my account:
premium blogad:
blogads:
advertisement:
advertisement:
sitemeter:
advertisement:
|
|
|
December 4, 2007
Book Review: The Book: Playing the Percentages in Baseball
What The Book lacks in style it makes up in pure content. Its most remarkable achievement is a universal system of baseball statistics grounded in base-out states and their consequences: run expectancy and win expectancy. The Book, written by Tom M. Tango (otherwise known as Tangotiger), former St. Louis Cardinals analyst Mitchel G. Lichtman, and Andrew E. Dolphin, achieves a union of context-independent measurements of individual player performance and the context-dependent analysis of in-game strategy through a single meaningful unit. The importance of this for statistical baseball analysis cannot be overstated. The sabermetric community has no shortage of player valuation metrics; one can take his pick from On-base plus Slugging (OPS), Win Shares, VORP, Marginal Lineup Value and a number of other tools. Situational acumen is often measured by batting average with runners in scoring position or Win Probability Added. But the difficulty in these traditional statistics is finding how they actually relate to each other—using them to paint the whole picture to find some meaning. For instance, OPS includes two completely unrelated statistics, on-base percentage (OBP) and slugging average (SLG), which are simply added to each other even though their value is not equal and they do not share a mathematical unit. A plate appearance that results in a home run nets a player four “points” of SLG but only one of OBP. The Book does not use OPS. Instead, it uses weighted on-base average (wOBA), which is adjusted to reflect the run value of baseball events (thereby connecting it meaningfully to context-dependent analysis), and scaled to look like OBP. Using wOBA and the 24 possible base/out combinations and their resulting win expectancies (with no outs and no one on, a team is expected to score .55 runs, and so on), the authors analyze a plethora of different baseball situations and attempt to verify or disprove the common wisdom associated with them. The authors write, “The unwritten Book is about to be written.” In The Book there are chapters on hot- and cold-streaks, heads-up match-ups, clutch hitting, batting orders, platoon effects, starting pitcher use, relief pitcher leverage, sacrifice bunting, intentional walks, stolen bases, and baseball game theory. These are not separate articles, but rather parts of a unified system. In each chapter there are a number of situations and questions that are explored briefly using a play-by-play dataset from the years 2000-2004. After each section there is a conclusion, marked off by a box titled “The Book Says.” Some of these conclusions will come as no surprise to most readers: “Expect a player’s performance in an upcoming clutch at-bat to be much more like his overall performance than his past clutch performance.” Others may surprise: “A runner on first with less than two outs is an enormous disruption to the defense. The batter gains 14 points to his wOBA.” Or how about: “Early in the game in a low run-scoring environment, it is correct to often sacrifice bunt with a runner on first and no outs. In an average run-scoring environment, you should occasionally sacrifice to keep the defense honest.” Such conclusions are not arrived at lightly. The crew uses its play-by-play data extensively, along with regression toward the mean to account for random fluctuation. However, there certainly were times reading the book when I got to one of these conclusions and failed to see how it came from the analysis. For the most part, though, a careful reader will be able to follow along with the writers and understand how they’ve used their methods to arrive at their conclusions. That is not to say that the book is especially readable. It contains 140 tables and plenty of typos. One paragraph in the “Toolshed” chapter reads, “Why not simply combine them as OBP plus SLG and call this new-age statistic named OPS? Might this statistic see.” See if you can count the errors. The book, which was published by Potomac Books, really could have used an editor. For Mets’ fans, there are some interesting tidbits. Jose Valentin has the worst platoon split versus left-handed pitchers for switch hitters during the book’s sample period. Scott Schoeneweis (you know what’s coming) has the worst platoon split for left-handed pitchers versus right-handed batters. In a table titled simply, “Least Trustworthy Relievers,” we find our old friend Guillermo Mota. In a list of “Players Who are Awesome,” we find Lastings Milledge.* For the casual reader, the book is kind of dull. Looking for a break from numbers and analysis? Want to hear anecdotes you can’t find anywhere else? Do you enjoy witty narrative and ironic prose? Put down The Book and look elsewhere. But if you want the best book on baseball statistics since Baseball Between the Numbers, The Book is just the thing. * List may not actually exist. Note: You can find the authors’ daily observations at their blog. John Peterson hates old players on principle. You can read his stylized ravings regularly at Blastings! Thrilledge. 13 Responses to “Book Review: The Book: Playing the Percentages in Baseball”Leave a ReplyYou must be logged in to post a comment. |
advertisement:latest poll:
recommended journals:
recent user journals:
advertisement:recent articles:
rss/syndication:your ad here:advertisement: |
Loading ...
You got it in one, John. There’s a lot of good information in Ze Book — the chapter on lineup selection was a highlight — but you’re right, it really isn’t that reader-friendly, even if by ‘reader’ we mean ‘analytical geek who isn’t afraid of spreadsheets.’
Who’s this Milledge guy again?
Brilliant. Is Umberto Eco a stat-head baseball fan? He ought to be.
I got the book and I must say its great. I guess it would be a lil boring for people not really into the numbers but if you are into sabermetics its a great book.
this doesnt surprise me. i never listened to y’all. :-p
I loved the last chapter of The Book, actually. Their brief discussion of game theory was mind-blowing and added a whole new dimension to games that I watch.
Is there any list about players whose grandmothers could strike them out if their grannies were lefties?
It’s probably filed under “Ryan Church”, y’know.
Today’s game isn’t a low run scoring environment.
I think i liked the lineup construction chapter the most….but yeah i also agree with dan about the game theory chapter.
Even tho some of what they talk about on their blog is a bit over my head, i go there on a daily basis.
Thank you for your honest review. It is much appreciated.
As for the typos, I’m looking at my personal PDF version (I’ll double-check the print version at home), but it has the word “it” in the sentence you copied, as well as a whole line is missing in your post. This is what it actually says:
“Why not simply combine them as: OBP plus SLG, and call it this new-age statistic named OPS? Might this statistic allow the deficiencies in OBP and SLG to cancel each other out? Let’s see.”
So, all looks fine to me in my final PDF version (which was used to source the book). I will double-check the original print edition, as well as the reprint edition from Potomac.
Tom
Ok, I’m home. The self-published version is correct, and matched my PDF, which I’ve reproduced in post 9.
The professionally-published one omits and entire line, which is why it looks as strange as it does.
The “it” however exists in all editions, and is therefore a typo on the reviewer’s part, I would figure.
Thanks for pointing out the typo. What I guess we should have done is NOT hired an editor, since we did fine without one!
You’re right, Tom. The line verbatim from the book is, “Why not simply combine them as OBP plus SLG and call it this new-age statistic named OPS?”, which is still a bit awkward but makes far more sense with the “it” included.
However, here is the whole paragraph in my version of The Book (I can’t speak to John’s version, though it may very well be the same:
“We have one statistic that is deficient in one area, and another one that is deficient in another. Why not simply combine them as: OBP plus that is deficient in another. Why not simply combine them as OBP plus SLG and call it this new-age statistic named OPS? Might this statistic see.”
The emphasis is mine. It looks like part of one sentence is mostly repeated, and that last sentence appears to have been cut off. This is exactly as the passage appears in my copy, though it may be that subsequent printings (or your edited PDF) have been corrected to reflect them.
Yours is in the reprint edition from Potomac.
The correct one is in the original, self-published by TMA Press.
As strange as it may sound, the more recent version is both wrong and done by professional editors.
If there are more strange passages like this, please email me tangotiger at yahoo dot com, and I’ll forward to the publisher.
Thank you.
Hah. I suppose it figures that the professional editors would get it wrong.