THE BOOK--Playing The Percentages In Baseball

The Schell book is very good, just be sure you get the one with SLUGGERS in the title, as opposed to his first book, with HITTERS in the title.

I just finished Schell’s book, and while I found it interesting and in many ways well done, I think that much of it is based on bad baseball history and logic, and no amount of clever math can make up for that.

He starts with,"Primary Assumption: After adjusting for ballpark effects, a pth percentile player in one year is equal in ability to a pth percentile player in another year for each basic offensive event.” I think that this is very wrong.  I am not arguing for a timeline, (though I believe in one), but that the way the game is played on the field has changed over time such that the assumption is incorrect.

Specifically lets talk homeruns.  Today almost all homeruns are out of the park, and a large majority of those are fly balls, with a few being linedrives, and not many groundballs!  It is obvious today that most batters, at least some of the time, try to hit fly balls in the hope that a useful percentage will be homeruns.  As I read the history this a change.  BBR, (that’s before Babe Ruth, not before baseball-reference), the conventional wisdom was that hitting fly balls was a bad idea.  Ruth changed this, but only slowly.  I don’t see any low average hitters who have most of their value in HR’s until long after Ruth.  Such as Kingman or Balboni.  So I believe that Schell is quite wrong to promote as top HR hitters batters such as Harry Stovey, Roger Conner, Charlie Hickman, as well as, Wagner, Cobb,
Lajoie, and Crawford.  Many of these were great hitters, but they were not fly ball hitters, so we can not simply say that if they had played their whole careers in NL 1977-1992 conditions they would be among the 100 greatest HR hitters of all time, as Schell says. 

To be continued;

In order to work from his Primary Assumption, above,
Schell would prefer that all his data follow normal distributions.  He goes to some length to transform the data to a more normal form.  I think that instead he should be asking why the HR data is much farther from normal in form than the BA data.  My guess is that much of this comes from most players not trying to hit HR’s.

As I think that he is wrong to adjust greatly the HR hitting of 19th century and deadball era players, I think he is also wrong to increase the adjusted numbers of Ruth, Hornsby, and others of that time because most others of that time were not trying to hit HR’s.  I think that his adjusted total for Ruth of 805 is wrong, as is the
idea behind the current title, The Year Babe Ruth hit 103 Home Runs.”

Schell does note the strong correlation between strikeouts and HR’s.  His adjusted figures have Ruth and Foxx 2nd and 3rd all-time behind Reggie.
I’m not as sure that this is wrong as I am with the HR’s, but again the context changed as more batters tried to HR and K’ed instead.  Also, of course, as the batters changed the pitchers did to in reaction.  In both cases the traits selected for excellence in 2000 are quite different than in 1900.

Schell is to be complimented for his explanations of his methods, though to completely follow everything a graduate degree in Statistics may be necessary. 

Worth reading in my opinion as an example of a certain type of incorrect argument.  Good math, questionable baseball analysis.

The process you are describing is flawed for the very reason you cite.

The ideal method would be to follow what is done here:
http://www.hardballtimes.com/main/article/changes-in-home-run-rates-during-the-retrosheet-years/

You simply cannot plop Babe Ruth in 1920, compare him to the distribution at the time, and then somehow link that to a distribution in 2008. 

The kinds of players in 1978 is much different than in 1998.  So again, you can’t compare how Mike Schmidt did with alot of gazelles around to how Sosa did with other power hitters around.  It’s a different kind of population of players.

Anyway, I really need to read the book now, if I’m going to be any more critical…