THE BOOK--Playing The Percentages In Baseball

Actually, Greinke is the best spokesperson because he’s actually good.

He’s just awesome. I can’t wait to see what kind of pitching coach he’ll make. I bet he’ll be a future Dave Duncan, and the first Brian Bannister.

There’s a difference between understanding DIPS and understanding how to be an effective major league pitcher.  It’s great that Bannister understands the former, but I’m not sure how much it helps him pitch well.  Thus, I don’t really see how he’s a sabermetric hero.  It’s a nice skill to have in a GM or your director of baseball ops, but if I’m looking for a pitcher or a pitching coach, it would be well down the list.

Morgan Ensberg seems much more like a player who understood how to use sabermetric-style thinking to help him do his job better.

I’d love to sit down and talk with Bannister for a few hours. 

Regarding Tango’s point about whether giving up long HR’s is “mostly the domain of bat pitchers”,
my approach to analyzing this issue would not to look at HR production or HR length as some indicator of bad pitching.  Rather, I would look at hitf/x data as one indicator of good/bad pitching, just as I would look at it as one indicator of good/bad hitting.  As Bannister himself indicated, there is certainly luck involved with pitching (and the same for hitting).  Better to look at outcome-independent metrics, such as how hard the ball was hit (batted ball speed) and vertical launch angle (line drives, fly balls, etc.) as indicators of a good hit.  Minimizing those good hits would be an indicator of good pitching.  Looking at the frequency of good hits would not be the only thing I would look at, but it certainly would be one of them.  Looking at how good hits are correlated with pitch location, movement, etc. are other things I would look at.

If I were advising an MLB club interesting in purchasing hitf/x data, that is the argument I would give them.

Alan, no disagreement. 

I should clarify that I was using the “true distance” on Greg’s site, not the actual distance.  So, IIRC, that would include the speed off the bat and inferred launch angle in that calculation.

Ideally, Greg would have all long-FB, not just those that ended as HR.

Greg’s “true distance” is the distance extrapolated to ground level.  He also has something called “standard distance”, which is the distance the ball would have traveled (to ground level) under some standard atmospheric conditions (sea level, standard temperature and pressure). 

If TrackMan were implemented in all stadiums and if the data were publicly available, then we would have the desired FB landing points.

But I am arguing that you really don’t need that information if you are trying to evaluate hitting (and, by extension, pitching).  You only need the launch conditions.  There is already enough data in the public domain (from April 2009) to correlate the launch conditions with outcome and therefore to establish what constitutes a well-hit ball.

Alan: thanks for the correction on true v standard.  Standard I think is what I would want.

Yes, definitely, all you need is the launch conditions to regress against something.  The launch condition will let you figure out distance, hang time, and spray angle, and that’s all you need to know as the chance a ball is a hit, extrabase hit or out, and how many runs that’s worth.

Ideally, you’d have a run value equation based on the launch parameters.

He says it while wearing a uniform and performing really poorly.

Best spokesperson?

I’m not the world’s biggest fan of xFIP, but his career xFIP is 4.85.  So he’s not horrible, but he’s not very good, either.

What struck me while doing a recent analysis of Dan Haren is how much Haren’s stuff is like Bannister’s.  They’re not identical pitchers by any means, but Haren’s made a heck of a lot more out of his 90-91 mph fastball than Bannister has made out of his 89-90 mph fastball.  Haren’s curveball is not very good.  So it’s not like he has physical tools that Bannister doesn’t have access to. 

Haren has learned a cutter in the last few years that has become a dominant pitch for him.  He’s got a splitter that’s a pretty good pitch, but it’s not a killer pitch.  Moreover, I don’t know how much I’d credit an offspeed pitch to the pitcher’s natural physical gifts as opposed to effort and learning and practice.

So it’s one thing to see Bannister as a saber darling in 2007 and 2008 when he’s using PITCHf/x and talking about FIP.  But two or three years later, I don’t see any evidence that he’s learned anything from that to help his craft.  I wonder if he isn’t taking the wrong lessons from PITCHf/x.

I’d much rather have a pitcher like Haren who’s honing his craft by a zealous work ethic and time spent with the bullpen catcher than somebody who knows the right saber language but doesn’t demonstrate an effective process for applying that knowledge.

@Mike Fast:

I’m not sure you’re being fair to Bannister.  Why don’t you think he understands how to be an effective MLB pitcher?  That he hasn’t managed to do it doesn’t demonstrate he doesn’t understand what needs to be done - he may just not have the ability.  I know I’ve seen at least one interview with him where he talked about trying out new grips and pitches and checking to see their pitch f/x movement to see if they might be promising new additions to his repetoire, and also using pitch f/x to try to improve his GB rate, knowing he’s not a strikeout kind of guy.

I also don’t like the comparison to Haren, look at something as simple as swinging strike % and you see a large difference, and combining that with Pitch f/x you can see Haren gets more movement on his fastball....so he might just have better stuff, in addition to better control/command.  Bannister would probably be a lot more succesful if he could post Haren’s strikeout numbers…

Anyways, what I like so much about Bannister is not just that he uses the metrics, but he understands it. He gets the statistical principles behind it that go over a lot of people’s heads.

"So it’s one thing to see Bannister as a saber darling in 2007 and 2008 when he’s using PITCHf/x and talking about FIP.  But two or three years later, I don’t see any evidence that he’s learned anything from that to help his craft.  I wonder if he isn’t taking the wrong lessons from PITCHf/x.”

From 2008 to 2009 he did use it to help his craft. He went from 37.5 GB% to 49.5. He introduced a cutter and I remember from an interview he said the goal was to get grounders. So from the research on himself, he developed a pitch to make him a better pitcher. And it worked in 2009.

I need to explain myself more fully on this, clearly.  Let me take some time to compose a better response and come back to this.

But I can address a few points quickly:

I also don’t like the comparison to Haren, look at something as simple as swinging strike % and you see a large difference

That’s saying that micro results produce macro results, which is true, but it doesn’t do much to explain why one pitcher is successful and another isn’t.  Swinging strike percentage is not a natural physical gift in the same way that fastball speed is, or that long fingers are that can grip the baseball in the odd way that Mariano Rivera does his cutter, or that a flexible body and abnormally long stride are for Tim Lincecum.  Each of those probably have a training and effort aspect to them as well, but they are also based on physical gifts that not everyone has.  Swinging strike percentage is not the same as those.

with Pitch f/x you can see Haren gets more movement on his fastball

I don’t see that at all.  Can you elaborate?

in addition to better control/command

Bannister has walked 7.6% of the batters he has faced in his career, and Haren 5.3%, compared to league average of 8.6%.  That’s not the perfect measure of command, but it’ll do for the sake of discussion.  I’ll grant that you’re correct about Haren having better command than Bannister.  But why?  Shouldn’t a smart pitcher be able to improve his command?  It probably has a physical component to it, but it surely has large mental and training components, too.  Haren has improved his walk rate somewhat since he came into the league.  Bannister, not really.

Anyways, what I like so much about Bannister is not just that he uses the metrics, but he understands it. He gets the statistical principles behind it that go over a lot of people’s heads.

I agree with this.  Bannister is clearly smart, and he has grasped a lot of current sabermetrics, and he’s articulate at explaining them.  That’s rare for a player.

Mike, in all honesty, I don’t think you’re being fair.

Take Cliff Lee this year.  I don’t think anyone would say that Lee has pitches that are, by speed or movement standards, amazing.  What’s amazing is his control.  But the fact that Lee is so unusual suggests that control is not something that can simply be learned, but is a natural talent. 

Haren has that talent to a greater extent than Bannister.  And that’s a key difference.  PitchFX might be able to tell Bannister which pitches he needs to try and improve his control on...but it doesn’t necessarily mean he is ABLE to improve his control.

Alternatively, PitchFX could indicate that he should use his cutter more often.  Which he did last year (which is why we give him credit).  And they may be all he can do! 

A better question for him is WHY ISNT HE STILL DOING THAT?  His GB% has fallen again and just a preliminary look at his PitchFX seems he’s gone back to the four-seamer.

Both Bannister and Haren have cutters, but there is only a 2 mph difference between Bannister’s fastball and cutter while there is a 4 mph difference with Haren’s, as well as with Carpenter and Wainwright. The difference in speeds may make the difference between the effectiveness. That difference in speed is a mechanical/grip issue that may not be related to Bannister’s approach.

@12 - I pretty much agree with what you said, I just reach a different conclusion I guess.

“That’s saying that micro results produce macro results, which is true, but it doesn’t do much to explain why one pitcher is successful and another isn’t.”

Definitely, and I don’t think ( I was trying to present it as such (at least it wasn’t my intention).  I was just using it as one of many quick and easy examples that shows a difference between them.  Why is there a difference?  I’m sure I could come up with a lot of factors, but ultimately, I really have no idea - there’s just too many things that could be at play and it’s impossible to isolate any individual factor to figure out how much influence it has...so I just don’t know why, I just do know there IS a difference.  Fangraphs has Haren’s pitch f/x movement on his fastball at 10.2 vertical, -6.0 horizontal for his career, Bannister at 9.9 and -3.3, respectively.  Again, that looks like a difference to me.  I have no idea what, exactly, causes difference or what impact it has, just that there IS a difference.

For command/control it’s still the same thing to me.  I have no idea what makes for good command/control - I mean, I’m pretty confident some of it is natural gifts like better “muscle memory”, some of it is no doubt mechanics, but really it’s just a complex set of factors interacting with each other and I really don’t know how the end result comes about.  Haren’s better at it than Bannister, that’s what I do know.

The conclusion that I reach when looking at this is there are a range of possibilities, from it being pretty much all physical/natural differences to it being completely mental/figuring out “how to pitch”.  It’s probably somewhere in between, but I really don’t know.  I just don’t think it’s fair to conclude the difference is just coming from things like work ethic, “honing his craft”, etc.  I don’t think we can really know enough to conclude anything except their end results are different. 

Of course, if you know for sure Haren is working harder than Bannister, you can conclude that is at least one contributing factor for the difference in end results, but I don’t think Haren’s better end results tells us he works harder....

Interesting discussion here about applying sabermetric knowledge to the game on the field. I’m only a college player, so this might not translate to the discussion here, but I’ll give it a shot:

I’m probably the most saber-friendly player going right now. I was the only player on my team that had heard of metrics such as wOBA and FIP, (much less understand what they are and how they should be used) and probably even something fairly mainstream like OPS. Ever since I read The Book, I’ve tried to tailor my game around what offers the team the most value. As an example, earlier in my career, I emphasized small ball techniques, such as sacrificing with a runner on second or trying to pull an outside pitch for a groundball to move the runner. At an early age, I was told that those plays were integral to winning baseball games and I worked to build those skills. As I grew older and discovered sabermetrics, I began to overhaul my game. I rarely bunt on my own anymore (usually just if it’s an attempt for a single) and instead try for the best wOBA I can accumulate.

The problem is, this is quite difficult. For a first basemen, I have zero power (I have a very lean frame- 6’6’’ 160 pounds) and I recognize that my lack of oomph will eventually be my fatal flaw. I’ve done everything I can to add more power to my game: I try and work the count, look for pitches to drive, and bulk up in the off-season but there’s no escaping my natural limitations. With that in mind, I have to defend Bannister. More than anyone (and far more than my coaches) I realize what I need to do to become a more productive player. But just because I understand what I need to do, even though I work to improve those skills, doesn’t mean that I can really do it. I’m sure Bannister, and any other saber friendly players, use their knowledge as best as they possibly can. At the end of the day though, the knowledge can only take a player so far.

it’s like the story about Jay Hook (paraphrasing), ‘he could explain how to throw a curve ball and explain the physics behind a curve better than anyone. He just couldn’t throw one.’ Eventually, everyone reaches their peak, and Bannister has probably found his.

Just my two cents.

Let me try to clarify what I was saying about Brian Bannister and sabermetrics.

If what is being claimed is that Bannister is smart, understands sabermetrics pretty well (far better than most players or coaches), and is able to explain the concepts accurately in a way that gets the media to pay attention, I have no disagreement.  I think that’s basically what Tango said at the beginning of this post, and I more or less agree with that. 

If people want to claim he’s the perfect spokesperson for sabermetrics, I’m not sure that’s quite true.  He would be a lot better spokesperson if he were more successful.  Greg Maddux, say.  Whatever.  That is what it is, and I’m not really complaining about that.

What I disagree with, and what some saber-leaning folks both here and on other sites seem to be claiming, is that Bannister is the ideal pitcher that sabermetrics would want in terms of making the most of his ability by leveraging sabermetric ideas.  Where I fail to see the evidence is for (1) Bannister having done a better job than any other pitcher with similar physical abilities or (2) that sabermetric ideas have made a noticeable improvement in his performance.

It seems to me that sabermetrics might find that things like compulsively throwing and working on pitches every waking moment like Haren is alleged to do might be more effective for a pitcher than studying sabermetric ideas.  I don’t know, and it wouldn’t be easy to study, but I don’t think it’s correct to assume that understanding sabermetric thinking is the best way to improve your craft as a pitcher.  It’s pretty clear to see how sabermetric understanding helps a GM or a manager perform parts of their job better.  It’s less clear to see how it helps a player.  I think Morgan Ensberg does a good job of explaining some of that.  I don’t see the same from Bannister.

It’s true that Bannister did add a cutter to get more groundballs.  (1) I’m not sure how well that worked, and (2) increasing groundball rate is a very marginal way to improve as a pitcher.  Dan Haren added a cutter (presumably a better one than Bannister’s?) and improved his strikeout rate from the 7 K/9 range into the 9 K/9 range, with the cutter as his new strikeout pitch.

The other thing that bugs me a little bit is the idea that Bannister especially, and to some extent Max Scherzer, are using PITCHf/x to its full capabilities in improving and honing their craft.  I don’t see it.  I love Dan’s site, and it’s fantastic that Bannister and Scherzer and others inside baseball use it, but what you get from BrooksBaseball.net is simply scratching the surface.  It’s like saying you’re a Class A sabermetrician simply because you know how to find stats on Fangraphs and Baseball-Reference.  Yes, that’s an important step, but knowing what to do with the data is even bigger.  I’m convinced that there needs to be a collaboration between ball-tracking analysts and teams that hasn’t happened yet, outside of probably the Rays with Josh Kalk, in order to apply this data in a useful fashion.  (Yes, I have a personal interest in seeing that happen, but one of the big reasons I want to do that as a business is because I care about seeing it happen in the game.)

Fangraphs has Haren’s pitch f/x movement on his fastball at 10.2 vertical, -6.0 horizontal for his career, Bannister at 9.9 and -3.3, respectively.  Again, that looks like a difference to me.

That’s because Haren throws a two-seamer and Bannister doesn’t.  The MLBAM algorithm for classifying pitches, which Fangraphs relies on, mostly classified two-seamers as four-seamers prior to this year.  It’s better at it this year, but still not quite there for Haren, at least.

But Haren’s two-seam and four-seam fastballs aren’t particularly impressive in any case, in terms of speed, stuff, or results.  It’s his cutter and splitter that are his difference makers.

If every pitcher could throw a cutter as well as Dan Haren, I’m pretty sure they would, Mike. Treating that as something less than a gift (a gift that has to be taken advantage of and worked on, obviously, but I think any pitcher who has thrown a cutter as much as Bannister has made the pitch as effective as they can) seems weird to me.

TCQ/19, why did that gift descend upon Dan Haren at the age of 28?

What I’m getting at here is not fundamentally a criticism of everyone’s favorite player, Brian Bannister, though that’s how people seem to be taking it.  I’m not alleging that he’s an idiot or a slacker.

Sabermetrics is the search for objective knowledge about baseball, is it not?

Many people are assuming, without evidence as far as I can tell, that sabermetrics helped Brian Bannister become a better pitcher than he otherwise could have been, and further, that other pitchers would be better off if they would adopt the same sabermetric approach.  I don’t see any objective evidence for either of those two things.

The main piece of objective evidence that I have previously seen offered for the value of sabermetrics to a pitcher is that Bannister only throws 89-90 mph and he does better than most pitchers who throw that hard.  But I don’t see evidence for that, and in fact, in Haren I see a very strong counterexample.

The addition of the cutter is the other piece of objective evidence that I have seen offered for the value of sabermetrics to Bannister.  I don’t see how it has been much of an improvement to him in run prevention.  I also see many other pitchers who have added cutters who have done much better with that new pitch, Haren being only one.  If Bannister’s addition of a cutter is the best that sabermetrics can offer a pitcher, wouldn’t most pitchers be wise to ignore sabermetrics and continue doing what they’re doing already?

If we’re all just happy that a major league player is a fan of our work, and that’s it, then great.  I’m not trying to bring Bannister down.  I’m sure he’s a far better pitcher than I could ever be and probably works harder at it, too.

I don’t disagree with your core point - if I had to pick a side, I might even be inclined to argue it. The cutter thing just seems wacky to me. Sabermetrics made Bannister decide to throw a cutter, and seeing as his best year was also the year he threw it the most, I wouldn’t say he was wrong. I don’t get that you can blame the information that Bannister’s cutter isn’t as good as Haren’s.

And I don’t think the idea that Haren had a talent that he just hadn’t taken advantage of for a while is very far-fetched at all.

TCQ/21, I’m not sure we disagree all that much, then.

I’m not saying that if Bannister worked harder or did something different, his cutter would be as good as Haren’s.  I don’t know if it would or wouldn’t.  We don’t really know why some pitchers can throw pitches better than other pitchers can throw the same type of pitch.  We can guess, but we don’t have an analytical method for evaluating whether our guesses are correct or how much various factors contribute.

The reason I brought Haren into this discussion at all is that during my analysis of him, some of the things I had been thinking about started to crystallize in my mind. 

I don’t know how much Haren’s hard work and dedication to throwing contributes to his success.  I don’t know how I would measure that.  Similarly, I don’t know how much Bannister’s sabermetric knowledge contributes to his success.

Just as in Haren’s case, I can say his throwing the cutter all winter to his bullpen catcher made the pitch great, and you can’t prove or disprove what I said, you (the general you, not you specifically, TCQ) can say that Bannister used a sabermetric approach to improve his pitching, and you can’t prove or disprove that or measure its impact in terms of runs.

You can wave your hands and use a very loose definition of connecting correlation and causation and say that Bannister’s cutter was responsible for his improved play in 2009, and I can say the same for Haren’s increased strikeout rate in 2009 and 2010.

Neither of those things is getting us anywhere in understanding how pitchers improve.

So I don’t buy it when people say that Bannister used sabermetrics to improve his pitching.  It may be true.  I’m not saying it’s false.  I’m saying that no one’s accurately measured that.  But I guess because Bannister is crediting the work of our community we fawn over that and assume that credit that is being given to us is deserved even if it can’t be measured? 

Correlation equals causation when we get credit for the causation, but if the causation comes from some other source, perhaps those traditionally credited by baseball insiders, we remain skeptical?

I think you’re looking at this in a bit of a narrow way; the cutter thing is weak from both sides (my participation in this thread stemming solely from thinking that was a point based from ambiguity).

I do think that we can say that Sabermetrics have helped Bannister, but from a different angle. Money quote:

“I’m just saying if you get in a stretch like this and you start changing a lot of things, you can really get out of whack.”

I think that’s a really big deal, and that’s something you get from DIPS theory. Changing mechanics when nothing is wrong turns bad luck into bad performance. And preventing that is huge (you could feasibly say that such a mindset could cause a player not to work as hard when they enter a slump as they would otherwise, but I doubt that very many players [only the supremely talented] make it to the bigs without working hard regardless of current performance).

But, basically, this conversation is based on the cutter thing, which I think we agree is not going to be anything beyond conjecture. Not much meaningful disagreement here, I don’t think…

@Mike Fast - I’m actually pretty much in agreement with what you’re saying.  I do think saber concepts have helped Bannister - how can understanding the concepts not help him some amount, after all - but I’m not sure if it’s been in a meaningful way or not.  And while I credit him for understanding the principles, that doesn’t mean he’s been able to do anything but scratch the surface of what’s possible yet.  Just look at how many bloggers are out there that have been working hard on improving sabermetric knowledge for years and how long it’s taken them just to get to this point.  It’s a process, it will definitely take time for players and coaches to figure out how to best use the knowledge once they’re open to it and understand it.  I do think it’s a great first step, though, and yes, it would definitely help promote “the movement” if he was much better and had results that impressed others to credit to sabermetrics.

I don’t really agree with Mike’s argument.  Specifically, I don’t agree that Bannister hasn’t used saber concepts and Pitch f/x to improve his performance. 

http://www.hardballtimes.com/main/article/brian-bannister-tbd/
http://www.fangraphs.com/blogs/index.php/bannister-love-pitchfx/

If he uses Pitch f/x to figure out that his cutter should be a good groundball pitch if throw more regularly and recognizes that groundballs = good, that seems to be the perfect defintion of applying sabermetric research to improve himself as a player.  His xFIP was 4.37 in 2009, which was vastly improved over his 2007/2008 self which saw him throwing more fourseamers

In fact this year, he’s reverted to throwing more fourseamers (according to the BIS pitch types, which could obviously be wrong), and his groundball rate (and his xFIP) have declined back to 2007/2008 levels.  I’m not why he’s throwing more fourseamers this year, but it doesn’t seem to gel with his revelation last year.  So maybe BIS is just mis-classifying, or the Royals are encouraging him to throw more fourseamers?

Nick:  “I wandered over to introduce myself to Brian Bannister, who waved me to have a seat, and was as gracious and engaging as you’d expect. We spoke about sabermetrics and Pitch F/X and the adjustments he was trying to make this season (fewer cutters, trying to change speeds more often) for about 15 minutes.”

http://www.ranyontheroyals.com/2010/07/good-day.html

I’m not sure why he’s doing this either considering it makes little sense for him, but it seems that BIS is correct in saying he’s throwing more 4-seamers given this statement.  If i had to guess though, he’s realized the limitations of using the cutter in that way and is trying, while on a team where him losing his job seems unlikely, to see if he can improve in another way.

Thanks garik.

It makes sense I guess, but I think it would have been smarter for Banny to stick with last years plan and see if it could be sustained for another year.

Also, it’s really amusing to see Rany so happy about the Royals and Dayton Moore.  I can’t wait till Dayton trades for Jeff Francouer or some crap.

I don’t really agree with Mike’s argument.  Specifically, I don’t agree that Bannister hasn’t used saber concepts and Pitch f/x to improve his performance.

That is not my argument.

What I said was this:

Where I fail to see the evidence is for (1) Bannister having done a better job than any other pitcher with similar physical abilities or (2) that sabermetric ideas have made a noticeable improvement in his performance.

I still fail to see the evidence.  I’m open to being persuaded, but by evidence, not because I want something to be true.

Correlation = causation is the argument being presented against my point #2.  That’s fairly unconvincing to me.  On the same level with Haren’s dedication to throwing being credited with his success with his cutter.  There’s probably an element of truth in both arguments.  But how much, specifically, in each? 

Did Bannister actually get more groundballs and less home runs on the cutter in 2009?  It’s irresponsible sabermetrics to assume that’s true and 100% the cause for Bannister’s change in performance level without evidence that shows that.

The cutter in general is a more-groundball-inducing pitch than the four-seamer (44% vs. 35% according to Harry’s numbers).  If he turned half of his pitches from four-seamers to cutters, on average we’d expect his groundball rate to go up by about 4.5%.  That’s not going to have a huge impact on his run prevention.  Now, he may very well have done something different that made the cutter work better as a groundball pitch for him, but without seeing that data, I’m very uncomfortable simply accepting the assumption on faith.

Also, no one is addressing my point #1.

More to the point, my larger argument here is not really about Bannister.  It’s the same argument that Morgan Ensberg made, that sabermetrics has not been tailored to be useful to the player in improving his play.  A sabermetric approach (i.e., evaluating evidence objectively) can certainly be helpful to players.  But 99% of sabermetric ideas, stats, and studies as currently presented are targetted at helping GMs, managers, and fantasy baseball players and not at helping the actual baseball players themselves.  If you restrict it to PITCHf/x sabermetrics only, the picture’s a little better, but still 90% of the stuff is not terribly helpful to players the way it’s available right now.

But Mike, regarding non-pitchfx sabermetrics, how much is physically helpful to players?

For example, I think DIPS theory would certainly be psychologically helpful to pitchers in that those who understand the theory are more likely to shake off some bad luck starts (being hurt by a lot of singles in a row for instance) and not make overreactions to such starts.  I think it might be helpful to make players realize that some principles are silly (a heavy fly ball pitcher pitching to contact in some situations). 

But helping the actual improvement of a player?  I’m not sure how much is there.  Analysts have worked on improving our MEASUREMENTS of players’ worth rather than how players themselves should improve.  I’m sure the metrics would show certain players that it might be worthwhile to go for power over batting average at times, but really how much can one improve via that?

PITCH/FX is different in that while we use it for sabermetric purposes, it really is just a computerization of pitch scouting.  So it’s obviously more helpful for players.  But other stuff isn’t, so what can be done? 

Back to the point of this thread, are you suggesting that the optimal spokesperson for sabermetrics would be a GM or coach who uses them to successfully lead a team to the playoffs multiple times?

Garik/30, I’m suggesting that sabermetrics that makes sense to Morgan Ensberg is the kind of sabermetrics that will find wider adoption in MLB moreso than the sabermetrics that makes sense to Bannister.

Ensberg is obviously a former player, whereas Bannister is a current player, so they’re not directly comparable.  But I really like the way Ensberg has shown that he thinks about all the game situations.

That’s putting a little bit of a fine point on it, though.

Don’t get me wrong, here.  I like that Bannister is looking at PITCHf/x data.  And he’s probably smart enough to do a decent job with it.  I think he could do better with it if he had the cooperation of his pitching coach, manager, and a good analyst in working through it all.  (I know, pipe dreams.)

But I don’t think that the Bannister approach is something that can really be leveraged for very many other players.  That’s why I don’t view him as the perfect spokesperson.  If he were to suddenly turn around and put up a 3.20 ERA every year, then other players would listen.  But they’re not going to listen because he can improve his xFIP from 4.80 to 4.30 for one year.  That may catch the eye of a sabermetrician who wants to dig deeper, but it’s not going to win the players over on the value of PITCHf/x data or understanding DIPS.

(I guess he did win over Greinke, but there again, it’s not clear to me if Greinke changed anything as a result, and if so, what the measurable impact was.)

But Ensberg’s approach, if we can format our data into that language and perspective, is something that ought to be more accessible to more players.

To be clear, I’m not arguing that all or any other saberists ought to change their language in the way I’m suggesting.

It’s something I care about.  Others may not care about it.

But I do believe that something like that is required to see wider adoption of sabermetric ideas among players and coaches, for those who care about that.

I should also say that I don’t think wanting to impact the game itself is the only thing that should drive looking at sabermetrics from a player’s perspective.

In fact, it’s not the only motivation for me.  I’m also finding in my quest to understand the game of baseball, I’m being driven more and more toward the players’ perspective rather than the traditional GM-oriented perspective.  That’s particularly true in trying to understand what makes certain pitches effective.  It’s hard to do that without understanding the many subtle dynamics of the batter-pitcher confrontation.

A little blurb from ESPN:

“As far as the big picture goes, his days as a major league starter could be winding down age at 29 as his home-run rate (22 allowed in 119 innings, 1.66 HR/9IP) and walk rate (3.25 BB/9IP) have swelled to career highs this season.”

http://espn.go.com/mlb/players/profile?playerId=6474 (Player News Updated 7/28/10)

Even though he added a cutter to generate groundballs, his G/F ratio is 0.77 which is worse from his 1.06 ratio from last year. His K/9 has also dropped from last year’s career high of 5.73 down to 5.37.

I know I’m using some crappy statistics here, not considering DIPS or xFIP or adjusting for this or that… but even something like a G/F spike downwards can’t be explained solely by luck, can it? If not, what else can explain it? Is there something in pitch f/x or the Royals defensive performance or pitch sequencing or something that can explain why Bannister is having a career-worst year? I understand the flipping coin analogy very well, but can luck really outweigh pitching skill to this kind of an extent?

Most everything is baseball has a component of luck and a component of skill, but you tend to see the effects of luck most over a pitcher’s season-sized sample of BABIP, and to some extent in extreme values of LOB% and HR/FB.

His BABIP is .310 this year, compared to his career value of .291 in about 2200 balls in play.  Regress his career value with about 3700 BIP at the mean value of .300, and you get an expected BABIP of .297 instead of .310, which has cost him a little over two hits, or about 0.15 runs/9.

His LOB% is 69%, and that’s a pretty typical value.  His career value is 67%.  I’m not sure of the details, but I don’t think his expected value would change a lot.  Someone else may know better about this portion.

His HR/FB% is 14.6% this year compared to 10.1% in his career.  I know that regresses fairly heavily toward the mean (xFIP regresses 100% to the mean).  If we use his career HR/FB, we get 16 expected HR instead of 23 actual.  Those 7 extra home runs have added around 0.72 runs/9.  That’s the rough explanation for the difference between his 5.53 FIP and 4.85 xFIP.

Scouts’ takes on various major-league players:

Royals right-hander Brian Bannister: “I think this guy has some upside as a back-of-the-rotation starter, but he’s just so inconsistent. He’s never been able to repeat his delivery and that’s been the root of a lot of his problems.”

http://www.baseballprospectus.com/article.php?articleid=11701

The sabremetric take is it’s bad luck and this particular scout thinks he is having problems repeating his delivery. Mike Fast #35 provided some good data on that. If he is having problems repeating his delivery, could that show up somewhere in pitch F/x or some other tool?

Richard:

I would think that Bannister would notice his inability to repeat his delivery using PitchFX, yes.

It could start with a look at his release point, which I am sure would dictate whether he had problems repeating his delivery or not. I am sure there are also some underlying things within the PitchFX function that could offer answers to that problem.

Just a guess, though.

If he is having problems repeating his delivery, could that show up somewhere in pitch F/x or some other tool?

If you back out the camera calibration issues, then yes, you can see problems with repeating delivery in release point data.  If you take the PITCHf/x data as is it is almost impossible to see this effect.

To me Bannister is the perfect spokesman for sabermetrics because he’s articulate and uses sabermetrics to explain the meaning of his performance.  Since he’s a major leaguer, he’s got street cred.

Sabermetrics isn’t going to give him 3 mph more on his fastball.  But it will help explain why his peripherals haven’t changed but his ERA has and how reporters should interpret it…