Sunday, July 12, 2009

The Dao of Strategic Assessment (28): Assess and Predict


The first step to predicting the future is to assess the data source and its reliability. The next step is to assess the data through regressive means. Understanding the data from an active view, a functional view and a grand picture view enables one to see the trends. In future posts, we will touch on each view.

#

July 8, 2009
Seeking a Way to Predict Baseball Injuries
By MICHAEL S. SCHMIDT

LOS ANGELES — Beyond the massage tables and stacks of athletic tape in the Dodgers’ cramped training room stands a dry-erase board with scribbled questions like “Fat? Skinny?” and “Country of origin? Region of the country they grew up in?”

Stan Conte, the team’s director of medical services and head athletic trainer, expects the answers to help him resolve a problem plaguing professional sports teams: injuries.

In Major League Baseball, where players are breaking down in record numbers, teams paid about a half-billion dollars last season to players on the disabled list.

The ability to predict how players’ bodies will fare is a holy grail. With an actuarial approach, Conte seems to have a head start in the pursuit. He is trying to build a formula that would give teams a competitive advantage and help them avoid players who spend their days in the training room and not on the field.

“The insurance industry has made millions of dollars off figuring out how, when, where and why people are going to die, and we are trying to figure those things out about injuries,” Conte said.

Every major league team and scores of independent analysts are trying to understand why injuries strike certain players. But Conte said his effort is more advanced because his data has been compiled over 15 years as a trainer for the San Francisco Giants and now the Dodgers. He has had hands-on interaction with hundreds of players before and after their injuries, and his observations and medical records inform his hypotheses.

Conte also draws on the analytical prowess of David Zes and Adam Sugano, statisticians who teach at U.C.L.A., and Matt Marks, an employee in the Dodgers’ baseball operations department. They are building mathematical formulas that they hope will show the chances a player will be injured within the next season.

“This project couldn’t have been done 5 to 10 years ago,” Sugano said. “There weren’t enough numbers, and there wasn’t access to lots of the numbers. The numbers weren’t being collected, and those that were weren’t the cleanest. And 20 years ago, there wasn’t the computational power. It didn’t exist.”

Billy Beane, the general manager for the Oakland Athletics, called it “the natural progression of statistical analysis.”

Beane was at the forefront of baseball’s first surge in statistical analysis, as detailed in the 2003 book “Moneyball,” when he embraced new ways of evaluating players’ talents. Now Conte is applying similar methods to injury research.

“Injuries are a huge part of the game and it makes sense that they are doing it,” he said. “I just don’t have the money to let someone spend all year looking into this.”

Conte considers many of his findings proprietary, but in a recent interview at Dodgers Stadium he provided a glimpse of his project.

About once a week for the past six months, Conte has sent an e-mail message with lists of players’ identifying characteristics to Sugano and Zes. They build logarithmic formulas and computer codes that test Conte’s hypotheses, such as Dominican players being more durable than Americans and whether high pitch counts lead to injuries.

“He has been working and seeing things over his career that he thinks are trends, and he needs someone to execute them,” Sugano said. “We are the executors.”

Marks helps Zes and Sugano by scouring the Internet and baseball databases for statistics on players’ performance and injuries.

Although Conte said the project is far from finished, he has begun to apply some of his analysis to advise the Dodgers’ front office on personnel decisions, including free agents.

When the team’s contract with relief pitcher Scott Proctor was due to expire last year, Conte said he and his assistants conducted a risk assessment that considered Proctor’s medical history, the number of pitches he had thrown and his frequent appearances two seasons in a row.

“We did a risk analysis — I can’t tell you what it was, but we saw him go 83, 83 in appearances two years in a row and had some concerns,” Conte said. “Without getting into specifics, we ended up not tendering him a contract.”

Proctor signed with the Florida Marlins. He sustained an elbow injury in spring training and needed major surgery. He may miss the entire season.

Conte’s injury analysis has evolved. He said it is more sophisticated and, he hopes, more accurate. Three years ago, he advised the Dodgers that signing pitcher Jason Schmidt to a three-year, $47 million contract was not a high risk. Since then, Schmidt has had two shoulder operations and pitched in only six games.

“Everyone asked me two, three times a day, ‘If you had to do it all over again, how would you analyze Schmidt?’ ” Conte said. “I didn’t have all the answers with Jason Schmidt, which has pushed me to get better knowledge.”

A nagging challenge is the absence of a centralized database of injuries. All Conte’s team has to work with is Major League Baseball’s disabled list, a relic that includes errors and misinformation.

“People in M.L.B. don’t believe it, and we still don’t,” Conte said, adding that teams sometimes manipulate roster moves by putting players on the list who may not be injured. “But it is the closest thing we have to anything.”

Unlike the N.F.L. and the N.H.L., baseball does not have a central database of league-wide injury data. The commissioner’s office has begun trying to help teams track injuries reliably and consistently. But that project may take several years.

“Stan has been all over us with this trying to get all the data into one place,” said Chris Marinak, Major League Baseball’s director of labor economics. “The difference between Stan and other trainers is that he has taken a real vested interest in this.”

Conte said he was conducting his research on his own time, not as part of his job with the Dodgers. Considering all the money teams have invested in injured players, the formulas could become lucrative if successful. Conte said he has no agreement with the three others on sharing any potential profits.

“For now, we are having fun and that’s enough,” Sugano said. “It’s something we have talked about, but we will be careful. We will let them taste the sauce. And it’s something that won’t work if they don’t have our algorithms; we have the recipe and it won’t work without it.”

It is not clear whether other teams are attempting similar research because they are not inclined to discuss something that may provide a competitive advantage.

Lonnie Soloff, the head trainer for the Cleveland Indians, a team considered to have a progressive medical staff, said he used a mathematical formula to assess injuries but admitted, “it is not an exact science and I am not sure it will become an exact science anytime soon.”

The Boston Red Sox, another team with a progressive medical staff, declined to let their trainer be interviewed.

Conte believes the long-prevailing belief in baseball that injuries are a matter of chance is misguided.

“I refuse to think we are doing all these things to get them healthy, and it’s a matter of luck whether lightning hits or doesn’t hit,” he said.

Copyright 2009 The New York Times Company
http://www.nytimes.com/2009/07/08/sports/baseball/08injuries.html

No comments: