clock menu more-arrow no yes mobile

Filed under:

Sabermetrics in the latest Baseball Research Journal

The latest Baseball Research Journal (volume 34) from SABR has been completed. I received a couple of copies since I am the author of one of the articles (does that mean authors of two articles should get 4 copies?). Other SABR members generally are supposed to receive it about mid-February. Jim Charlton is the editor. Anyway, there are some interesting sabermetric articles in it. Here is a brief rundown for non-SABR members, in order as they appear in the journal.

"Normalized Winning Percentage, Revisited" by Bill Deane. An update of an article he wrote several years ago. "NWP basically measures how much a pitcher has exceeded his team's performance, divided by how much he could have done so..." is how Bill defines it. So if a pitcher's team has a .600 pct and he has a .700, his NWP is .625 because he went .100 above his team while the max would have been .400. His .100 is 25% of the possible and if you increase .500 by 25% you get .625. He goes on to list the best pitchers ever in this as well as from 2005.

"Has Greg Maddux employed the "Bagwell gambit" in his career?" by me. Last year in Newsweek magazine, George Will wrote that Maddux once let Bagwell hit a HR in lopsided game only to strike him out in a close game two weeks later on the same pitch. My research indicates that Maddux has not really made a habit of tricking hitters this way. Also, Bagwell has hit Maddux very well in his career. If you email me, I can send you a copy of the article.

"Are Traded Player's "Lemons?" by Phil Birnbaum. He finds that it is possible that players who get traded may not reach the level of performance that was projected (at the time of the trade) for the remainder of their careers.

"In Search of Clutch Hitting" by Tom Ruane. He looks at how players batted with runners in scoring position (RISP) vs. non-RISP situations. His study includes all players with 3000+ ABs from 1960-2004. He discusses some of the data and statistical problems in looking at this kind of clutch data. The actual clutch performances by players was similar to what he found in random simulations. So he concludes that we might be able to ignore "clutch forces" without much loss of accuracy. This article is on line at the Retrosheet website. Just click on research. It also has plenty of data tables.

"Cumulative Home Run Frequency and the Recent Home Run Explosion" by Gabriel Costa, Michael Huber & John Saccoman. They look at how the career HR %'s of the players in the 500 HR club changed as they aged. For most, it rises then levels off. But for many of the most recent entrants, it kept rising as they aged.

"Should a 22-Game Sweep Have Occurred? An Examination of Season Sweeps and Near-Season Sweeps" by Bruce Cowgill. He looks team vs. team for a season and at some famous actual cases and then gets into probability theory. The odds are good that it could have happened over a 22 game schedule and he thinks it might happen in 19 games.

"Do Batters Learn During a Game" by David W. Smith. He tracks performance of hitters in each plate appearance as the game progresses and finds that they do tend to hit better. An earlier version of this is at the Retrosheet website. Just click on research. It also has plenty of data tables.

"Which Great Teams Were Just Lucky" by Phil Birnbaum. In this article, luck has three dimensions: teams that won more games than expected based on runs scored and allowed, teams that scored more runs and allowed fewer runs than expected based on other stats, and teams that saw an unusual number of players have "career years."

"World Series Winners and Losers: What's the Difference?" by Kent von Scheliha. He creates a "team strength index" based on runs scored and allowed and how many standard deviations from the league average each pennant winner was. He then looks at the biggest upsets in series history based on his TSI as well as who the best and worst teams were. There is also a comprehensive table for all series teams.

"Deconstructing the Midas Touch: Gold Glove Award Voting, 1965-2004" by Robert Reynolds, Steven Day and David Paculdo. They present a statistical model in which various fielding stats are used to predict who won the award. Other variables in the model include appearing in the post season and winning the award the previous year.

"Do Some Batters Reach on Errors More Than Others" by Tom Ruane. This is probably not random. He takes hitting with men on base, the park and the handedness of the batters into account. An earlier version of this is at the Retrosheet website. Just click on research. It also has plenty of data tables.