Rob Johnson's CERA and Kenji's

So there's a 'Point of Contention' regarding Rob Johnson's defense, whether his CERA is plus, and even whether a plus CERA is something that can happen in major-league baseball.

Notice Jeff's inversion of the debate in his article at LL:

What's funny is that, if we're just scrounging for as much evidence as we can find, there's a lot more evidence that Kenji was bad than there is that Johnson is good. Kenji was here for four years. In three of those years, his numbers were a lot worse than those of his backup(s), and in the fourth they were about equal.

Now, we are not talking about Jeff S here, whom you know we consider to be an absolutely top-flight analyst...

But this reminds me of the fact that many (other) sabermetricians get wrapped up in this clear logical fallacy: they're 100% ready to believe that an MLB player can be worse-than-average in a hard-to-measure skill, but 100% cynical that any MLB player could be better-than-average at it.

For example, you'll search far-and-wide to find a sabermetrician who believes that Mariano Rivera can "dial it up" in the clutch and throw 16 shutout innings because he's in a zone. But any sabermetrician will be perfectly willing to hear evidence that a Brandon Morrow chokes in the 9th inning...

Saberdudes dislike the idea that a particular catcher has a positive CERA ability. Many, however, are very friendly to the idea that a particular catcher has a negative CERA ability.

The catch comes in that they argue CERA effect does not exist at all on Monday, but come Tuesday to talk about the guy who's bad at it, now the effect is possible...

=== Six o' One, Half-a-Dozen o' th' Other Dept. ===

Returning to LL's article, Jeff S is open to an argument that Johjima had a problem for four years.

If that's true, we don't need to go any farther than that. The 2010 Mariners would stand to improve their pitching. Johjima was part of the 2009 team. He won't be part of the 2010 team.

Sully re-frames the debate in more harmonious terms. If Johnson's skill is never going to be anything other than an argument, then perhaps Johjima's disconnect is an easier way to talk about the same thing.

I mean, if you're going to talk about 2010 at all, then talking about the catchers' effects on the pitchers has to be part of it. What would you think about analysis that ignored outfield defense?

=== Name-Dropping Dept. ===

The tools scouts -- including Don Wakamatsu, kiddies -- tell us that Rob Johnson is a plus defender, who is getting better.

The pitchers -- admittedly, perhaps overenthused to have a problem removed -- universally tell us that Rob Johnson makes them better. Does Rob Johnson help Ryan Rowland-Smith pitch better? When I've reached the point that Rowland-Smith's own opinion is irrelevant, that, gentlemen, is going too far. :- )

....

Because the math doesn't tell me much either way, and because the subject-matter experts (SME's) in uniform tell me very emphatically -- I therefore, tentatively, believe that Rob Johnson is a plus defensive catcher with a plus CERA effect.

Incidentally, he looked plus to Dr. Detecto in 2010. I thought he visibly out-thought the guys at the plate on a game-in, game-out basis. I'm not saying that lightly. It's my own tools-scout report, FWIW, that Rob Johnson seems to create a Jamie Moyer "passive and confused hitter" effect behind the plate.

My guess is that, if Johnson's healthy and playing, the synergy between the M's pitchers and catchers will be improved next year. I could be wrong.

Cheers,

Dr D

Printer-friendly version

Add new comment

Filed Under

Other

Rob Johnson

Kenji Johjima

Replies

Sullivan

Sullivan is pretty specific in saying that we really have no meaningful quantitative data on Johnson one way or another. 80 games in a single season just don't give us enough to go on, so the only measurement we have is the qualitative one from his coaches and pitchers. It could be that in three years, we will be able to say with all confidence that Johnson is a net positive with CERA...or we could look back at the 80 games in 2009 as an aberration.

I can buy that Johjima had a negative impact. We heard late in the year - from Drayer and others - about his "only call for breaking pitches when the pitcher is ahead in the count" philosophy. That's a philosophy that can make a pitch selection a little too predictable for MLB hitters.

The CERA distribution

What does a statistical distribution look like if everyone is really really good at something, but some are a little better than average and some are worse?

In the case of a stat where good is 0 and bad is unbounded high, it looks like a Weibull curve. Severely skewed right with a median left of the mean.

In otherwords...statistically speaking...we should EXPECT that there are lots of catchers who are a little better than the mean ERA and a few catchers that SUCK. Johjima SUCKED. We have strong data-driven and scouting-driven evidence that Johjima was one of the worst game-calling catchers EVER (yes...really). It is far more likely that Johjima was really really bad at catching than it is that Rob Johnson was really...really good. In fact I would guess with rather high personal confidence that Johjima really was knocking 1-2 runs a game off of our chances to win based on his PITCHER-NEUTRALIZED! OPS allowed vs. his back-ups (if you're curious...I went and counted the number of batters faced by pitchers with Johjima and with his back-ups from 2006-2009...the weighted average pitcher throwing to Johjima had an ERA of 4.46...to his back-ups the ERA was 4.61...the approximate OPS split should slightly *increase* from the observed .809 to .687 split we have in our data banks when we neutralize for the quality of pitchers Johjima caught vs. his back-ups). I really believe that Johjima was so bad at catching that he may have cost the Mariners 40-80 runs a year or more on defense.

That is how I picture it, a Weibull curve

Which is also how I picture the MLB player population re: defensive impact.

40-80 runs, I think that will light Taro up LOLOL, but to the extent that's true, the 2010 M's are looking at a great year. Like Taro says, if that's the case, then bank your playoff tickets right now.

I dunno. I hope it was 40-80 just as an M's fan in 2010. But that would be a weird proportion. Baseball percentages tend to be subtle.

You're wrong about fielding skill.

Statistically speaking...fielding achievements do not follow the right-skewed weibull curve. Why? Because their performance is not bounded on the left by zero...fielding (at a position) is not governed by total runs allowed because no fielder can control total runs allowed. Each position contributes a little bit. As a result the walls are so far away from the distribution that it looks more bell-curve like even for a stack of players who are, as a group, the best in their field.

On top of that, fielding in baseball is not the only selector for who has a job.

Many players who get fielding time are not among the best fielders in the world...whereas catchers all have to be good game-callers because if they aren't their pitching staffs mutiny (see: Bedard on Johjima...instantly declaring him not a major league capable catcher)...many guys are on the team for their bat and do indeed play defense much more poorly than guys playing in A ball or even in American Legion ball.

So no...fielding skill is not a right-skewed Weibull. It's potentially slightly skewed...but much more symmetrical than CERA skill.

More symmetrical, yes

I'm speaking conceptually, comparing one SS to another SS, as I should have clarified; we're not talking about Adam Dunn here. We're comparing Cedeno to Everett.

[img_assist|nid=27893|title=Weibull curve|desc=|link=popup|align=left|width=407|height=311]Conceptually I'm talking about this kind of curve as opposed to this kind. The 30 major leaguers at SS hit that flat part of the curve, IMHO, to where it's awfully tough to get a lot better; you can't play shortstop 20% better than Jack Wilson plays it. But you could play it 5% better, perhaps.

And yet, it's possible for a player to drop off the left side of that kind of curve -- as Yuni seems in danger of doing. This is the point that you made that I was reinforcing -- that the glove specialists can hit a point of diminishing returns at the R side, and yet hurt their teams badly on the L side.

................

By the way, did you read my policy statement about rudeness? We can make our points without inciting confrontation. Your confrontational post was in response to a pat on the back.

What is confrontational about anything I said?

Unless the word "wrong" is offensive to you...I see absolutely nothing in any of my words that should be looked on as confrontational. Nothing at all.

If you'd specified that you meant at the glove-first positions...then I'd have agreed with you. In LF, 1B, 3B it's a different-looking picture because there are lots of guys playing for their bats only. At SS-2B-CF...yes there is probably a distribution of results that has some catastrophically bad performers who quickly lose their jobs and a chunk of very solid ones not far from the MLB mean.

At any rate, I was not trying to be confrontational...I was offering what I felt was an important mathematical correction to your conceptual model relating to the non-skill positions.

If you don't know, you better ax somebody :- )

But your statement of intent is noted. We all try to do as we'd be done by.

I agree that your qualification is important, that the non-defenders create a very different picture if they are part of the discussion at hand. I was thinking about Jeter-vs-Wilson, vs-Betancourt but didn't say that.

How is that helpful?

If I said something to offend you...isn't it better to take the time to explain why you felt I was being confrontational? I'm getting really really really REALLLLLY tired of the Seattle blog/forum-o-sphere telling me what a jerk I am and not telling me what I say that has them so steamed.

Blunt contradiction is simply rude

Check Dale Carnegie's "How to Win Friends and Influence People." He does a great job of explaining why it is important not to humilate other people.

............

We don't have a rule against bluntly contradicting people. Truth matters more than expediency. But there's nothing wrong with a nod towards friendliness, either.

As an abstract principle, it is almost always possible to take the other side of a question, without leaving the other person feeling like an idiot.

.............

The vast majority of your posts actually do follow Carnegie's principles. A little confusing that you should assert that you're not familiar with them. :- )

BTW, I'm not steamed at all, just trying to periodically nudge us all along the direction towards enjoyable atmosphere.

OK

I have never actually read the book, though it's on my list. I wasn't thinking of the implications of the word "wrong" when I used it in this case as being bluntly contradictory. I believed I was speaking collegiately...that's how discussions tend to go between scientists in panels at conventions. They might sometimes say something like "I might suggest you look at Pearlman's paper on..." but usually, they go right after you if they think you've made a mistake somewhere or they can offer an important clarification.

I guess the difference is the lack of face to face contact or something...because I didn't even think of the word as being combative.

I've gotta admit, I get irked

I've gotta admit, I get irked whenever I hear that Johjima "sucks" at game-calling. Joh was a positive factor in the NPB and Team Japan spanked the rest of the world with his game-calling.

Is it bad game-calling or did other factors lead to these results (or is it just a fluke)?

None of the above

It's just dissonance.

If Matt Hasselbeck isn't on the same page with T.J. Houshmandzadeh, neither one is guilty of lack of skill. They're guilty of lack of camaraderie.

Housh might be quoted saying that Hass doesn't throw the ball right, as Felix said that Joh doesn't call the game right. But Housh would then watch from the sidelines as Hass DID "throw the ball right" to Nate Burleson. And the M's pitchers can watch as Joh does "call the game right" and Japan kicks the U.S.' tails in the WBC.

.........

You've no need to get riled up about Johjima's inherent game-calling skills. Those are above reproach.

OK...clarification needed.

I'm a statistician. I use the work "suck" when I want to describe a massive deviation in the wrong direction in the player's consistent statistical performance from the average. It's entirely possible that Johjima is a good game-caller under the right circumstances. In major league baseball, however, he's a sucky one. Statistically.

With that caveat I see what you're getting at, mate

In practical terms, the net effect *was* lousy.

I mean, those CERA penalties were colossal. It wouldn't be right to brush them off, either.

.............

I'm quite sure that, catching Matsuzaka or Darvish or in fact any pitcher who was receptive to him, who let him dance the lead, Johjima would be more effective than Johnson.

But the realities of the last four years are there, too.

I remember...

...the frustration we had with Johjima catching King Felix back in '06 and '07. Every time Felix shook off a sign, the hitters knew a fastball was coming and Johjima stubbornly refused to take action to change the pattern to avoid that situation...resulting in 450 homers (while Felix stubbornly refused to adapt to Johjima's pitch mix strategy). This was also happening with Washburn...until 2009. Washburn decided after working with Adair in '09 to throw more variations on his fastball and curve and slightly increased his rate of throwing change-ups...so he was more amenable to working with Johjima on the rare occasions that they were paired. But before that, Washburn had his gameplan and got really REALLY frustrated that Johjima wasn't listening to him and adapting. And you could tell Johjima was frustrated too...because I'm sure he was probably right more often than he was wrong regarding what the smartest pitch to throw in a given situation would be but the guys weren't listening to him because they didn't have faith in those pitches. So they'd do what he wanted a few times...get pulverized because they didn't have a positive belief in the pitch when they threw it so they were more apt to mess it up, and then they'd go "SEE! My way is better!" It was just not a good situation for Kenji...and not a good situation for the Mariners.

Articulately stated

and fair enough.

::c-points::

right, the article was admirably qualified

Re-reading it, do you think I read the tone wrong?

Isn't the general implication as stated above -- Johnson isn't valuable unless you buy into his CERA, and there's not much reason to buy into his CERA?

I'm not saying that's an unreasonable position. I'm just wondering whether you think I read it wrong.

It Takes Two

Wonderful to see everyone dancing around (and on top of) the point I was planning to make. The basic unspoken assumption with EVERY stat in baseball is that it is 100% in the athletic control of the player. The discussions ASSUME that when you talk about "strikeout pitchers", that all pitchers are effectively 100% in control of the result of strikeout or not. Clearly, this is not the case, as umps have an effect, the catcher (may) effect the outcome. The defense can impact K totals, (missing outs in the OF leads to more K chances for the pitcher).

Matt's holy grail Matrix is an attempt to provide INDIVIDUAL context that accounts for all these variables. (and bravo for the attempt). But, in "general" terms, we accept that over time, given sufficient sample sizes, most of these variables even out. Pitchers don't face the Yankees EVERY series. They get to face the Yanks and the Sox and the Rangers, etc. Everyone "kind of" faces everyone -- even with an unbalanced schedule, there's enough variance that we're willing to accept that in the majority of cases the skew isn't large enough to really matter, so we pretend the influences aren't there. It simplifies the discussions. And that's okay.

But, whereas pitchers will face a couple hundred different hitters every season, of the thousand plus innings each catcher receives, he's typically giving 800 of them to about 5 guys. If you had a hitter - ANY hitter - and you said, "Okay, 80% of your plate appearances this year will be against Clemens, Maddux, Unit, Pedro, and Schilling" - (all in their prime, of course). Who is gonna bet that the hitter is going to put up a "normal" production line? He's not. Normal production lines ARE a mixture of facing cream and curdle.

So, when talking about CERA, the REAL dynamic is BATTERY ERA. As noted above, catcher and pitcher not being on the same page is *NOT* about physical ability. Heck, it's not really about mental ability. It's likely more about emotional IQ than any other thing -- but it is unavoidably couched in terms of RELATIONSHIP - rather than an individual trait. Moss was always good -- but Brady/Moss, *THEY* were special.

The big "duh" in tracking CERA is that it's following ONLY the catcher - and there is an underlying assumption that his horrid synergy with pitchers A, B and C is going to transfer in full to the next set of pitchers. Yes, it "might". But, it's certainly not a given. If you've ever worked with someone you just didn't get along with, (oil & vinegar), did that mean your next job you were going to have a similar experience? If so, the problem is likely you - not him. But, the Grinch and Scrooge both changed their perspective, (and mood), given the proper circumstance.

With Johjima, I've stated repeatedly that the cultural disconnect was the likely rotten egg in Denmark. This is supported by his wonderful results when playing with countrymen. Hence the problem in TRACKING CERA across time. In the modern world, teams swap out pitchers at an incredible pace, (except for aces). So, a new battery-dynamic is formed each time. The error is in thinking calling a good game is about pitch selection to try and fool the opponent. It isn't. (And this may draw some serious rebuttals). But, it has ZERO to do with calling curve vs. FB vs. change.

It is about knowing *THIS* pitcher. What can THIS pitcher do? What spots can HE hit? What is HIS weakness? When should you call for that? What does HE have *TODAY*? It's very, very, very personal.

There is *NO* magic bullet of the "proper" way to call a game. And while I have generally stayed out of the conversations about "calling the game backward" or the discussions on how you could "fix" things if you just changed the FB% of a pitcher -- IMO -- it's all a bunch of baloney.

I watched the Braves pitching dominate for a decade. You'd be hard pressed to find three pitchers with more disperate styles than Maddux, Glavine and Smoltz. And they ALL had success. Why? Because they pitched how they INDIVIDUALLY needed to in order to be successful. The club tailored the game calling to each pitcher -- they tailored the defensive pitching based on those game plans -- and the catchers were universally on the same page with the organization and the pitchers.

In the end, I do believe in CERA. I believe, (as a rule), it's easier to identify the guys who really suck in a given situation. But, I suspect that most pitchers, catching those quintet of main guys likely tend to have "favorites" and "others", so the aggregates tend to wash out, and the CERA becomes very noisy. I expect that when a catcher happens to become really, really good handling an ENTIRE staff -- the credit ends up being applied solely to the pitchers, so the truly gifted catchers remain hard to spot. But, when you end up with a catcher who has nearly universal disconnects, he gets a lot easier to spot. I think Johjima was such a case -- but expect he'll be wonderful, (and produce stellar CERAs) in Japan.

The important point here is that CERA isn't a static skill -- like whiff rate -- or eye ratio. It's inextricably linked to the specific pitchers you are paired with -- and to maintain horrid results across many, many pitchers is almost certain to result in demotion. So, it almost never happens.

Sandy, w/r/t Maddux

and Glavine and Smoltz, how much would you guess each pitchers' ERA's would have changed (if any) had they been throwing to catchers they strongly disliked?

Glavine - 1-K per game?

Glavine's final season in Atlanta, (first time), he went 18-11 with a 2.96 ERA. His peripherals were solid normal (for him); 0.8/HR; 3.1-BB; 5.1-K.

Next season, at age 37, he's with a new regime and catcher. He goes 9-14, 4.52 ERA (worst ERA since a rookie), and had a peripheral line of: 1.0; 3.2; 4.0. The stock explanation was Glavine got old, and had "lost it". Problem is ... he didn't. His walk rate didn't budge, but his K-rate plunged to a career worst.

In his next-to-last season with the Mets, (at age 40), he went 15-7 with a 3.82 ERA and a peripheral line of: 1.0; 2.8; 6.0. Okay, you've gotta expect SOME variation over time. But, he didn't miss a start in his entire tenure with the Mets.

How to explain the massive swoon in his first season with the Mets? Simply put, Glavine had pitched a certain way for his entire career. The Mets had a different philosophy, and different defensive schemes and plans. His pitching coach actually said in an interview the he didn't even TRY to get Glavine to change his approach for most of 2003, because he felt it would be either pointless or actively detrimental. Instead, he WAITED for Glavine to become frustrated enough with results that he became WILLING to adapt. He beat those 2003 numbers for the next 4 seasons, (even at age 41 -- barely).

Not blaming catcher - but emphasizing that the pitcher, catcher and defense ALL need to be on the same page to thrive. And this is where CERA gets really problematic. If a catcher has a disconnect with 1 pitcher, then the backup handles him -- no problem. If he has a disconnect with two, but is really in tune with the ace, no CERA problem. For a CERA problem to become visible, the catcher has to have a disconnect with the majority of his battery mates, (4 out of 5?). THEN, the numbers emerge.

Joh doesn't mesh with Felix ... you adjust. Joh doesn't mesh with Felix and Wash ... things get complex. Joh doesn't mesh with Felix or Wash or Bedard or RRS ... now you've got a major problem that is showing up in results. BUT ... maybe Joh DOESN'T have a problem with Snell and Fister and Vargas. You 'might' swap out pitchers and see the CERA problems vanish ... and the conclusion? CERA doesn't exist. It was just luck or a standard abberation (sic).

Of course, the whole discussion is rife with pitfalls. What if the pitcher is just losing it? (aging, injured, in the midst of a divorce).

The argument to dump Joh was never that Joh "couldn't" be a good catcher. The argument was that the club couldn't swap out 5 pitchers "hoping" that Joh would mesh with them ... which is especially risky if you suspect that the root of the problem is cultural, not intellectual.