America's Pastime for a Beautiful Mind
Apr 10, 2012 03:23PM ● Published by Anonymous
A baseball game can be watched from two basic perspectives—from the heart or from the head. You can “root, root, root, for the home team” and “if they don’t win, it’s a shame.” Or you can look at it as a numbers game between rival theorists. In reality, most seasoned observers of the game look at it both ways—they have a favorite team and favorite statistics and satisfy both the longings of their hearts and the thoughts in their heads. Novices have their favorite teams, but know little about the validity of the statistics behind the numbers game.
If you are a novice, why not add another dimension to your game—try watching baseball like a statistician. If you decide to do this, you will immediately encounter a problem. A statistician is not a statistician is not a statistician. There are different ways to rate the performance of teams and players and there is no agreement on which is the best. Different statisticians look at it in different ways.
The four most widely used measures of a hitter are batting average (AVG), on-base percentage (OBP), slugging average (SLG), and on-base plus slugging (OPS). They are routinely quoted and regularly discussed by television commentators and newspaper columnists. Other measures have been developed—total average (TA), linear weights (LWTS), and runs created (RC), for example—but because of their technical nature they are not generally quoted or discussed by the press.
How can a novice decide what statistics to use if the experts disagree? The answer stems from a fundamental baseball fact—no run has ever been scored, or can ever be scored, unless the runner travels around the bases from first to second to third base and, finally, to home plate. Bases are the building blocks of runs. The further around the bases a runner gets, the more likely it is he will score. Thus, the team that accumulates the most bases will generally score the most runs.
Five of the seven measures of hitting use bases in their calculations. Two of them—batting average and on-base percentage—do not count bases. Batting average counts hits and on-base percentage counts times on base. Statistical studies show that these two measures correlate least of all with runs scored.
The granddaddy of the five measures that count bases is slugging average. It counts one base for a single, two bases for a double, three for a triple, and four for a home run. Unfortunately, slugging average excludes walks and hit-by-pitch—two events which contribute to the scoring of runs. On-base plus slugging reflects the weaknesses of on-base percentage and slugging average. Total average applies the same weight to singles, walks, and hit-by-pitch, but with runners on base singles can advance runners two bases whereas walks and hit-by-pitch can only advance a runner one base and only when there is a runner on first base. The calculations for linear weights and runs created are more complicated and too time-consuming for novices.
My book, The Runmakers, also uses bases in its calculations. It counts bases made by batters for themselves and also counts bases earned for advancing runners on the bases and for driving in runs from the bases. The result, potential runs per game, correlates best of all with actual runs per game.
In this article’s sidebar is a form for counting bases. Before setting out to use it, however, remember that Rome was not built in a single day. Look at counting bases as a process. It is not as easy as it may look. Adopt an incremental approach. Start out by counting bases that batters accumulate by themselves. Then, in subsequent games, gradually add the other steps.
Once you become adept at counting bases, you need to interpret the results. Remember that position in the batting order is very important. Number three, four, and five batters have more opportunities to advance runners and drive in runs; number one and two batters have many fewer opportunities.
But what about pitching and fielding? I have not yet developed a comprehensive theory of pitching and fielding. Branch Rickey said there wasn’t much one could do with fielding. Some progress has been made since then, but fielding remains a very difficult subject.
Perhaps the most important point about pitchers is the need to categorize--starters, middle inning relievers, set-up men, and closers. Earned runs-per-inning is the most important statistic for pitchers in each category. Games won or games lost are irrelevant because they reflect the contributions of entire teams from both the defense and the offense. Other pitching statistics—strikeouts, walks, hits, etc.—are secondary to earned runs. If a pitcher has a lot of strike outs, his earned runs should be lower; if he has a lot of walks or hits, his earned runs should be higher. Games are won or lost based on the number of runs not the number of strike outs, walks, or hits. I do, however, favor one adjustment to earned runs. If pitcher A allows three earned runs in six innings (.50 per inning) and pitcher B allows four earned runs in nine innings (.44 per inning), the latter has clearly pitched a better game. The format for recording pitcher performance is at the end of this article.
Looking at baseball from the perspective of a statistician may seem tedious to a novice. But if you stick with it, the rewards will be great. In the meantime, don’t forget to enjoy the game by rooting for your favorite players and teams. Baseball will always be a game of the heart and the head.
Dr. Frederick E. Taylor is the author of The Runmakers: A New Way to Rate Baseball Players, published in 2011 by The Johns Hopkins University Press. He is a resident of Annapolis and was educated at the University of Rhode Island, the Fletcher School of Law and Diplomacy and Georgetown University which awarded him his Dr. of Philosophy degree. He worked for the Departments of Commerce and Defense and taught at several universities. The Runmakers is his first book. It is available in bookstores and on-line.