Apr
07

Final review of The Fielding Bible

By

The subject of defense is a hotly debated one in baseball analytic circles. For years we were stuck with just errors and thusly judging fielders on their ability to cleanly field batted balls. As statistical analysis has progressed, so has the measurement of defensive ability. These new defensive metrics — UZR, ZR, the plus-minus system — are interesting to consider, but they don’t give us the same level of insight as offensive and pitching stats. Yet they are interesting because they are not perfect. We’re still trying to figure out the best way to measure defense, and imperfect stats make great conversation starters.

Last week we discussed some Yankee-related issues from The Fielding Bible Volume II, John Dewan’s updated effort at pegging defensive ability. Using stats, articles, and scouting reports from the book, I was able to piece together articles on Robinson Cano’s range and Derek Jeter’s clean fielding skills, but those clearly aren’t the only cases you can make after reading through the 400-page tome. It contains all sorts of stats and leader boards that can provide varying levels of insight on defensive ability.

The volume can keep a fan occupied for days, even weeks. It’s filled to the brim with data, from raw numbers to processed stats. Dewan rates fielders on all sorts of criteria, and most of it is presented in these pages. How well they ranged to their left and right, how they turned the double play, how they fielded when playing shallow, deep…it’s a truly comprehensive look at a player’s fielding ability. Even if you want to know how many outs a player recorded in 2008, The Fielding Bible has that.

One of the most interesting parts of the book, which we didn’t get to touch on in the Jeter and Cano discussions, is the Defensive Positioning Chart. As Dewan notes, these are not spray charts. Rather, they represent the distribution of hits which can be reasonably turned into outs. That, of course, requires a level of subjective judgment, a complaint I’ll address in a minute. For now, let us revel in these colorful charts which depict a player’s hitting tendencies. One of my favorite charts is of Joey Gathright. The charts make a case for an Ortiz-like shift for the speedster, but to the opposite side. He almost never hits the ball down the first base line either to the infield or outfield. (The problem, of course, is that he’s so fast that the first baseman can’t stray too far off the bag.) According to the chart, if a manager plays his second baseman pinching towards second, his shortstop playing slightly toward second, his third baseman slightly to the left of normal, his left and center fielders shallow, and his right fielder well off the line, he should be able to contain Gathright considerably. It’s really neat stuff both from an analytical and visual perspective (The colors, children! look at the colors!).

Another amazing aspect of the book is the six-year register. As the name implies, this section contains data for nearly every Major Leaguer over the past six seasons. As Dewan explains it, this helps mete out poor single years in favor of a more comprehensive, thus more accurate, view of the player. For instance, he says, Mark Teixeira had a pretty poor season with the glove in 2007, but when you look at his six-year marks he still emerges as one of the best first basemen in the league. This is one aspect which will help advance the accuracy of defensive statistics. We’ve only started to seriously measure them recently. Larger samples will do wonders for analysis.

The book isn’t without its faults. While Bill James’s chapter on Defensive Misplays and Fielding Good Plays was interesting, he’s still using some subjective judgments. That’s not wrong in itself — we make subjective judgments about baseball every day and sometimes we’re right — but when the goal is an objective system to judge a fielder’s ability to make a clean play, any hint of subjectivism detracts from the case. Still, I like the idea of stripping defensive misplays (the act of misplaying a baseball, not James’s stat) of as much subjective judgment as possible. We all have our own ideas of what “ought to have happened,” meaning five people making the judgment could call it differently. As such, we either need to qualify those who do make the judgment, or else need to strip the subjectivity out as much as possible. James’s effort is certainly commendable, and I hope in the future he’ll not only release a more comprehensive list of his results, but also go deeper into his methodologies.

Through the efforts of curious minds like Bill James, the baseball analysis community has evolved over the past three decades. We now have not only a better idea of which basic stats tell us a better story about a player (OBP, SLG), but some great minds have developed advanced statistics which put them into an even greater context (wOBA, EqA, VORP). Similar efforts have been made to harness pitching stats. We’ve learned that pitchers have little control over what happens to a batted ball, hence the focus on peripherals like strikeouts, walks, ground balls, and line drives. These help us not only determine which pitchers have better and more sustainable abilities, but also to help us construct a more complete story of the player. The hope is that over the next few years we make this kind of breakthrough with defensive stats. So observe, analyze, criticize, scrutinize. It’s how we come to new conclusions about previously unknown issues.

You can get The Fielding Bible–Volume II from Amazon.com for $16.29. That’s our Amazon Associate code, so if you buy the book from that link you’ll kick us a few pennies.

Categories : Reviews
  • andrew

    I think it’s an interesting point you make on Gathright, however the difference between him and guys like Ortiz and Giambi is that he could lay down the bunt and beat that difference 75% of the time.

    • andrew

      defense*

  • http://www.newsday.com/media/photo/2009-02/45088362.jpg Mike Pop

    Great work, Joseph. This series has been interesting.

    • http://twitter.com/tsjc68 tommiesmithjohncarlos a/k/a Ridiculous Upside

      Great post, Joseph. Compelling, and rich.

    • http://actyankee.blogspot.com Matt

      Word.

      You know what else was great work? Me beating you, Mike Pop, in the RAB Bracket Buster pool.

      • andrew

        hahah damn you both. I came in 6th, despite being 1/4 on the final four.

      • http://www.newsday.com/media/photo/2009-02/45088362.jpg Mike Pop

        Bitch, why don’t you go with your heart!!! I was telling you how UNC was going to take down UConn, all I heard from you was Adrien will take him down. Haha, congrats man.

        /walks away pissed off

      • http://www.youtube.com/kevyyankees Kevin G.

        You remind me of the guy who won the RAB fantasy football tournament.

        • jsbrendog

          burn

  • viridiana

    “We now have not only a better idea of which basic stats tell us a better story about a player (OBP, SLG)”

    What nonsense. These are the trendiest stats and in many ways misleading. Totally underestimate the role of speed and advancing runners, equating walks with hits and totally ignoring (even claiming the non-existence) of clutch hitting. Are they legit stats? Yes. Are they more revealing than a whole bunch of older metrics — not to mention the superior judgment of trained scouts? No way.
    This blog would be so much more interesting if you guys abandoned your search for the statistical Holy Grail and trained yourself to see the sorts of things scouts see.

    • Moshe Mandel

      You know what scouting is? A total crapshoot. There are tons of rounds in the draft, and every so often they hit on a sleeper. More often than not, prospects fail, because scouts are guessing based on tools. In regard to stats, why is it not possible that newer metrics are more revealing than older ones? No one said they are perfect, just that they tell a more contextual and comprehensive story.

      • Moshe Mandel

        Let me just add that advance scouting, which is a totally different thing, is not rocket science either, and I do not think anyone here at RAB advocates not observing players and just using numbers.

    • steve (different one)

      OBP is “trendy”?

      AOTIBI

      • http://www.youtube.com/kevyyankees Kevin G.

        AOTIBI?

    • andrew

      i dont think this was a joke, but let me just throw out an ietc anyway, just in case my sarcasm meter is failing me

    • Stephen

      I think you’re grossly misrepresenting the views of the authors of RAB. They all rely to a different degree on stats to make their point, but none of them have advocated using only statistics to evaluate a player, and in-fact they do make their own scouting reports. They may not be trained ‘scouts’, but they do watch a lot of baseball.
      Anyway, you also misrepresent sabermetrics as a whole. The point of sabermetrics is to get a more accurate assessment of players and looking at all the factors that lead to run scoring, winning, etc. Nobody (or almost nobody) advocates using only stats, but there ARE better stats.
      Also, the view is not that clutch hitting does not exist, but that it is either doesn’t exist as a repeatable skill or very few players possess that ability. That is also not to say that being unclutch doesn’t exist, only that very few players are demonstrably unclutch.

      • Moshe Mandel

        Exactly, Tom Tango has found that clutch does exist, just not as a statistically significant element of the game.

      • http://www.riveraveblues.com Benjamin Kabak

        Just to further Stephen’s point, the difference between a “trained” scout and any us who watch hundreds of baseball games a year is rather minimal. Scouting isn’t rocket science.

    • jsbrendog

      two things.

      1, it is obvious that you are seeing firsthand the tarnished socialist reputation of this site and its authors.

      2. a pitcher for my college, who was not very good, was drafted in like the 17th round or something. did i mention he wasnt very good when he was at my school? oh, and we were and are a small school where no athletes come from and our baseball team wasnt really that good and were division 1aa i believe. so this pitcher was panned, not very good etc etc. now he is in the bullpen for the a’s and threw a scoreless innning yesterday. obviously he sucks since he didnt strike out the side.

      but i guess those scouts who didnt think he was very good and such were right then.

      • http://www.riveraveblues.com Joseph Pawlikowski

        Brad Ziegler?

    • http://actyankee.blogspot.com Matt

      These are the trendiest stats and in many ways misleading.

      If I’m not mistaken, OBP and SLG have been around for a really long time and have, numerous times, been shown to be the two factors most conducive to run scoring.

      Totally underestimate the role of speed

      The role of speed is obviously important but there’s a difference between running willy nilly and doing it intelligently. Bobby Abreu stole 22 bases last year. That’s a good mark, right? But he was also caught 11 times–awful. That’s a 66% success rate which is bad. No one says we want slow players on our team. Rather we want guys who will break the 75% mark because that’s been shown to be the “break even” point in terms of run scoring. Why are the Phillies such a good base stealing team? Not because they run a lot, but because they’re seldom caught.

      advancing runners

      Strategies used to advance runners almost always include making an out. Making an out means your team just brought itself closer to the end of the inning and the end of the game. Even moving a runner to third can lessen your chances of scoring if it comes with making an out. Want proof? Go tinker w/the WPA calculator or look at a run expectancy chart. And before you get all up in arms about those things as arbitrary or whatever, the WPA calculator uses data on literally every game situation from 1977 to 2006 to calculate the chances of winning so it’s not like it’s theoretical or anything. Run Expectancy charts are probably a little less accurate, but they’re much more reliable than the “conventional wisdom” of baseball thinking. In fact, last season, there wasn’t one situation in baseball where a team would benefit from “moving the runner over.”

      http://www.baseballprospectus......cid=310118 Run Expectancy Chart

      Win Expectancy Calculator: http://winexp.walkoffbalk.com/expectancy/search

      equating walks with hits

      What’s better when leading off an inning–a single or a walk? There are times when hits and walks are equal. Usually, hits are more valuable since you’re unlikely to bring a run home by a walk, but to say that a hit and a walk have some huge amount of space between them in importance is ridiculous. The goal of the batter isn’t just to get a hit, but rather to not make an out. People who walk a lot (Swisher, Dunn, etc.) are valuable because they are good at not making outs. Guys who get a lot of hits and a lot of walks (Pujols, Chipper, etc.) are better because they can do both well.

      and totally ignoring (even claiming the non-existence) of clutch hitting.

      That’s because many traditional definitions of clutch aren’t that great. Newer “clutch” stats like leverage situation and WPA (posted on earlier today) are much better because they adjust for the context of the game. Runs in the first inning count just as much as runs in the ninth inning.

      Are they more revealing than a whole bunch of older metrics — not to mention the superior judgment of trained scouts? No way.

      Yes, they are better than the old metrics. Batting average tells you whether or not a guy got a hit. It completely ignores walks and reduces the value of a home run to that of a single. On Base Percentage includes not only hits but also walks and HBPs. I will, however, admit that it is not perfect because it doesn’t include adjustments for power hitting like Slugging Percentage does. SLG isn’t perfect because it doesn’t include walks and HBPs either. However, both OBP and SLG are far more informative than AVG.

      I’m not going to get in to the whole scouts vs. stats debate because to use one and not the other would be wholly stupid. Scots should be used to inform stats and vice versa. For a great perspective on this, check out the final chapter of BP’s Between the Numbers. There doesn’t need to be an either/or approach to evaluating baseball talent.

      • tim randle

        nobody on single vs a walk

        if you were a hitter, say , ARod, and you were watching another hitter, hypothetically, Abreu, and you could pick between a first pitch single or an at bat that rates sportscenter’s pitch count graphic as they fast forward through 13 pitches only to have abreu walk, what do you think would be better?

        if you’re the yankees, and you know you scored most of your runs in the 7th inning last year (2nd reliever?), do you want to go ahead and give an innings worth of work on one batter earlier in the game?

        or a first pitch single?

        (and no comments about abreu then getting thrown out please…)

  • http://dylankidd@earthlink.net dkidd

    wait for it…

  • steve (different one)

    Are they more revealing than a whole bunch of older metrics

    of course they are.

    i don’t even know what this means.

    • steve (different one)

      double fail