Forecasting a Winning - or Losing - Season
Robert M. Saltzman of San Francisco State University says a team's won-loss record at the All-Star Break is the best predictor of a good season. Prof. Saltzman will deliver a paper, "Wait Till Next Year? Some Analyses of Win-Loss Records from 1960-1998," in the Cincinnati Convention Center on Tuesday, May 4 from 4:30-6pm.
"Several conclusions can reasonably be drawn from basic statistical analyses," he says. "First, there's no need to follow baseball until Opening Day: The exhibition season is not important to your team's ultimate success. Second, you should be concerned if your team isn't playing well by the All-Star Break: They're quite unlikely to make a major recovery during the second half of the season despite Yogi Berra's admonition that 'It ain't over till it's over.' Finally, should you optimistically 'wait till next year' if your team has a bad season? Yes, you can be somewhat hopeful about your team's chances of success next year after a poor performance this year, but you'll have to be equally insecure about a letdown following a good year."
Although the strongest predictor of future success is a team's winning percentage at the All-Star Break, it's not a sure thing, Prof. Saltzman notes. In 1997, the Philadelphia Phillies climbed from an absolutely terrible .286 winning percentage at the All-Star Break to an almost respectable .420 by season's end. This was the best second-half improvement on record since 1960. In contrast, the Cincinnati Reds' hopes were dashed after mid-season in 1991, when they went from a commendable .550 to a disappointing .457, making theirs the ninth biggest drop in winning percentage since 1960.
Prof. Saltzman also examines changes in baseball since 1960 and their effect on the game. Expansion, in which existing teams lose only a few non-star players, he argues, has had only a short, negative impact on overall predictability. On the other hand, the amateur draft begun in 1965 initially equalized the teams by spreading out the good new players. By 1970, though, the correlation between success in one year and the following year returned to its pre-draft range. Free agency, though, has brought less stability to team rosters and has had the greatest impact on predictability - although in some respects that instability has made the game more exciting.
Sports Time Machine
Scott M. Berry and C. Shane Reese of Texas A&M and Patrick D. Larkey of Carnegie Mellon compare baseball, golf, and hockey greats from different times by taking into account important changes in the games and players. The authors present some unexpected conclusions in a paper, "Bridging Different Eras in Sports" at the Cincinnati Convention Center on Wednesday, May 5 from 9:45-11:15am.
The authors use what are called bridge and additive models to estimate the innate ability of players, the effects of aging on their performance, and the difficulty of playing in each year/season. Their operations research accounts for different players aging differently and the changing population in each sport. They look at the changing talent pool within each sport. They characterize players by their career profile, rather than a one-number summary of their career.
The paper has been named the best applied paper of the year by the American Statistical Association.
Sometimes, the authors' conclusions overturn classic sports views. In hockey, for instance, they rate Mario Lemieux as number 1 among top 25 peak players and place Wayne Gretzky at number 2. In baseball's home run contest, they rate Mark McGwire as number 1. Hank Aaron is only rated number 23, although he hit the largest number of home runs in major league history.
Locally, Cincinnati Red Pete Rose placed 15th among the top 25 peak baseball players for a batting average study. Looking at golf greats, the authors place Ohio-born Jack Nicklaus as No. 1 among the top 25 peak players in their golf study.
Among the authors' important observations are the following:
Baseball
- The ability of players to hit home runs has increased dramatically over the century. Many of the greatest home run hitters ever are playing now.
- Batting average does not have the same increase over the century. There is a gradual increase in the ability of players to hit for average, but the increase is not nearly as dramatic as for home runs.
- The best players are increasing in ability, but the median and bottom 10% of players are increasing faster over the century.
- It has gotten harder for players of a fixed ability to hit for average. This may be due to the increasing ability of pitchers.
Golf
- Golf players' abilities have increased substantially over time. The bottom 10% of golf players have improved by about two shots over the last 40 years, with the rest of the major competitors remaining of the same ability overall.
- The optimal age for a pro golfer is 34, though the range of 30-35 is nearly optimal. A golfer of 20 is about equivalent to the same golfer at age 50, with both about 2 shots below their peak.
- Examining a representative course and competition, the Augusta National, the equipment and conditions of today have made the course about one shot easier than in 1950.
Hockey
- Career scoring totals for individuals are greatly influenced by the era in which they played.
- In hockey, the enormous physical demands of the sport make the optimal age for a hockey player 27. There is a sharp decrease after the age of 30.
- A hockey player at age 34 - which is the optimal age for a golfer - is at only 75% of his peak value.
- Many of the greatest scorers of all time are playing now. NHL Hockey is a sport that has greatly expanded its talent pool in the last 20 years, and the number of great players has increased as well.
Operations Research
INFORMS® is holding its semi-annual convention in Cincinnati from Sunday, May 2 to Wednesday, May 5. The theme is "Delivering to the Global Consumer." It takes place at the Cincinnati Convention Center and the Omni Netherland Plaza. The convention includes sessions on topics applied to numerous fields, including commuter transit, e-commerce, health care, information technology, energy, transportation, marketing, telecommunications, and sports. More than 1,300 papers are scheduled to be delivered.
The General Chair is Professor David Rogers of the University of Cincinnati. The convention is underwritten, in part, by a generous grant of $125,000 from Procter & Gamble. For additional information on the conference, including a full list of workshops, visit http://www2.informs.org/Conf/Cincinnati99/
The Institute for Operations Research and the Management Sciences (INFORMS) is an international scientific society with 12,000 members, including Nobel Prize laureates, dedicated to applying scientific methods to help improve decision-making, management, and operations. Members of INFORMS work primarily in business, government, and academia. They are represented in fields as diverse as airlines, health care, law enforcement, the military, the stock market, and telecommunications.