Beyond CERA: The Sabermetric Case for Rob Johnson
I am new to this forum, so it has taken me some time to become acclimated. Visiting a blog community is much like visiting a foreign land -- one has to make an effort to learn the language and customs. Like a traveler in a far away land, however, I feel that not only do I have much to learn from you, but you can learn from me as well.
One custom I have noticed here is that everyone seems to enjoy putting down Rob Johnson. I think in a community like this one, it is easy to develop a sort of groupspeak echo chamber. One person puts down a player, and then another, and then another, and before long, the entire community is down on that player without really thinking about it and evaluating their play objectively. I feel that has happened here with Rob Johnson.
I know you all like to poke fun at Rob Johnson's batting average, and his CERA in order to "prove" that he is a below average major league baseball player. While I think those statistical tools have their place, I believe that we can all work a little harder to learn more statistical tools and refine our understandings. I think if you all open your minds and make a concerted effort to educate yourselves, you will change your view of one of my favorite players -- Rob Johnson.
Many baseball analysts and, frankly, many fans, use statistics like errors and fielding percentage in order to evaluate defense -- even catcher defense. But I am here to tell you that those tools do not tell the whole story. Statisticians have much more sophisticated tools now that break down individual plays and can help tell us which players on a field at a given time are responsible for which outcomes. This is important because it helps illuminate repeatable skills, which tell us how players will perform moving forward.
One of the things that Rob Johnson does well, for example, is catch curveballs. There is a lot that goes into determining the success of a pitch (speed, location, sequencing, delivery), but for a curveball the catcher is of the utmost importance. The catcher of a curveball can affect how much the spin causes it to deviate from a spinless trajectory, and is commonly broken up into catcher horizontal and catcher vertical components. So the catcher vertical movement is how much a pitch sinks or rises compared to the same pitch with an average catcher. The catcher horizontal movement is how much a pitch tails horizontally compared to expectation with an average catcher. Positive catcher horizontal movement indicates a catcher who causes a pitch to tail away from a RHB and in to a LHB.
I have taken the initiative, using a MLB service called Gameday PitchFX, to record the vertical and horizontal movement of curveballs thrown to Rob Johnson, versus those thrown to the other Mariner catchers: Kenji Johjima and Jamie Burke. Here are the movements of curveballs caught by those catchers. The gray dots are the movement of all curves for reference.
As you can see, curveballs thrown to Rob Johnson have a wide range of movement, but generally lots of horizontal movement (tailing away from RHBs an average of about 6 inches) and not much vertical break. Curveballs thrown to Burke have lots of vertical and horizontal movement, sinking and tailing away from RHBs by almost 10 inches in each direction. Johjima-caught curves have little movement what-so-ever (his curves are very close to 0,0).
These advanced statistics, Catcher Vertical Curveball Movement (CVCM) and Catcher Horizontal Curveball Movement (CHCM), tell us something we cannot glean from rudimentary statistics like CERA. If you were a pitcher, and you could throw to one catcher who does not improve the break of a curveball (Johjima) and another catcher who does (Johnson), which one would you choose?
Jamie Burke-caught curveballs have the most movement of all, but he is not considered a very good candidate to be the Mariner catcher because he has low CFBV (Catcher Fastball Velocity). That is to say, that Jamie Burke-caught pitches have a high correlation with low fastball velocity. This is a glaring weakness in Burke's game and will be the subject of a subsequent post.
But the comparison between Johnson and Johjima is telling. I am confident that if this community were to have an open mind to advanced statistics such as this, it would share my view that Rob Johnson gives the Mariners the best chance to win. Let's all try to improve our statistical analysis, or we will fail to appreciate that one of the better catchers in the game is right here in our own backyard.
***
UPDATE: I really appreciate the outpouring of support in the comments section. Clearly, I underestimated the openmindedness and sophistication of this blog community. It is great to see so many people willing to admit that they were wrong before.
99 comments
|
27 recs |
Do you like this story?
Comments
I'm actually fairly certain this is a joke
and if it is, it’s one of the best jokes in LL history
by seattlebruin on Jul 16, 2009 3:05 PM PDT up reply actions
This is not a joke
Why would you say that?
by Attractive Nuisance on Jul 16, 2009 3:08 PM PDT up reply actions
Because it feels like an extended exercise in irony
by Graham MacAree on Jul 16, 2009 3:10 PM PDT up reply actions
Irony?
Please do not make fun of what you don’t understand. If I have not done a good enough job explaining some of these new statistics to you, I would be happy to go into more depth. But you should not just insult a post that you don’t understand.
by Attractive Nuisance on Jul 16, 2009 3:12 PM PDT up reply actions 3 recs
It's a joke.
Statistics are not a joke
You can attack the speaker, but the facts are unassailable.
by Attractive Nuisance on Jul 16, 2009 3:15 PM PDT up reply actions
So here's a question
If, as you say,
Statisticians have much more sophisticated tools now that break down individual plays and can help tell us which players on a field at a given time are responsible for which outcomes
Why aren’t you using any of them? Curveball movement? Seriously?
Nice Guys Finish Third - My semantics are a waste of time.
by pdb on Jul 16, 2009 3:11 PM PDT reply actions 1 recs
If this is a joke then well played, though
Nice Guys Finish Third - My semantics are a waste of time.
These facts may alter your perception of reality, but...
…it is your perception of reality that should change. The facts will not and cannot.
by Attractive Nuisance on Jul 16, 2009 3:16 PM PDT reply actions
Ok, I'll be the straight man.....
Hmm, that’s interesting. Tell me more! Have you calculated Burke’s velocity-dampening effect by pitcher? Does he sap more precious velocity from Felix or from Vargas?
And how much additional break on a curveball does Johnson impart to pitchers like Vargas who do not throw curveballs? Or is that one of the great imponderables in your line of work?
I'm going to start calling you Dick Smothers
you’re a natural straight man.
Nice Guys Finish Third - My semantics are a waste of time.
Answers
I have not controlled for pitcher. And, frankly, I’m not your personal research assistant. If you have a question like this, why not go out and do the research yourself. I would be thrilled to read your findings.
Regarding additional break to non-curveball throwers: in the study above, I only examined curveballs as a whole; so non-curveball pitches from both pitchers who do and do not throw curveballs did not factor into my results.
by Attractive Nuisance on Jul 16, 2009 3:32 PM PDT up reply actions
Whoops a doodle.
I didn’t mean to insinuate that you would do the work for me. I couldn’t possibly do research at this level, and I’m just trying to figure out how best to utilize your (and Johnson’s :) ) talent (just to be clear, the previous parenthetical included one smiley face before the closed parens; I don’t really know how to make that clear. I’m not saying that you or I or Johnson are smiling with a double chin or something.).
Another thing I’ve been thinking about – and this is not a request or command – is doing a WOWY analysis on Johnson’s presence. Do pitchers pitch better when he’s nearby, even if he’s not catching? Do you think this is a promising application of Science?
Interesting
Traditionally, it has been difficult for sabermetricians, and statisticians to quantify intangible factors like “chemistry.” That said, I suppose if I could get access to a dataset on dugout player proximity to mound, I might be able to kick the numbers around a bit and see if I can observe something with statistical significance. Perhaps, for instance, Rob Johnson in the dugout does have a positive relationship with pitching success.
And I’ll add a note here: by “pitching success,” I do not necessarily mean the traditional statistics like ERA or Wins. We want to isolate fielding independent pitching success. The best way to do this, in my experience, is to use a statistic called Fielding Independent Pitching (FIP).
So, if I used FIP as the dependent variable, and dugout player proximity to the mound as the independent variable, I could potentially observe some sort of proximity effect if there is one. The real issue here is that I am not sure I can get my hands on that kind of dataset. I would imagine that if a team had a dataset like that, they would keep it to themselves for private use, rather than give their secrets away to the public.
If anyone has ideas on this, I’d be open to them.
by Attractive Nuisance on Jul 16, 2009 4:02 PM PDT up reply actions
To get a rough idea
you could use Rob’s days on the Bereavement list…and assume that any other days he was active with the team that he was in the dugout. Your independent variable then only has two cases…Rob Present or Rob Missing. Might still correlate with FIP though, would be interesting.
by appleshampoo on Jul 16, 2009 4:05 PM PDT up reply actions
Well, traditionally it has been difficult for sabermetricians to quantify Catcher Curve Vertical Break
but you didn’t let that stop you. You seem like a guy who isn’t bound by tradition, and I really admire that.
I’ll admit it WILL be difficult. One might need to scour the broadcasts to see if Johnson’s in the dugout, if he sits near to the pitcher between innings, if he’s adding value by bouncing suggestively/stretching/discussing curve balls with Johjima, etc.
The best way to do this, in my opinion, would be to contact BIS or STATS, which are firms compiling play-by-play data. They have people at the game categorizing balls in play, so it would add nothing to their budget to instruct these stringers to ALSO collect data on player proximity. You’d have the data from each game, and I’m almost positive there’s a statistically significant improvement in a pitcher’s FIB or whatever if they’ve been near Johnson or even touched him.
Sabermetrics are going to measure things in 5 years that we never even imagined now.
by marc w on Jul 16, 2009 4:12 PM PDT up reply actions 1 recs
Honestly...
…I doubt that one of those firms would have much interest in compiling the kind of dataset you are proposing. While, I thought your dugout proximity theory was creative and I wanted to encourage your continued effort and curiosity, I have serious doubts in the project you are proposing bearing fruit. The reasons are twofold.
First, even if Rob Johnson proximity had any kind of correlation with FIP, we have to think common sensically about these relationships and why there might actually be causation. For something like proximity to pitcher, I am not yet convinced that there is any reason to believe that it has a real effect. I played baseball in high school, and I actually pitched a bit for the junior varsity team. While there were all sorts of combinations of people at a variety of distances away from me all the time, I do not remember noticing any impact at all. Though, I’ll grant you, to my knowledge Rob Johnson was never in close proximity to me during that time so I cannot be sure of his effect.
Second, let’s say that Rob Johnson actually touching the pitcher, as you suggest, actually was found to have a statistically significant effect. it is unclear how a team might leverage that information to their advantage. It would be against the rules of baseball, for instance, to have Rob Johnson out there by the pitching mound as a tenth man touching the pitcher while he pitches (though, I’ll grant you, perhaps MLB should reevaluate this rule). Maybe you could remove an outfielder and substitute Rob Johnson by the pitching mound, but then your outfield defense will be much worse. As this site and others have proven quite convincingly in my opinion, outfield defense is important.
by Attractive Nuisance on Jul 16, 2009 4:43 PM PDT up reply actions 6 recs
There are many, many avenues a sabr-minded team could explore to maximize the value
of a team member who made people better by touching them.
1) Johnson would touch the pitcher between innings, either through backrubs, or direct hand-to-throwing arm contact.
2) Johnson could accompany the manager and/or pitching coach on any trips to the mound. This, much more than anodyne ‘advice’ or rah-rah BS, could really affect team performance, as these trips often occur when the leverage index is high.
3) Johnson could be moved to first or third base coach for the games he’s not catching. I can’t imagine there’s a whole lot of skill involved in those roles, and he would get a lot closer to the pitcher. If rickpo’s right, this would have an exponential effect on pitcher performance.
I appreciate this discussion, and again, I’d like to thank you for keeping an open mind. There’s so much that just gets jumbled together in the black box labeled ‘chemistry’ that a really smart sabermetrician could tease out, disaggregate and put to use. Mike Sweeney’s been a highly important member of the bench, but what if the M’s acquired a Jew who grew tired of being told to ‘Go with Christ’ all the time? Would the tension affect only the newcomer, or would Sweeney’s effects on the rest of the roster be diminished somehow?
How has Hannahan affected clubhouse morale? Most have never played with him or been around him, and some may find his gigantic penis a distraction. Others might feel motivated, I have no idea.
I’m just frustrated that I don’t know how to test these theories. And I’m not asking you to either, I’m just reiterating that this is a brave new world for sophisticated analysis, and sophisticated analysts like yourself.
by marc w on Jul 17, 2009 12:19 PM PDT up reply actions 12 recs
HAHAHA
Turn green now.
The Yankees suck-a-doodle-doo!
by JamMasterJesus on Jul 17, 2009 8:42 PM PDT up reply actions
Oh I see...
…you are hypothesizing that Johnson-contact has a lasting effect. My initial understanding was that you thought the Johnson-contact had to occur while the pitcher pitched, which sounded to me like it might get awkward.
Another idea that you have not considered is we might be able to put Johnson in a harness and drop him down from the air near the mound. This would have been easier in the Kingdome, but perhaps we could use a similar apparatus to that used for the aerial camera view during football kickoffs. If that fails, we could probably just use a crane — there are plenty of those in the vicinity of Safeco.
I think the bottom line is that you and I have proved this information would be valuable to obtain. We just need to put in the legwork now and get the data. Does anyone know Dave Cameron’s phone number? I think we should get him involved.
Another point I’d like to make here, is that we ought to keep this information secret. Like just between us here on this blog. I’m just worried that if we do find that there is some statistically significant effect, that Johnson will be classified as a PED.
Regarding your other suggestions, those sound pretty dumb.
by Attractive Nuisance on Jul 18, 2009 12:50 PM PDT up reply actions 3 recs
I would think
the catcher’s effect on the break of pitches decreases inversely with the square of the distance from the pitcher. So it’s not enough to know if Johnson was in the dugout. You’d need to know where in the dugout he was sitting.
What're ya gonna do with those pies, boys?
Actually, Baseball Prospectus has this exact data (but only for the past three years).
You have to pay a premium for it, however. Only a two teams in the league currently subscribe at for data at this level. I could certainly get you access if you really want to go here. Let me know.
by Sinking Away on Jul 18, 2009 9:41 PM PDT up reply actions
One way to definitely measure Johnson's value while sitting in the dugout would be to measure the FIP of the pitcher in an inning which follows a time when the pitcher was in the dugout and Johnson spent the majority of that time on the bases.
If this ever happens, we will then be able to measure it.
by Kenny Knows Sports on Jul 17, 2009 3:15 PM PDT up reply actions
Please tell me that the rest of your material is this good
Nice Guys Finish Third - My semantics are a waste of time.
Started out funny
But I think you milked it for too long. I like the idea though.
That's what you get...
…for wasting your time reading some random dude’s Fanpost on a baseball fan blog.
by Attractive Nuisance on Jul 16, 2009 3:48 PM PDT up reply actions 1 recs
You just blew my mind.
And it made my left foot tickle..
...and now I'm here
In the wrong thread even?
Fuck.
The Rise of a Superstar:Justin Upton-.398wOBA, 21 years old.
by Goose on Jul 16, 2009 4:04 PM PDT up reply actions 1 recs
Bravo.
I will smash your face into a jelly.
by Phil Hatzenbuehler on Jul 16, 2009 3:59 PM PDT reply actions
I don't get it....?
Aaron King is still my homeboy... iffy mechanics and all
If Dustin Pedroia played in Seattle, not many people would be talking about him.
GET THAT VORP SH!T OUTTA HERE!!!
Great post!
I’d love to know, can anyone access this pitchf/x data? If so, where can you find it? This post has inspired me to explore my own theory that Rob Johnson’s massive personal attraction increases the velocity of fastballs thrown at him.
Gravitation
Which would fit with the dugout proximity theory. Decreases with the square of the distance away.
Chicks dig the long ball.
Thanks!
To be honest, I really was not sure this community would be open minded enough to consider new, more advanced, analysis. But the recommendations and encouraging comments reflect as well on you all as they do on me. Congrats to Lookout Landing for swallowing its pride and embracing education.
by Attractive Nuisance on Jul 16, 2009 7:52 PM PDT reply actions 1 recs
This proves once and for all that...
Jamie Burke is the Mariners best defensive catcher since Dan Wilson.
"Where's my doctor?"
Holy shit.
Andrew Raycroft for backup? Does not compute.
by Woodinville_12thMan on Jul 18, 2009 11:52 AM PDT up reply actions
Question:
I don’t get it when you refer to the velocity. How does the catcher affect a ball’s velocity in any way?
Well, we know Rob Johnson sucks.
Think about it, and you’ll see the answer to your question.
...and now I'm here
by CapSea on Jul 17, 2009 4:59 AM PDT up reply actions 2 recs
I think that if we are going to build a community here that is committed to learning and improving our collective baseball knowledge,
we need to make an effort to make well-supported, thoughtful posts. Your post here says “how does the catcher affect a ball’s velocity in any way?” This strikes me as a lazy question and I think it provides a good opportunity to make this point.
First, catchers impact a ball’s velocity all the time. Every time they throw a ball, catch a ball, hit a bill, or nudge a ball with their little noses, they affect the velocity of the ball.
Second, I think you may have been trying to make a point about how a catcher might affect the velocity of a pitched ball while it is on the way to the catcher. But if you are going to make a negative inference like that, you might try and back it up with some of your own statistics, or at least some reasoning. Don’t just appropriate conventional wisdom for your own a if it was in a community fridge, or you will often find that you have stale wisdom. Why wouldn’t you think that a catcher has an affect on the velocity of a pitched ball?
by Attractive Nuisance on Jul 17, 2009 9:30 AM PDT up reply actions 4 recs
We have at least one example of a player who could affect the velocity of balls in play
Thanks to some research from Derek Zumsteg.
Really, if you’re going to be presenting yourself as a sabermatician, you should be aware of the research that has already been done.
by Chris Hafner on Jul 17, 2009 9:33 AM PDT up reply actions
That link is to a post about Bugs Bunny!
I don’t know who this Zumsteg guy is, but his post does not make any sense. Bugs Bunny is a cartoon! It’s not real. It is just some idiots drawing pictures and putting them in sequence to an audio track. Thus, the laws of physics that rule our world do not apply. I really do not see how cartoons such as this have any bearing on the statistical analysis I have been advocating here.
by Attractive Nuisance on Jul 17, 2009 9:42 AM PDT up reply actions
Wow, just wow
I was in shock and disbelief reading this all the way to the last paragraph when a light went off (yeah, I’m a bit slow).
well done. I will never doubt Rob Johnson again.
The problem with RJ is not his CERA, which is awesome,
it’s that it transfers into the batter’s box. Opposing pitchers get to benefit from it too.
Well done, sir.
by Bearskin Rugburn on Jul 17, 2009 8:47 AM PDT reply actions
I agree with your conclusion, but not your methodology
This was a nice try, and I applaud your desire to learn sabermetrics, but I have some quibbles with your process.
For example, you mentioned above that you would use FIP (Fielding Independent Pitching) to measure pitching success. As I’m sure you’re aware, TRA does a much better job of filtering out other data that clouds pitcher evaluation. However, even TRA has its limitations – there is still too much random statistical variance built in. Using either of these metrics is like using a slide rule when a supercomputer is available.
This is why I use PIP (Pitching Independent Pitching). It removes the last bit of noise from the statistics by eliminating pitcher contribution from the evaluation process. This is the ultimate metric, unclouded by statistical noise and volatility.
Judging Johnson vs. Johjima, we find that they rate exactly the same according to this metric – if we remove the pitching component, we see that no pitches were thrown, caught or hit. While the nature of these results make it difficult to calculate WAR for either player, we can at least assume that Johjima’s and Johnson’s impact on defensive and catching are nonexistent and, thus, equal. Thus we can calculate their relative value by comparing their contracts and the impact on the team’s payroll. Johjima makes dramatically more money than Johnson, so, by comparing contract value versus CPIP (Catcher Pitching Independent Pitching), Johnson is the more valuable catcher. You arrived at the right conclusion, but with stone-age analysis. Once you apply better process, your results should also improve.
One could also extrapolate CPIP and PIP to evaluate the run prevention impact of catchers and pitchers across MLB, with startling results, and I’m currently in the process of developing PIP-Against, which measures hitter performance once pitching is removed. However, all of this is a discussion for another day.
by Chris Hafner on Jul 17, 2009 9:29 AM PDT reply actions 11 recs
Sorry to break it to you,
but there is a serious flaw in your methodology. Pitching Independent Pitching likely tells us nothing. You have not shared your equations with me, but it seems to me that you are most likely using the same variable (pitching) as both a dependent and independent variable. Your results confirm this suspicion.
Let me try to help you with an example. Let’s say you are trying to figure out the relationship between a dependent variable Y and an independent variable X. So your equation will look like this:
Y = beta * X
Where beta is a constant.
You can then plug in a dataset full of Xs and Ys, run a regression and find out the value of the constant beta. Vwala, you now know more precisely the relationship between X and Y.
But, if your equation looks like this:
Y = beta * Y, then we know before we plug in any data at all, that beta = 1. Beta will equal 1 for all players in all circumstances.
Accordingly, it does not make any sense to control for all pitching when your independent variable is pitching. Your analysis is rather silly, and would make more sense as some kind of joke or spoof post rather than as a part of a serious discussion.
However, if you were trying to make some sort of joke, I must give you credit. It was quite humorous. Though, if you were not making a joke, no offense. We are just not all cut out for this.
by Attractive Nuisance on Jul 17, 2009 9:53 AM PDT up reply actions 4 recs
New advanced metrics have always hit opposition from those unwilling to drop their preconceived notions ...
… but I’m disappointed to see that behavior from somebody seemingly so dedicated to breaking new analytical ground.
I would like to help you correct some of your errors above, but unfortunately I can’t reveal the inner mechanism of PIP (or its variants, CPIP and PIP-Against) as I am currently licensing its use to several organizations that are interested in advanced statistical analysis—among them, the Houston Astros, New York Mets, and the Kansas City Royals. In any event, I would recommend taking a statistics course from an accredited university before you begin judging the work of others.
However, I will add that by writing off my findings as a joke, you’re engaging in the same behavior with which others have responded to your your well-intentioned but ultimately erroneous analysis in this fanpost. I hope we can all grow up and treat this analysis with the level of seriousness that it deserves.
by Chris Hafner on Jul 17, 2009 10:08 AM PDT up reply actions 4 recs
I dont mean to question your methodology
but it looks like your graph was made in ‘paint’ with the spraycan tool.
Thug Life
Because Matthew doesn't use MS Paint when he makes heatmaps
by seattlebruin on Jul 17, 2009 10:01 AM PDT up reply actions
Nope
I think we have rooted out a non-statistician.
by Attractive Nuisance on Jul 17, 2009 1:28 PM PDT up reply actions
Matthew is the worst analyst ever
he didn’t even post anything this weekend
by seattlebruin on Jul 20, 2009 11:00 AM PDT up reply actions
In that he's white, plays catcher for the Mariners and is younger than 40?
Yes, he reminds me of a younger Dan Wilson in all those ways
by seattlebruin on Jul 18, 2009 3:33 AM PDT up reply actions
Time to submit for publication
I think you have the making of a lead story for the Journal of Irreproducible Results. Nobel prize to follow?
Chicks dig the long ball.
No no, that would be the Ignoble Prize
Google it. A couple years ago it went, among other things, to a DARPA project to make a ‘gay bomb’. The idea was that you could spread some kind of pheromone that would make enemy soldiers go gay for each other, and while they’re busy doing each other’s nails (and the nasty) your troops come in and kill the shit out of them.
This was a real project. Your tax dollars at work.
by Bearskin Rugburn on Jul 18, 2009 7:29 AM PDT up reply actions
Sadly a similar kind of approach is still being used to keep women out of combat
The theory being that women in fatigues and combat boots in the front lines will distract the male soldiers. Like mortar shells landing all around you and bombs going off and AK47 fire puts you in the mood? This kind of crap is insulting to the guys (and gals) risking their lives.
Chicks dig the long ball.
I didn't mean this to be political
Just as an example of thinking/fail. If I crossed the line I apologize. Please don’t nuke me Graham!
Chicks dig the long ball.
This must be the new mod they brought in.
Career Safeco Record: 6-1 (3 of which were started by Felix)
by .Taylor on Jul 18, 2009 5:27 PM PDT reply actions
I found this even more funny because..
On the Mariners forums, they were talking about Rob Johnson, and one poster said,
“there seems to be a more advanced discussion going on here”, and he cited this post hahahahahahahahahaha. Im trying to find it, I think its simply called “Rob Johnson”.
by Pessimistic Optimist on Jul 26, 2009 10:43 AM PDT reply actions
I don't understand why you think that is funny.
This is an advanced discussion. Did you see the Johnson in a harness proposal?
by Attractive Nuisance on Jul 26, 2009 11:53 AM PDT up reply actions
And I have credibility because...
I made a graph.
by Attractive Nuisance on Jul 26, 2009 11:55 AM PDT up reply actions 2 recs

by 












