Games » Toronto Blue Jays
Jun8Why Ned Yost won't pinch-hit for Alcides Escobar
Lee Judge
None
Ned Yost says he’s not going to pinch-hit for Alcides Escobar no matter how much fans want him to. Ned is thinking big: He believes Escobar is the best defensive shortstop in the game and the Royals will need him on the field when they’re playing for a championship.
Yost believes that if he starts jerking Escobar out of the lineup now, Escobar never will learn to hit in those situations. The more at-bats Esky gets, the faster he improves, and Ned believes he will improve. Ned is looking at the big picture, and I’m pretty sure he doesn’t care if the rest of us can’t see it.
The problem with numbers
I don’t want to get in a protracted debate about advanced metrics (so naturally I’m writing something that might start a protracted debate about advanced metrics). I thought I got bogged down in a nonproductive back-and-forth last season over this subject and wanted to avoid that this season. But then a reader sent me a link to an article about Alcides Escobar and how bad his numbers were offensively, so I figured I should explain where I’m coming from on this issue.
If you want the short version of what I think, here it is:
People who enjoy advanced metrics should go right ahead and enjoy them. They can be useful and tell you interesting things about the game. Despite using a “system” and “points,” I’m not doing anything like advanced metrics here. I’m using a very old system that was put together by an old-time baseball guy and designed to measure contribution to a team’s success and reveal patterns of play. If what Ron Polk’s system reveals is helpful to you, great. It’s made me think about the game in new ways. If you don’t like it, ignore it and try to enjoy the other aspects of this website.
Despite having said that on many occasions, there still are people who insist on bringing the metrics argument to this site, insisting that the way I’m looking at the game is incorrect. Of course that means Ron Polk doesn’t know much about baseball either, and one look at his resume tells you that’s not likely.
So, why don’t I wholeheartedly embrace advance metrics and forget this Ron Polk stuff?
UNSEEN FACTORS
My first exposure to advanced metrics was an article by Bill James about a player’s numbers on grass versus AstroTurf concluding that the player was a “grass” player (or it might’ve been AstroTurf … can’t remember … I was on grass) and that he should play for a team that provided the surface that would give him the best chance for success.
Even then, I thought, “Who was pitching?” If you face me on grass and Justin Verlander on AstroTurf, you’re probably going to look like a grass player.
The same thing happened early last season. I wrote that Rick Ankiel was a very good outfielder, and an outraged reader talked about Ankiel’s range factor versus the range factor of Mitch Maier. Which left out the fact that Rick had put together his range factor while playing for the Cardinals, with two Cy Young candidates, and Mitch had played for the Royals. Because of the difference in pitching staffs, Mitch was getting more balls hit to him.
One night I watched Jason Kendall give up three stolen bases, but he wasn’t responsible for any of them. On the first one, the pitcher fell asleep and the guy was halfway to second before the pitch was delivered. The second one was a blown rundown in which Jason never touched the ball. The third one happened with runners at first and third and the Royals had a play on that had Jason feinting to second and looking to third. So statistically Jason had a bad night. The stats didn’t record reality.
One more example: Vernon Wells of the Angels has a new centerfielder playing next to him, Peter Bourjos. Bourjos has great speed and is getting balls that used to be in Wells’ territory. Being the centerfielder, he gets to call Wells off. So through no fault of his own, Wells’ range factor probably will go down this season and someone will write about him losing a step.
I asked a friend of mine, Mike Keefe, to take a look at some of the baseball metrics and tell me what he thought. Mike has a B.S., M.S. and all but the dissertation for a Ph.D. in mathematics and taught college-level probability and statistics at the University of Missouri-Kansas City and Penn Valley Community College. (Mike also quit math and became a political cartoonist, so you might take this with a grain a salt … although he did just win a Pulitzer Prize, which annoys me to no end.)
Here’s part of the email I received back from him.
Godel’s Incompleteness Theorems (paraphrased):
No system of logic that is complex enough to contain within it the axioms of the natural number system (or the logical equivalent) is capable of proving all facts within that system. In other words there are true statements within the system that are unprovable.
In any such system of this complexity, the consistency of the system itself is unprovable.
(See? I told you he was annoying, but here’s the problem expressed in plain English, once again in the words of Mike Keefe.)
A Deterministic Universe:
It was once believed that if a person could know the state of the universe at one particular instant — all the forces at work, all the conditions present —then he could predict with certainty the state of the universe in the next instant. Also known as the “Clockwork Universe,” it was frequently cited to support the idea that God set the universe in motion, with all outcomes predetermined. No free will. On several philosophical grounds and based on what we know about quantum mechanics, the idea of a deterministic universe is hogwash. It’s basically flat-Earth thinking.
(OK, here’s the key line.)
Any predictive system that chooses a limited number of factors from the nearly limitless number available, and weights those factors arbitrarily, is at best a flawed system.That’s why in statistics you hear the terms ‘margin of error’ and ‘confidence level.’ (When the media quotes a statistician’s poll results, for example, they always include the margin of error. But they don’t often quote the level of confidence, indicating the reliability of the estimate.) No system can predict complex outcomes with absolute certainty as if the universe were deterministic. Informed best guesses are all we have.
(See: http://en.wikipedia.org/wiki/Confidence_interval)
And that’s brings me to my next problem:
PSUEDO-SCIENCE
When I asked how it was determined that Yuniesky Betancourt was the worst shortstop in the history of mankind, I was told about a system in which the field was divided into sections. I said, “Stop right there. I don’t see any lines on the field. How are they determining where the sections are?”
They eyeball it.
Then the person recording the statistic determines what kind of ball was hit into that section, pop up, line drive, etc. and I said, stop right there. How are they determining what kind of ball was hit? I don’t see any radar guns are laser beams measuring speed or trajectory.
They eyeball it.
OK, so how is that any different than an old-time scout with a chaw in his cheek saying, “Tough play, but the kid can go get it”? It’s not. Both people are eyeballing it, but one attaches a number to his guesstimate and it seems more scientific. (And if that’s not really how it’s done, please enlighten me.)
MISUSE OF NUMBERS
Someone might say (and did), “How could the Royals lay down a sacrifice bunt? Don’t they know sacrifice bunts make it less likely to score a run?” Well, first of all, I wouldn’t trust any number until I see how it was determined, but let’s grant the argument that sacrifice bunts make it less likely to score a run.
In all situations?
Let’s say a move succeeds an average of 40 percent of the time. Focus on the word average. That means sometimes it has worked more than 40 percent of the time and sometimes less. The people and conditions involved can change the success rate. But people will grab that number and act like it works 40 percent of the time in every situation. Coaches do what they do because they know a lot about the people involved and their chance of succeeding.
The other thing to think about: If a move works 40 percent of the time, but the other options work 30 percent and 20 percent, the 40 percent move is the right move … even though it will fail most of the time.
AND FINALLY: ATTITUDE
(If I haven’t made you mad yet, this should do the trick.)
The people who play and coach this game for a living are generally pretty humble. Clint Hurdle, the manager of the Pittsburgh Pirates, says he will never know the game. You hear the same thing from pretty much everybody out here (and if someone thinks they’ve got it all figured out, they soon find out differently). The game is so complex and varied that something new always arises. Every game is different, and everyone needs to be in learning mode because you might figure something out in a meaningless game that helps you win a bigger game down the road.
In my opinion, too many of the people involved in advance metrics deliver their opinions with a sneering, “Boy, I’m so much smarter than everybody else”attitude. Why is that necessary? If you’ve got an interesting statistic or new way to consider things, why not let that stand on its own?
When the guys at Baseball Prospectus title their book, “Baseball Between the Numbers: Why Everything You Know About Baseball is Wrong,” it throws up a wall between them and the people who play the game for a living and fans who might otherwise be interested.
Really? Everything we know about baseball is wrong? There aren’t three outs in an inning? Four balls don’t entitle the batter to a base? A title that said, “Hey, we’ve looked at the numbers and found some really interesting things!” would probably be better received and allow everyone to approach the new information without resentment and anger.
OK, that’s it. I’ve said what I have to say (at length) and realize not everyone is going to agree with it. Here’s hoping we can all approach the game with an open mind and treat each other with respect.
If not, you can leave a comment and tell me I’m a jerk.
Why you should continue to read this website
OK, let’s say what I just wrote made you angry and you really disagree with me. Why should you continue to read this website? Well, we’ve tried to create something unique here. I am not doing statistical analysis (despite what the Star says in the promos). I’m recording patterns. You can find out who has taken an extra base the most times, what pitcher has walked the most leadoff batters or who tends to strike out looking.
The more they play the more the patterns reveal themselves and you get a better idea of how these guys play the game.
The other thing that’s worth your time is the thoughts of the players themselves. What they were thinking at the time a great play or a mental mistake happened. You can also ask a question and sometimes I can get an answer for you, straight from the horse’s mouth…or if you ask it of me, straight from the other end of the horse.
Anyway, I know that some of you, a lot of you (most of you?) might disagree with what I have to say about metrics. That’s fine. To each his own. Even if we disagree on that subject, I hope there’s still enough interesting content on this site to keep you coming back.
Butler
Cabrera
Hosmer
I never thought of the defensive range like you explained it... they eyeball a "zone" then say their opinion if the player should have gotten to it or not. I think there are a lot of people that want to look at a chart and say "See Betencourt does have terrible range here are the numbers to say so" or using some other numbers to say a hitter was bad or what not. Those numbers do forget things. Like you said... they don't count in extra bases taken, 8 pitch at bats, hard hit outs, and the other various things you discuss and put into your grid. Any sabermetric lover can go to those websites and look up the numbers quickly themselves. Most of us enjoy what you are doing and love hearing the stories from the players and from the ball park that we wouldn't have access to other wise. Keep doing what your doing Lee(or Mr. Judge whichever you prefer haha).
I enjoy reading this website because of the interesting things you bring up about the game, not because 3 out of 10 times you deliver a well-written article. The thing I hate about numbers is that they always make sense, until they don't. This guy is 0 for 100 in his career--oops he hit a game-winning homerun. What would the advance metrics people have to say if Ned Yost always pinch hit for Escobar? There would be no data! Yes it's great to know that on Saturdays during games that start after 3pm, on grass cut 3/4 inch and a sprinkle of dirt on home plate, Player A is likely to fly out to the recently called up prospect in right field foul territory with a spectator reaching in his glove...but honestly, who cares??
Wow, a 2000 word post on advanced metrics, I thought I was reading Posnanski's blog for a minute there. Of course the Poz would never disparage the numbers so I wasn't confused for long. I read Joe's blog fairly regularly but lose interest quickly when he goes to "Bill James' World". I really appreciate the simplicity of the Polk system. It's nice to be able to glance at a straight-forward chart rahter than busting out mathematic formulas and an acronym decoder. Thanks for the great work you've been doing. A game-day doesn't go by without me checking in to catch your analysis. (I didn't mean to be hard on Mr. Posnanski. If it wasn't for his many years of superb Royal's coverage I wouldn't be nearly the baseball fan I am now.)
I was hoping to read something about the IBB to Bautista. I don't agree with taking away the R-R matchup to go with the R-L matchup. Both Bautista and Lind are power hitters and I would have liked to have seen Adcock pitch to Bautista instead of giving Lind the opportunity to hit the Grand Slam.
I wouldn't say that I'm obsessed with all the sabermetrics, but I do find them interesting. My wife could care less about WAR, BABIP, or UZR. However, she simply loves reading about the Polk system because it just makes sense. It was either a mental mistake or it wasn't. It was either a normal play, an error, or an outstanding defensive play. Good stuff.
Sorry to drop a term paper on everyone. That piece had been sitting around waiting to get used and I decided last night was the right time.
It's just that once in a while someone will drop a 'don't you know about his UZR rating?" comment on me and I figured sooner or later I should comment on that.
I don't know anybody in the game who completely dismisses numbers. They just think they need to be taken with a grain of salt and a big dose of observation.
Generally, I don't think people come to this site for my thoughts on metrics though. What I'm trying to do is bring the way the players think about the game to baseball fans.
That's one of the reasons I like Ron Polk's system also. It's the way a great baseball coach thinks about the game and it gives us a look at baseball through his eyes.
So now that I've got that out of my system, I'll try to get back at what I should be doing anyway.
Joshua: Yost thought he was between "a rock and hard place." He could pitch to Bautista "maybe the best hitter in the American League" or Lind "one of the best 15 hitters in the American League." (The quotes are from Ned.)
Apparently, Adcock has got good numbers against lefties, so Ned was more comfortable with that matchup.
The real problem was walking McCoy. That painted the Royals into a corner and forced them into those bad choices. We often focus on the last thing that happened and lose sight of the previous thing that made the last thing happen. (Does that sentence make sense?)
Like, "Why did he throw that pitch 3-2?" when the real problem was what he threw 2-1 to paint himself into the 3-2 corner.
If I hadn't been going all Bill James on you guys, I would've mentioned the three walks that scored in front of the home runs.
Throw strikes to those guys and you probably win.
Erin: 3 out of 10 articles are well written? That's a .300 average! So you're saying I should be in the Hall of Fame?
I couldn't really tell what side you're on or if you want to be on a side (do we all have to take a position on everything?), but thank for reading the website. I hope you continue to get something out of it.
Patrick: It's 'Lee' and I'm glad you're enjoying the work I do.
P.S. Can you guys tell I'm not going to work this morning (day game) and I'm sitting here responding to you one by one to increase my page hits?
Tom: Love Posanski to death, he's a good guy, but he once wrote that after seeing some advanced metrics that batting average really didn't matter.
I remember thinking, OK, two down, bottom of the ninth, tying run on third, winning run on second, first open and a pitcher who's not going to walk anybody...does batting average matter now?
Blanket statements tend to get you in trouble. (And if you're going to let a political cartoonist get away with saying that, you're missing a good chance to make fun of someone.)
I've been hanging around at Royals Review a little and trying to learn some of the nerd stuff and do agree with you about the attitude, and would add that the stats are historical, rather than predictive, and tend to lag current realities, and also don't seem of much use for weighting the value of defense, which is half the game. Statheads love Billy Butler, who plays no defense, drives in few runs, scores few runs, and probably runs slower than either of us, but the numbers say he is a great player because of high OBP and he hits lots of doubles. Escobar, on the other hand, looks to be among the handful of the worst players in the game by the same stats, even as the eyeball tells us that he saves more runs with his glove than Butler creates with his bat.
Of the positives from the stats, I do have a greater appreciation of OBP, which tells us why Matt Treanor and Chris Getz seem so often involved in late rallies, even with low BA, and I've come to accept that OPS may be the most useful single offensive stat. What I don't agree with is that career history has much to offer for tonight's game, preferring more recent ranges of performance for filling out a line up card. Good post, Lee, thanks.
Tyler: Nick Scott, who writes for the Royals Authority website and probably disagrees with what I've said, but nontheless is a good guy (hang on, I'm wandering near the point of this sentence), said he likes the "Quality Start' stat for the same reason: it's understandable.
It's not perfect, but it let's you know how many times a pitcher keeps his team in a game.
When stats are so convoluted, even if they're accurate, you pretty much have to take the numbers on faith, which I'm unlikely to do. Faith is defined as 'belief without evidence' and that's a pretty scary territory.
Mike Keefe (the math whiz quoted above) made a very interesting point: a lot of the metrics were trying to be 'predictive' (tell you what was going to happen), Ron Polk's system is 'descriptive' (tells you what HAS happened).
So once these guys play for a while, you can figure out who's running the bases well or making mental mistakes.
I wouldn't take what I'm saying to the bank either. Some of it's subjective, but if a guy is crushing a category, it's probably not a coincidence.
I don't need any numbers(advanced metrics or batting averages) to tell that Esky struggles with the bat or that Hosmer doesn't. Numbers don't tell me that Dyson can run and Butler can't. I think you've said that if you'll take your eye OFF the ball, you might really see what's going on out there.
Jim: Good point. Yost made the same one last night. Esky's hitting .200, if Ned has a .300 hitter available on the bench, while the odds are better, that's still only one more hit in 10 at-bats. There's no guarantee it's going to happen.
And the numbers will change depending on the people involved. So looking at a number that reflects the past history of EVERY situation encountered may have little to do with what should be happening tonight.
I just discovered your site this year and now am a faithful daily visitor, because I really enjoy the unique insights into the game that you provide from your conversations with the players and the coaches. With respect to today's column, as someone with a Master's degree in Operations Research (trying to solve real-life problems with mathematically-based techniques) who has been plying his trade for a few years now, I whole-heartedly agree with you that advanced statistics have their place (they tell a part of the story), but they are not predictive, not always reliable (see your Betancourt example), and they can be misrepresented in order to tell the story that someone BELIEVES (based on gut feelings) to be true.
THIS article, for some reason, perfectly explains why I have greatly enjoyed reading your blog this year. Excellent stuff, and this is my first article to read each day when I come to the sport's section.
While numbers can be a great painter of history, they can never tell what is going on (chemically) inside a player's head. While Butler may hit ball after ball over the wall during BP, his results will vastly differ during a game even with the same pitch thrown due to "what he's thinking". If he tries just 5% harder to swing, (even if he doesn't mean too; just the game pressure) he might crush it, or roll over the top of it or be slightly under it. While you will have another number to add to the mathematical system, it still won't add to the future...
Anyway, I completely enjoy this blog and thank you for taking the time to enlighten us regarding our Royals!
Love your page Lee. Personally I ride both sides of the fence when it comes to stats versus what scouts see, so to speak. While stats are a useful tool, they simply can't and don't tell the entire story for reasons that you already mentioned. Those that are wise will use both. Obviously the Royals do.
Too bad most of corporate America relies almost exclusively on bean counting (stats) these days. No small wonder we're getting our butts handed to us in a global economy.
Thanks to everyone, I was hoping to write something reasonable and I'm pleased to get such reasonable responses.
I'm probably in the same boat as everyone else: you need both numbers and observation. Of course some of us are sitting in the middle of the boat and some of us are hanging over the side, but learning to share the boat with others makes the journey more pleasant.
(Wow, wore the hell out of THAT metaphor, didn't I?"
Great column Lee. Love the comments that resulted. Shows that this site is really for the fans who care. Keep on giving us opportunities to explore (both positive and negative) what is happening on the field.
On the Ron Polk system -- do you also grade the opposing team? I'm curious about the relationship between the team that grades higher on the Polk scale and the team that actually wins the game. I'm metrically challenged and have typically only looked at the runs scored vs. runs allowed in the standings to see how teams are doing. Thanks.
The only stat that really matters is W-L on May 5 the Royals were 17-14. Since they upgraded to Hosmer at first they are 9-22 for a winning pct of .290 or 47 wins in a 162 game season. How can this be explained?
Uh, wasn't there a baseball game yesterday that should have gotten a write-up ? I appreciate the statistical breakdown and all, but got enough of that from your buddy back in school (yeah, I was a student). As an analyst, you can take any set of numbers and slant them the direction you want. As Lucy once told Charlie Brown...."tell your statistics to shut up!". I like what Paul just said...W-L is what counts.
I do enjoy this season's blog much much better than last year. You lend interesting insight with a sense of humor that, obviously, people are catching.
Your "problem with numbers" post is exactly why I like your perspective on baseball in general and the Royals specifically. It's helped me appreciate baseball in a new way.
Joel: Thank you. I hope the site has improved this year as I've figured out where to concentrate my efforts. I think the conversation with readers has also improved the site.
Pat: No, I don't do this for the other team so you won't be able to compare. Sorry, but it's pretty much killing me to keep up with the Royals.
Paul: Maybe if you don't have Hosmer's .310 average and Gold Glove-level defense at first they lose even more games than they have since his arrival. If you're looking for a culprit, I'd suggest looking at the starting pitcher's ERA during that period. Picking one factor out of all those involved in those games and assuming that's the reason you got the results you got, seems like a good way to jump to a false conclusion.
Scott: I apologize for doing something unexpected. I'll try to be more predictable in the future.
I have about the same level of math training as your expert, and I have dabbled about with SABRmetrics over the years, and I have largely the same opinions. The key, for me, is making the proper qualifications to the work. That is why Bill James is, to me, still the gold standard. He is very careful about the design of his data, and also very careful about his conclusions.
In particular, on defense, I agree that is you look at any one play, then you may have a hard time knowing what qualifies as a line drive verses a fly ball, for example, or whether it is in this zone or that zone. But they are looking at every single play. They are looking at every single pitch. And, with defensive metrics in particular, we aren't very good yet. And I think everyone who builds the stats are clear about needing multiple years worth of data to have a sample size large enough to have much confidence.
Even if I knew everything available to an outsider - followed every stat imaginable, knew who was a grass player and who was a turf player and why Brian Bannister always did better during day games - there is always information we didn't know. Maybe the numbers say Maier should be pinch hitting for Treanor in this situation, but I didn't know he was running a low grade fever that night, or that he drilled a foul ball into his ankle during batting practice, or who spent last night with a sick kid in the emergency room. So I have no doubt that the manager always has access to information I don't know. The manager always has a reason; I may not agree with the reason, but we should always start with the premise that there is a reason behind what is happening.
Amen Curtis. It is always easier to go back after the fact and say "you should do this or that". The inside info plays a big role in the decision-making process, and you can't replace it.
I agree with your sentiment on the attitude of SABR nerds, but think the same can be true of the folks who completely spurn the numbers in favor of the scout's opinion. Sometimes when I talk with those kinds of fans, I feel like I'm trying to explain evolution to an evangelical. They just don't want to hear the numbers, even when they make a lot of sense.
All that said, I think I'm with everybody else here -- neither side is capable of telling the whole story. We need the data in the box score to back up the scouts' opinions, and vice versa.
Curtis, Nick: Excellent input. I think we're all in agreement, you need both. Curtis makes an excellent point when he says there are factors you can never know about that change the situation.
There's nothing wrong with numbers, it's jumping to false conclusions from numbers that causes problems.
It's entertaining as always, Lee. I would have enjoyed your perspective into the game that was played, however. btw....there is nothing predictable about you ! lol