Tag Archives: sabermetrics

Baseball Research Journal 1980 edited by Bob Davids: a review

The 1980 BRJ, like most editions, is now available on SABR’s website.

This is the strongest Baseball Research Journal edition yet. It features one of the classic works of sabermetric analysis, some excellent biographical portraits, a look at minor league umpiring practice in 1900, a glimpse at the 1880 opening of the original Polo Grounds, some analysis of why Fulton County Stadium was a launching pad, and an interview with Joe Oeschger about the longest major league game and other memories. Authors included Stew Thornley, Seymour Siwoff(!), David Smith, Ted DiTullio, and Richard Cramer. This issue has, of course, some variations in quality, but there’s really nothing you could fairly characterize as filler in this edition.

The classic sabermetric piece is Dick Cramer’s “Average Batting Skill Through Major League History,” which uses a moving average based on Cramer’s Batter Win Average (BWA) statistic (discussed in the 1977 BRJ) in an attempt to discover whether player performance has increased or decreased over time. I’d read this piece before, and been concerned about Cramer’s treatment of outliers, but this re-read has convinced me that he handled my concerns adequately. There is room for some quibbling, of course, but his basic point–that the talent level of major league players has increased substantially over the years, and that it continues to increase–is well argued and well supported by his data. This article is often cited as an important early analytical effort, which it is; someone should apply modern methodologies to the question and see how things hold up.

The early chapters of the 1980 edition are devoted to the post-playing careers of Billy Sunday (by Robert Muhlbach), Alfred W. Lawson (Lyell Henry), and Frank W. Olin (Tom Hufford). Henry’s Lawson piece is particularly interesting.

A number of events are commemorated in this journal, including The Last Tripleheader (A.D. Suehsdorf), a 1880 night game between two department store teams (Oscar Eddleton), and the Polo Grounds opening mentioned above (John J. O’Malley). Seasons considered include Joe Bauman’s 72 HR campaign (Bart Ripp) and the 1884 St. Paul Unions (Stew Thornley–I think his first SABR publication). Biographical treatments, besides those mentioned above, included Negro Leaguer Cannonball Dick Redding (John Holway) and nineteenth century star Jim Sheckard (Gregg Dubbs). There’s also a cute little piece on Rube Waddell playing college ball, writen by Harold Esch.

More or less sabermertic pieces, besides Cramer’s effort, included Seymour Siwoff’s accounting of some previously-undiscovered RBI records, David Smith on stolen bases (using Maury Wills data–and evidently Smith’s first SABR pub), and John Schwartz’ look at Intentional Walks.

Robert Kingsley’s look at Atlanta home run rates and Richard Burtt’s similar look at Pittsburgh triple rates both seem inadequat, as both explicitly discount what seem to be obvious causes (altitude in the Atlanta case, field dimensions in the Pittsburgh case). I think this is partly a case of we better understand these dynamics nowadays, and partly willful blindness on the authors’ parts.

There are, of course, the usual array of lists-with-explanatory-paragraphs; I shan’t list them here. Al Kermisch’s Researchers Notebook looked, among other things, at a postponed game in the 1918 World Series, Silver King’s no hitter, the 1889 Louisville player’s strike, and a recording error in Harry Schafer’s fielding records. Along the same lines was Arthur Ahrens attempt to pin down a story about an oft-misreported Bill Lange catch–Lange supposedly went through the outfield fence. Ahrens offers a plausible reconstruction which suggests that a couple odd incidents in a single game got garbled as folks retold the story.

So there’s some biography, some excellent historical work, a couple minor league pieces, and a key analytical piece. Something for every SABR audience.









This review was originally published on LibraryThing.

Revision History:

Bill James Presents The Great American Baseball Stat Book (1988) edited by Don Zminda and Project Scoresheet

My review of the first book in this series set attracted a note from Geoff Beckman, who helped edit both editions. It would likely be worth your while to read Geoff’s comments before reading on….

Geoff considers this the weaker volume, and on the whole he’s right, as it lacks the zing of the initial edition. There were 338 player essays (up from 240), but on the whole they’re less interesting. Too many read like spring training player profiles, emphasizing the player’s tools and potentials while downplaying his demonstrated weaknesses. This is not meant to imply that the essays are all bad. For instance, principal editor Zminda portrayed a stubborn and still-valuable Carlton Fisk on a Chicago team that didn’t much like him. Craig Wright’s commentary on Bret Saberhagen gives clues about the value Wright provided the Rangers as an early professional sabermetrician. Beckman’s own Mel Hall essay used six methods to frame Hall’s value and is absolutely delightful. Mike Kopf made an effort to get into Bo Jackson’s head. And Susan Nelson offered a fine look at Dennis Eckersley in transition from starter to reliever.

Nelson’s Eck portrait, in fact, illustrates the book’s unplanned theme: In 1987 baseball’s pitcher usage was in transition as most managers had largely abandoned the four man rotation and were retreating from their long-held preference that starting pitchers finish ballgames. In retrospect it’s pretty clear that the late eighties were a turning point, formalizing long-developing pitcher usage patterns in ways that few would have anticipated.

The second edition of the book added team essays, which were uniformly forgettable. These were followed by a set of truly interesting, but unsystematic, manager essays. For this reader, these justify the book’s existence.

There’s some worthwhile stuff in the back of the book. Gary Gillette–or perhaps the scoresheet project in general–offered some interesting measures of defensive ability, reworking range factor to measure opportunities more precisely. I’m not sure whether anyone followed up on this effort, but it’s certainly interesting. I’d like to see more work along these lines.

Gillette and Dave Nichols did something similar with base runners, measuring steals in terms of opportunities rather than attempts. This, too, seems to be a one-off effort, and again it would be interesting to see further work in this vein. The same authors also took a brief look at baserunner advancement on hits, not offering much analysis but presenting a few tables.

Mark Pankin followed up on his discussion, in the previous edition, on Markov Chain Analysis, mostly presenting better data but not really extending the earlier essay. Matthew Lieff and Gary Skoog considered a similar, but less calculation-intensive, approach to using the same data to examine the effectiveness of in-game strategies.

In other end-of-book essays, David Gordon looked at Quality Starts and found them meaningful. In contrast, Merrianna McCully offered scads of data about the weakness of contemporary pitching–a piece that I’d characterize as more entertaining than informative. And Brock Hanke described Whitey Herzog’s playing career, and pondered how it shaped the Cardinals organization when he became the team’s GM. The Hanke essay was a good preview of the work he’d later do elsewhere. It was also the book’s only piece I still remembered when re-reading 25 years later.

This was the first time I saw Hanke’s name in print. The same is true of several of the book’s other contributors: John Benson, Sherri Nichols, Stuart Shea, and Sean Lahman come quickly to mind. (Tom Tippett was there, too, as one of the project’s programmers.) Just bringing these folks (and Pankin, in the previous edition) to my attention is plenty of justification for the effort.

This was the last GABSB produced in this guise, though Gary Gillette would resurrect the title for another project a few years later. It was certainly a worthwhile experiment, but apparently wasn’t sustainable.









This review was originally published on LibraryThing.




Revision History:

Percentage Baseball by Earnshaw Cook: a review

It’s likely my first contact with this book was Jim Gallagher’s review of the Waverly Press edition published in the September 12, 1964, issue of The Sporting News:

If I didn’t know Earnshaw Cook, I’d think his book was one of the loveliest hoaxes of all time … 120,000 words, plus 108 tables, 35 diagrams and countless fearsome formulas … all aimed at befuddling the pseudo-intellectual sophisticates who are taking over the business of sports comment. But although Mr. Cook is a humorous and good-humored man, he is as serious about his baseball theories as an orator at a political convention. [punctuation as original]

[snip]

Cook, however, like all theorists, tends to be intolerant of [traditional baseball wisdom about randomness] and feels that baseball managers unwilling to adopt the methods indicated by his statistics are unenlightended … perhaps even stupid.

When a new edition was published by MIT Press, TSN editor C.C. Johnson Spink took notice in the April 23, 1966, issue of baseball’s bible. Spink’s commentary began “It would appear that major league managers have neglected their education if they have not studied college-level mathematics.” After a few comments about the book’s difficult reading level, he goes on to summarize the book’s most interesting suggestions–regarding bunts [don't, except for pitchers], batting order [best hitter first; worst last], pitcher usage [3 per game, with the "starter" pitching the middle 5 or 6 innings], and platooning [don't do it]. Spink also repeats, without comment, Cook’s contention that following his suggestions would improve a team’s annual scoring performance by 273 runs.

Spink would again mention the book, clearly favorably, in a May 1, 1971, editorial about pitcher usage. I didn’t find any subsequent editorial mentions of the book, though Waverly Press advertised the original edition for years, even after the MIT Press edition was available.

In the modern sabermetric community, Cook is generally acknowledged as a pioneer baseball analyst who did some useful work. A few analysts have apparently even extended his best work. But the consensus can be fairly characterized as dismissive.

Here’s the thing, though: The book, whatever its merits, is unreadable. The math usually seems to work, if you can follow it, but following it is difficult. Problems include the use of uninformative variable names [for instance, scoring index is Dx], calculations that generate not-obviously-meaningful numbers, and the book’s organization which unhelpfully splits the statistical discussion between the main text and the appendices. There are also numerous off-topic digressions, most of which demonstrate Cook’s belief that deadball-era baseball was superior to the modern “cheap homer” brand (a legitimate opinion, but not a helpful issue in this context). There are a few howlers in his analysis, of course; the most telling are his discussions of batting order and platooning, neither of which stand up to careful examination. Most trying, though, is the man’s sheer contentiousness: He doesn’t suffer fools, and he considers everyone foolish.

So why read the book? I can think of two reasons. You might want to mine the book for its equations; people have certainly done that, and found value. Or, like me, you might just want to read everything sabermetric. Unless you fall into those categories, you really don’t want to go there.









This review was originally published on LibraryThing.

Revision History:

Baseball Research Journal 1978 edited by Bob Davids: a review

Bob Davids, with help from Emil Rothe, collected the usual range of serious research and lightweight work for the seventh BRJ, which the editor notes was published late in the season (September, evidently). The best work in this edition is really quite good.

The issue kicked off with a pair of Braves-related pieces–Bill Price’s history of Braves Field, which is typical of the genre, and Randolph Lindthurst’s relatively short note about the relatively old (and temporarily very successful) rookies on the 1937 Bees’ pitching staff, Jim Turner and Lou Fette. The “lists with supporting commentary” department included Tom Hufford’s well-done piece about pitching appearances by position players, Ted DiTullio’s short commentary on players whose long major league careers occurred on a single team (for some reason he disqualified players whose teams shifted cities), Ron Liebman’s piece about pitcher winning streaks (the table’s more interesting than the discussion, methinks), Ray Gonzalez on Lou Gehrig (relatively weak, for Ray), and Paul Doherty’s essay on forfeited games. Bill James contributed a disappointing (to me) piece about what he’d later call the “Approximate Value Method”–which, uncharacteristically, he doesn’t fully define in the article.

Other pieces include essays about Chino Smith (John Holway), fielding feats (Rothe), nicknames (Stan Grosshandler), Fred Toney’s 17 inning minor league no hitter (Jack Rudolph), and Arthur Ahrens’ great exploration of the Western League’s turn-of-the-century transition to the American League, with a focus on Charlie Comiskey’s Chicago White Sox franchise. The Ahrens piece is worth the price of admission.

The issue featured two excellent sabermetric pieces. Pete Palmer’s essay concerned park effects in the American League, and is as good an introduction to the topic as I’ve ever seen. Pete’s later projects gained some sophistication, but this effort touches nearly all the basic issues. And Irv Matus, who apparently counted pitches for all the Mets’ games in 1976, authored an excellent examination of the impact of pitch counts on pitcher performance. I doubt this was the first time such an effort had been made, but if anyone published such a well-thought-through analysis before Matus I’ve not seen it.

SABR’s membership’s interests are varied, and its audiences are many. They were well served by this issue.









This review was originally published on LibraryThing.

Revision History:

The Baseball Analyst Issue 6: a review

Bill James published 40 issues of The Baseball Analst in the 1980s. All are available on SABR’s website.

James edited this issue, which includes five articles (one apparently frivolous); James added an editorial comment at the end. Three of the articles are followups to articles in earlier editions.

Paul Schwarzenbart completes the ballpark study he began in the first issue of the Analyst by documenting the American League parks. He again concludes that the main playing impact of artificial turf is that it reduces infielder errors, but conceded Craig Wright’s argument (Issue 2) that Schwartzenbart’s data also shows a similar, albeit lesser, impact in the outfield. The infielder impact is a strong effect, but the author points out that a superior fielder can be shown to overcome the effects of a difficult infield. He also comments on the best of the then-current crop of A.L. infielders. The numbers in this article are interesting.

Dan Heisman’s contribution is not so much a research report as a contemplation of the relative merits of long careers vs. high-peak careers. Some of the discussion is provocative, though James’ concluding editorial comment takes issue with one of Heisman’s main points. The note concludes with a tabulation of five-year peaks for some notable players.

Dallas Adams follows up on his first-issue study of run scoring distributions by demonstrating that his summary data correlates highly with won/lost record. He incidentally supplies three ways to calculate a team’s winning percent based on runs-per-game/opponent’s-runs-per-game data. This essay’s decidedly less ambitious than Adams’ previous efforts, but it’s presented well.

Pete Palmer extends the Jim Reuter essay, on park factor calculations, from the previous issue of the Analyst. Palmer’s first extension demonstrates that Reuter’s method works better if the calculations are based on innings rather than games, and shows how to make the necessary adjustments. He also suggests a further, individual-player, version of the factor which he doesn’t fully describe. The second extension begins with the observation that a team’s offensive “park effect” is partly the product of not facing the local pitching staff; he adds adjustments to allow for that. James’ editorial comment takes issue with some of Palmer’s assumptions, pointing out that James and Palmer differ on the implicit meaning of context as applied to calculating park effects.

Finally, someone claiming the name Cuthbert Magnolia offers a method for comparing pitchers, apparently intending to create a measure similar to Runs Created. I’m not sure I understand the method, and the essay itself is deliberately insulting. But the results seem to make sense.

Revision History:

The Baseball Analyst Issue 5: a review

This issue, like all 40 issues of this journal, is available on SABR’s website.

This edition begins with a plea from Bill James for more material. The issue’s four articles are by Dick O’Brien, Jim Reuter, Pete Palmer, and Dallas Adams.

O’Brien’s contribution is a quick-n-dirty examination of the relative importance of hitting and pitching when a team’s record improves or regresses by .031 (5 games). He concludes that teams whose records improve might show improvement at either offense or defense (or both). While collapsing teams have generally similar numbers, it appears more likely that the cause is a decline in offensive production. He goes on to examine ballpark effects. (O’Brien’s getting better at writing up his studies, which is good.)

Jim Reuter takes on Park Factors in this issue, arguing that simply splitting the season’s ballpark effects into home/not-home distorts the impact of extreme home ballyards. A better method would treat all league ballparks equally, at the cost of a slightly more complex calculation. The impact of the corrected calculation is, he shows, small (1% for Wrigley), but worth getting right.

Pete Palmer takes a look at the impact of balls and strikes on batter performance, based on data collected for 31 mid-1970s postseason games. He finds, unsurprisingly, that batters OBA improves as the ball count goes up; more surprising is the discovery that strike counts impact SLG far more than balls do. I’m sure this study has since been replicated with larger and more representative datasets, but don’t recall having seen that.

Dallas Adams gets more than half of the issue to complete his Issue 4 study of the distribution of run scoring, exploring some ways the data can be applied. Again, this involves fairly sophisticated math, but Adams is extremely good at explaining his methods. Definitely worth reading.

Revision History:

The Bill James Baseball Abstract 1988: a review

This was the last edition of the Abstract, and the weakest of those published nationally. Unlike most of the issues, this one has no overarching theme, and offers little in the way of pathbreaking research. One new analytical tool, Pitchers’ Game Scores, is introduced in this book, but Bill doesn’t seem to recognize that folks would put it to use. The overall tone’s valedictory, from the dedication to the closing essay.

That’s not to say the book’s a washout. There is, for instance, an interesting examination of platooning, as practiced in the mid-1980s. While there are wide variations, Bill reports, basically all hitters have a platoon advantage against opposite-handed pitchers of between .020 and .030. Breaking things out further, he finds few player types with greater-than-normal advantages–power hitters, for instance, aren’t much different from normal; neither are players with high strikeout rates. The only apparently-significant groupings he reports are older hitters, and perhaps hitters who work pitchers for walks, both of whom have higher platoon biases than normal. This seems to have been the first systematic examination of platoon advantage based on game data. On the whole, it mostly confirms conventional wisdom.

In other essays, Bill presents an argument that the minor leagues should be freed from major league control, and makes some predictions about the effects of the 1988 changes to the strike zone enforcement rules. One of the essays describes Game Scores, apparently for the first time; it’s since become a staple of sabermetric analysis.

The team essays are, as in the 1987 edition, actually focused on the teams; most, frankly, are pretty dull. The Twins essay did a fine job of dissecting their success, though, and a followup essay skewered the notion that the Twinkies were unusually dependent on two pitchers. The Oakland chapter is largely devoted to trying to understand LaRussa’s quirks, which turned out to be an ongoing sabermetric theme. The excellent Cards essay triggered a second excellent essay which used Herzog as an excuse to examine the field manager’s job. And the Astros essay is one of the finest analyses of a team’s season anyone’s written.

The best team essay, though, is about the Indians, and how folks could have predicted success for a team which played so horribly, and about the team’s prospects in the near and middle term. This is Bill James at his best–witty, sarcastic, and right on target analytically.

The player essays are different from any prior edition. Bill only wrote about players who interested him in this issue, most of whom got several-paragraph essays. None really struck me as special, though the Tim Raines essay is a reminder of the wonderful skills Rock brought to his team.

The book ends with an essay, “Breakin’ the Wand,” which reviews Bill’s career as a baseball analyst–and discusses his impact on the sport, both positive and negative. It’s worth reading, but not worth going out of your way to read, methinks. It is clear, at the end, that James expected he was planning to pursue other interests and would be leaving baseball analysis behind. That’s not, of course, exactly what happened.









This review was originally published on LibraryThing.

Revision History:

Win Shares by Bill James with Jim Henzler: a review

Win Shares is Bill James’ best effort to synthesize everything he knows about baseball into a single measure. A decade after the book’s publication, Win Shares has become an accepted tool in sabermetric analysis; by that measure this volume is a success. Nonetheless, the book is a famously difficult read.

I originally read Win Shares just after the book was published, then studied it a year or so later when I adapted its framework for a minor league research project. I found that the practical application was really helpful to understanding how the pieces fit together (although, sadly, it didn’t much help my project). This decade-later read benefits from familiarity, now, and from watching other folks apply Bill’s methods. Nonetheless, this is a difficult book.

The first problem is the book’s arrangement. The first dozen pages give some background and present a fairly unhelpful system overview. This is followed by nearly 90 pages of equations and explications, with occasional explanations; you’ve really got to take this section on faith if you’re reading the book from front to rear. This is followed by 20 pages which attempt to explain, not always successfully, how the equations fit together, and the assumptions behind them. These pages also directly address some of the questions James heard from others when he’d discussed the project while he was developing the system. 150 pages of “random essays” follow: Some of these address issues raised during the system’s development, but more just illustrate of the sorts of questions Win Shares can help a researcher study. Some of these studies are immensely interesting–in this way Bill’s work is always worthwhile. Finally, 250 pages into the work, we get the long lists of organized and comparative career and season numbers which make this a useful reference work; these constitute nearly two thirds of the publication and are the actual product of the system.

The second issue is that the book’s emphasis on Defense overwhelms the main message. In terms of the Win Shares project, Hitting and Pitching are treated as Solved Problems; those sections of the system are built around Runs Created and Component Earned Run Average. Since those methods were addressed in detail in earlier books, they are given only slight explication in this volume. Assessing Defense, on the other hand, is a new problem, so James gives it the bulk of his attention. Worse, for many readers, the system looks like a Rube Goldberg machine–or a house of cards. While it’s clear that the apparently arbitrary numerical constructs described in this book make large concessions to deficiencies in the statistical record, the justifications underlying the calculations are largely built on research which James describes elsewhere in the work. Since they’re initially presented without that context, accepting them on faith seems a lot like a leap of faith.

All this discussion of D overwhelms, in pages and evident intellectual effort, James’ description of the overall structure of Win Shares proper. Win Shares is a team-based method for assessing player seasons, and its core is allocating the overall team success to the players who made up the team. Many readers likely lost that message in the relentless march of difficult detail. Personally, I think he should have relegated the defensive discussion to an appendix, and presented the framework without the distractions.

A third concern is that James doesn’t convincingly articulate his basic argument about assessing defense. This matters because many analysts seem not to have absorbed his contention that defense is a team function, and individual defensive efforts make sense only within the context of their team’s defense. I recognize progress on this front, but we’re still far from really understanding how to measure this component of the game.

A fourth matter of concern is the decision to award three win shares for each team win. I fully understand Bill’s reasons, but the number itself is peculiar enough to create a barrier to understanding.

A final difficulty, which is less important, is the constant comparison of Win Shares’ defensive assessments with those generated by Pete Palmer’s Linear Weights. Some comparison was clearly necessary, but Bill’s not one for half-measures when he presents an argument. It’s pretty clear that Bill was aware that this was problem.

All that said: This is an immensely influential book, and deservedly so. Would it was less to difficult to master, but the effort’s worthwhile.









This review was originally published on LibraryThing.

Revision History:

Curve Ball by Jim Albert and Jay Bennett: a short review

I’d planned, and even drafted, a long review for this book, but decided not to post that. Instead I’m posting this short note.

Curve Ball is a look at sabermetrics by a pair of professional statisticians. It begins by discussing how statisticians view numbers and analysis, then moves to baseball’s. The authors are apparently familiar with every important sabermetrician, including the commonly-cited sabermetric predecessors, and with many (perhaps all) professional statisticians who study baseball. Some of the book’s chapters are overviews, while others examine specific topics.

The chapters I found most interesting were a series about modeling baseball offenses. On the whole, these guys give the leading sabermetricians good marks; in particular, Bill James’ Runs Created and Pete Palmer’s Linear Weights are given high accolades.

The book’s enjoyable if you’ve some background in academic statistics, but it’s likely difficult reading if you’ve not encountered that notation and vocabulary. I worked my way through the discussions, but was rummaging through four-decade-old memories from time to time. It’s certainly an essential book if you’re interested in serious baseball analysis.









This review was originally published on LibraryThing.




Revision History:

The Best of Baseball Prospectus: a review

This is a big, sprawling, two-volume book which offers a “best-of” selection from the Baseball Prospectus website. The work is generally of very high quality, and is well-organized, but the website origins of the essays occasionally cause some orientation issues. Moreover, the selection was deliberately biased toward more recent writings, apparently because the editors believe the context has rendered much of the older analysis obsolete–a belief I share, by the way.

The best stuff is classic. BP published Voros McCracken’s “How Much Control Do Hurlers Have?” early in 2001, which is likely the most influential sabermetric essay published in this century; it’s here, as are several author’s reactions. Rany Jazayerli’s delightful, twelve-part exploration of the free agent draft is reproduced as written; it’s fun and informative (though this is one of the places where a the book’s web origins really show; a rewrite would surely make things more coherent). Keith Woolner and James Click explore the areas sabermetrics had not, as of their essays, examined; everyone should read these essays for an overview of the discipline’s landscape. There’s a representative selection of Christina Kahrl’s delightful Transaction Analysis columns; I always looked forward to those. Besides the current staff, Joe Sheehan, Doug Pappas, Nate Silver, Gary Huckabee, Jonah Keri, and Dayn Perry are all represented; Derek Zumsteg, sad to report, is not.

The first volume’s largely about the game itself, often (but hardly exclusively) from a sabermetric perspective; the second volume could be said to be about the business side of the sport. Both are worth reading; both are often fun. A good book.








This review was originally published on LibraryThing.

Revision History: