest that the pitcher is less well-suited to be a closer than you would have thought before the lousy debut. Information from teams and knowledge of particular pitcher's makeup give a similar indication. 

So the question is how much you weight one against the other. In general, I think it's better to rely on the former type of evidence, since it's the one that's been proven to have the most straightforward predictive effect. That does not mean there aren't people out there who can pick out—before the fact—the pitchers who lack a closer mentality. But if those guys are out there, what they are able to do is very different from observing after the fact that a guy failed when thrust into the closer role and assuming from there that he lacked the closer mentality. And if those guys really do exist, I'm willing to bet they're gainfully employed by major-league teams and not sharing their secrets.