Skip to content

An interesting statistical note about Yadier Molina

Nov 15, 2013, 10:55 PM EDT

yadier molina getty Getty Images

The beautiful thing about science — any field of it — is that you start out with a hypothesis and you use data to confirm or reject your hypothesis. You don’t simply accept something as fact because a lot of people believe it or because someone in authority told you it was true. Science does have its flaws, especially when performed by humans, but it’s quite a good system, one that has served us well over the years. Sabermetrics, or more generally the practice of statistical analysis, has brought that into baseball, offering us some sometimes cold truths about what we thought we knew about the game.

For years, we have accepted that Cardinals catcher Yadier Molina has a bunch of intangibles that make him better than any other catcher in the game. And it still may be true that he has them, but perhaps the impact of those intangibles has been overstated. Over at FanGraphs, Jeff Sullivan dug into the numbers expecting to find that pitchers performed much better with Molina behind the dish than with other catchers. He was surprised by what he found:

With Molina, the pitchers averaged a .746 OPS against. With non-Molina, the pitchers averaged a .757 OPS against. That’s a difference of 11 points, and if you just look at the guys who had samples of at least 500 plate appearances instead of 250, the difference is 15 points. There’s a difference that exists, but it’s hardly massive at all, and it might be entirely explained by Molina’s quality framing. For as good as Molina’s reputation is when it comes to guiding a pitcher through a game, these numbers right here suggest he’s hardly a wizard. Or, if he is a wizard, then a lot of catchers are wizards.

[...]

But maybe game-calling also just isn’t that big of a thing. By which I mean, maybe there isn’t that much of a spread, once you get to the majors. As in all skills, the majors are selective for elite ability, and calling a game is a big one. There do exist poor sequences, but big-league catchers probably won’t call them. It’s possible they’ll end up with similar approaches, especially given that they tend to work with the pitchers in advance of their games.

Molina was one of the three finalists for the NL MVP award, eventually taken home by center fielder Andrew McCutchen of the Pirates. He was barely edged out of second place by Diamondbacks first baseman Paul Goldschmidt. Molina finished fourth in 2012′s voting as well, behind Buster Posey, Ryan Braun, and McCutchen.

  1. amaninwhite - Nov 15, 2013 at 11:08 PM

    Many coaches will call pitches from the bench, so pitch selection is rendered moot when evaluating most catchers.

    • rje49 - Nov 16, 2013 at 8:37 AM

      Great piece of scientific research, but maybe it’s just a matter of the other Cardinal catchers do a pretty good job, too. You do realize that statistics can be used to “prove” just about anything.

      • raysfan1 - Nov 16, 2013 at 10:06 AM

        Always remember, statistics do not prove anything. They provide evidence for a hypothesis or they fail to provide evidence. All statistical analysis has error built into it as well, usually 5%. Thus, if one tests hypothesis A, and the test validates the idea, the statement is not “A is proven true,” it’s “The hypothesis A appears to be correct with 95% confidence.” If the hypothesis test rejects the hypothesis, it again means only that it did not stand up to that particular test with that particular margin for error–the hypothesis could actually still be true.

    • paperlions - Nov 16, 2013 at 11:31 AM

      Very few pitches are called from the bench in professional baseball. Indeed, one of the lamentations of MLB organizations is that catchers often have little experience calling games when drafted because of heavy handed HS or college coaches that insist on calling the game from the bench even though they have much less information on which to base their calls that the catcher or the pitcher. In the majors, catchers call pretty much everything except for pitch outs, pick offs, and intentional walks.

  2. Carl Hancock - Nov 15, 2013 at 11:21 PM

    Molina has averaged 137 games a season the last 4 seasons. Simply comparing the stats for his starts vs. starts made by his backup simply aren’t going to tell you a definitive picture of Molina’s impact on the pitching staff.

    The simple fact is most of the time Molina only sit’s out a game here or there. He doesn’t sit out for long stretches of time. So the impact of him not being behind the plate simply isn’t going to be felt or be able to be accurately quantified by trying to compare WITH and WITHOUT Yadi as far as OPS goes.

    Comparing a tiny sample against 135+ games is just silly. Molina’s impact is still felt even in the games he doesn’t start. Just because he isn’t behind the plate doesn’t mean he isn’t actively involved in how the game is played, how the pitchers have prepared, the research and video work that goes into preparing for each and every opponent. Do you think Molina is simply uninvolved in the games he’s sitting out? You think he’s in the locker room sitting in the hot tub drinking a mojito and smoking a stogie? Come on.

    Lose Yadi to a year ending or career ending injury and you will see just how important he is to the Cardinals organization.

    The Cardinals pitchers have 110% confidence in their catcher. While they will say the same about his backup, it’s only lip service. Yadi is the heart and soul of the Cardinals and he is a big reason why the Cardinals success… especially this past season which featured a large number of very young and very inexperienced players thrive on the big stage and in high pressure situations.

    This is a perfect example where sabermetrics breaks down and analytics simply doesn’t tell the entire story.

    It’s things like this that hurt sabermetrics in the eyes of those that oppose it. I’m not against sabermetrics, not by any means. But I think there is more to a player and a team than just pure raw numbers. There are a lot of forces at play and you have to take more than pure analytics into account when evaluating both players and making decisions when it comes to running a professional baseball team.

    • jonrox - Nov 15, 2013 at 11:42 PM

      That’s an awful lot of words to say “I’m a St. Louis fan and I disagree with the results of your statistical analysis.”

      • sportsfan18 - Nov 16, 2013 at 7:50 AM

        He didn’t just want to say that he disagreed. He wanted to explain WHY he disagreed with the results of the statistical analysis.

      • clydeserra - Nov 16, 2013 at 5:09 PM

        maybe s/he did want to explain why, but s/he didn’t.

    • cur68 - Nov 16, 2013 at 4:27 AM

      Carl, disputing statistics with rhetoric and made up numbers (110% confidence, eh? Really? You know this how?) isn’t invalidating an argument based on facts. The whole point of the post was to get to the meat of the matter: are Cardinals Pitchers noticeably better with The Angry Molina vs without? I do not think that the slight difference in OPS Against is a significant difference (without the raw data one can’t tell) but its not a difference that should be disregarded.

      Ultimately, you’d do your fave catcher a BIG service by understanding that his pitchers ARE better with him catching, just not a HUGE amount better. Factor in what else he brings (hitting, throwing out steals, scowling) then you have a more apt argument that speaks to the heart of Yadi’s stats: he’s not just a good glove, so his pitcher’s OPS against needn’t be hugely different than when he’s not catching. He can overcome pitching problems with his bat and arm better than his counterparts. So Yadi STILL wins the saber argument for why he’s the best catcher in the league and makes his pitchers better.

      • gibbyfan - Nov 16, 2013 at 8:39 AM

        Guys, In statistics sample size is everything. Period ! Carl mkes the excellent point that the author ignored.

      • sophiethegreatdane - Nov 16, 2013 at 9:05 AM

        You are right, sample size IS very important. But so is *reading the article*.

        Because if you read the article, you’ll see there does not appear to be a sample size problem, as the author of the article in question has data spanning nearly a decade, with thousands of at-bats analyzed.

        Naturally, there is more data in Molina, but the data on his replacements is by no means a small sample size.

      • gibbyfan - Nov 16, 2013 at 9:50 AM

        “……….The cutoff is arbitrary, but note that most pitchers came with much bigger samples. Chris Carpenter, for example, faced about 4400 batters with Molina during the window, and about 1100 batters with other catchers….”

        Also -Virtually all of the commnentary I heard on Molina had to do with his ability with working with a staff of almost all rookie or second year pitchers…that is an entirely different sample srt altogether.

      • paperlions - Nov 16, 2013 at 11:34 AM

        All cut offs are arbitrary.

        The only point is of the work is that you can not tell the difference in terms of pitcher effectiveness between Molina being behind the plate and other Cardinal catchers being behind the plate. That is all.

        Of course it could be that Molina helps put together the game plan and so he has an effect even when not catching, but that was the point of this cursory glance at potential effects.

      • spudchukar - Nov 16, 2013 at 12:39 PM

        The is one important caveat. Let’s say Molina is asked to sit versus crappy teams, especially offensively challenged ones. If his replacements only face inept offenses then the numbers would be skewed.

        A similar skewing would occur, if Yadi rested when aces were hurling. I have no evidence that either of these options occur, but it could make a difference.

    • macjacmccoy - Nov 17, 2013 at 12:03 PM

      You are 100% correct and 40 people have no clue. Im a Phillies fan who cant stand the Cardinals but your right. Its like people never heard of the butterfly effect. An infinite amount of everyday occurrences affect the way people perform. Unless you can replace Yadi in those games that he played with someone else, then a raw numbers comparison means nothing. Results of statistical comparisons as a way to say who outplayed who and by how much is meaningless unless those players played in the same exact game and you only use numbers from that game to compare them by.

  3. yousuxxors - Nov 15, 2013 at 11:22 PM

    maybe cardinals fans will get off his dick now.

    • brucewayne3561 - Nov 16, 2013 at 1:00 AM

      Hey yousuxors! Is that what your boyfriend says to you? Jealous?

    • brucewayne3561 - Nov 16, 2013 at 1:02 AM

      Hey yousuxxors! Is that what your boyfriend says to you? Jealous?

  4. Carl Hancock - Nov 15, 2013 at 11:27 PM

    The Fan Graphs article also has a great comment that explains another intangible that no sabermetrics stat can quantify…

    http://www.fangraphs.com/blogs/searching-for-the-value-of-yadier-molina/#comment-4236892

    It’s very true. This goes back to many people equating Molina to like having an extra pitching coach on the field. Because that’s what he effectively is for the Cardinals. He’s as close to a player-coach as you can get without actually being given that title officially.

    • paperlions - Nov 16, 2013 at 11:38 AM

      That is the point though. Yeah, it seems great to have “a pitching coach on the field”, but the numbers indicate that it hasn’t actually affected pitcher performance. Since 2004, Cardinal pitchers have been just as effective when Molina was on the field as when he was not on the field. If having a pitching coach on the field resulted in more effective pitching, you would expect to see that in the data, but you don’t.

      So….while it is nice that he can trot out there and “settle a guy down” or whatever, there is no evidence that it results in better pitching than when any other Cardinal catcher trots out to the mound.

      • stevequinn - Nov 17, 2013 at 1:26 AM

        From 2004 until early in 2011, the Cardinals had Dave Duncan as their pitching coach. Many people feel Duncan should be in the HOF as a coach. He turned many careers around in a positive way.

        There was no need for Molina to act as a pitching coach on the field during those periods.

  5. eightyraw - Nov 16, 2013 at 12:30 AM

    I believe this is the original WOWY: http://www.tangotiger.net/catchers.html

    And two early WOWY articles: http://www.insidethebook.com/ee/index.php/site/comments/with_or_without_you/

    • jwbiii - Nov 16, 2013 at 2:11 AM

      It seems that what Jeff Sullivan in doing is the same thing as Craig Wright did for Mike Piazza in the 2009 HBT Annual. Wright’s conclusion for Piazza was that he had a horrible arm but that his balance of other skills made him better than his cast of backups.

      Two years ago, I did something similar with a number catchers, including Yadier Molina. Instead of doing matched innings/seasons of a given pitcher with Molina v. not Molina, I matched innings/seasons of a pitcher with Molina v. Tony Cruz, v. Bryan Anderson v. Gerald Laird, v. Matt Pagnozzi, etc. The sample sizes were still huge, except for oddities like Nick Stavinoha and David Freese. Instead of using batting statistics of opponents, I used pitching statistics. This made it easy to subtract out the value of a catcher’s throwing arm, which in Molina’s case is significant. The SB/CS dynamic doesn’t count for OPS, which is fine. What I found was that Cardinals’ pitchers were better with other catchers, as a whole, than with Molina, after the value of their throwing arms were removed.

      • jwbiii - Nov 16, 2013 at 3:33 AM

        Checking back and reading over my comment again, it seems harsh. I didn’t mean it to be. Including Yadier Molina’s extraordinary throwing arm, he was the best defensive catcher that I looked at and he deserves all those Gold Glove trophies that he has on his mantle. Judging defensive statistics is difficult; judging a catcher’s defensive statistics is particularly difficult. Controlling the running game is easy to measure, and Molina is the best in the game at that. But what of the other components of a catcher’s defense? Pitch framing (there are people working on quantifying this), calling pitches, handling/mentoring the pitching staff, blocking pitches, fielding pops and bunts. . . I’m sure I’m missing things. If his arm is so much better, and it is IMHO, most everybody else’s and his statistics, and he shows up worse in the sum of the other facets of the defensive game then his backups, what do we have? What do we know?

    • tangotiger - Nov 16, 2013 at 9:00 AM

      Three down-votes for posting relevant links?

  6. 461deep - Nov 16, 2013 at 2:31 AM

    Molina 13 X 137 games equals 1,781 on his favorable OPS allowed differential which saves a good number of runs so this article adds favorably to his results. Also, he is more likely to be catching his pitchers poor games which mitigate against him. Pitchers are the main source of OPS allowed as the
    catcher is of course secondary unlike throwing out base stealers of which Molina does [above average. Sabermetrics must consider a players ability to impact the game in different situations in relation to their activity. Keith Hernandez is consider to be one of the best fielding1st basemen ever. But I’ll bet Ozzie Smith and Al Weiss helped.

  7. louhudson23 - Nov 16, 2013 at 5:17 AM

    Maybe they should just put a spreadsheet behind the plate next year….

  8. jss1330 - Nov 16, 2013 at 7:30 AM

    I thought Molina was praised for his work with young pitchers. The 250 batter limit eliminates most if not all rookie starters and probably 75% of all relievers. Since veteran pitchers should be able to manage the game better I don’t find this surprising. Maybe doing a side by side with rookie starters would be illuminating with the standard SSS disclaimer.

    • paperlions - Nov 16, 2013 at 11:45 AM

      They have young pitchers every year, those young pitchers become old pitchers. This wasn’t a study of 2013, but of his career. He may have improved over the years, but the contention was that he was fantastic when he first came up. So, if he was so extra-ordinary, a signal should manifest regardless. One didn’t. This, of course, doesn’t mean he is not fantastic, just that this cursory approach couldn’t find evidence of it.

  9. sfm073 - Nov 16, 2013 at 8:38 AM

    A real scientist would call this quackery.

    • paperlions - Nov 16, 2013 at 11:49 AM

      I’m a real scientist, and I would call it a preliminary analyses that didn’t provide any evidence to suggest that Molina’s skills result in better pitcher performance. Again, the assumption is that Molina is other-worldly awesome at being a catcher, as a Cardinal fan, I love having him back there. If he is so wonderfully fantastic AND if that awesomeness results in better pitcher performance compared to backup catchers (not even other starting catchers but to backup catchers) then even this cursory evaluation should evince some signal of that awesomeness. Such an analysis would be sufficient to quantify the strength of the signal, but it should be able to identify its existence. The results suggest that if a signal does exists, it is small and not worth very many wins/year above replacement level catcher defense.

      • fearlessleader - Nov 16, 2013 at 1:40 PM

        Hey Paper—you keep saying that this article compares Molina to “backup catchers,” but I’ve re-read it a couple of times, and it seems to me that Sullivan is comparing him to ALL other catchers (not just to non-Molina Cardinals). Thus pitchers who have pitched for other teams would be counted using their stats with other first-string catchers, and Cardinal pitchers from early in Molina’s career would be counted using their stats with Matheny, among others. Am I misunderstanding?

      • paperlions - Nov 16, 2013 at 3:34 PM

        You are correct, it is compared to all other catchers for the same pitchers. However, much of the non-molina catchers would still be cardinal backups…not that it necessarily matters as most back up catchers exist for their defense, not their offense.

    • stex52 - Nov 16, 2013 at 2:15 PM

      I’m a real engineer with a degree in the sciences. It’s not remotely quackery. It was a hypothesis for explaining a perceived phenomenon. It didn’t show much. That applies to 99% of all research science. Personally, I think his model is just too simple.

      Or his counter idea could be correct. All catchers can call a decent game by the time they get to the MLB or they are not allowed to face a batter 250 times. I am speculating, but it seems to me that there is a random element to calling a game. The hitter is also guessing the next pitch, and a smart hitter also knows the kind of pitch a smart catcher is likely to call. A couple of other suggestions came up:1. Mike Matheny was one of those other catchers. He was an exceptional pitch caller. 2. Molina might have improved over the course of his career and this study wasn’t designed to pick it up.

      The only thing you can say is the researcher picked up a simple hypothesis and the hypothesis did not pick up a strong correlation. Doesn’t mean it isn’t there. It just says this study was not sufficient to pick it up.

    • cardsman - Nov 18, 2013 at 12:23 PM

      thats the most scientifically accurate comment that has been made. sabermucktrix sux

  10. gregbeau - Nov 16, 2013 at 9:28 AM

    I wonder how much better he would be without that neck tattoo.

  11. watchfullhose - Nov 16, 2013 at 9:55 AM

    And he gets a neck tattoo for every fight he gets into in the field

  12. macjacmccoy - Nov 17, 2013 at 11:46 AM

    Jeff Sullivan dug into the numbers expecting to find that pitchers performed much better with Molina behind the dish than with other catchers. He was surprised by what he found:

    People use stats like this all the time but they mean nothing. Each situation is different from the next. It doesnt matter if pitcher A actually did better with catcher b then catcher a. Even if they were playing the same team with the same players and doing exactly the same thing, because there’s no way of knowing if catcher b played instead of catcher a on that exact day if he would have done better or worse. Each game is a brand new experience. Weather, stomach aches, runny noses, the amount of sleep you had, how early or late you got to the stadium, etc, etc, etc, are all things that change the dynamic of how each person will perform on that given day.

    So unless you have a time machine and go back in time and replace catcher a with catcher b in every game they played, then you will never be able to compare a given players performance to another’s accurately.

    Its like playing a slot machine and then walking away and then the next person who plays it wins the jackpot 3 turns later. Just because that person won doesnt mean you would have. The factors are different, in slot machines its the odds built into the system, in baseball its the effect of the entire world on every player that plays.

    And even though its a fact they cant, just like how people believe there is such a thing as a slot machine that’s ready to pay out even though its nonsense, people believe that there is such a thing as an accurate comparisons of players using their performance in similiar but totally different situations.

Leave Comment

You must be logged in to leave a comment. Not a member? Register now!

Featured video

Three legends off to Cooperstown
Top 10 MLB Player Searches