SI Vault
 
Baseball's Next Top Models
ALBERT CHEN
April 06, 2009
Statistical models, that is. Systems for analyzing defensive play are feeding the new rage in the poststeroid era—run prevention. Teams are investing big money not only in slick fielders but also in the advanced metrics that are the game's last stat frontier
Decrease font Decrease font
Enlarge font Enlarge font
April 06, 2009

Baseball's Next Top Models

Statistical models, that is. Systems for analyzing defensive play are feeding the new rage in the poststeroid era—run prevention. Teams are investing big money not only in slick fielders but also in the advanced metrics that are the game's last stat frontier

View CoverRead All Articles

ASK TORII HUNTER where he ranks among the best defensive centerfielders, and he appears perplexed. "You mean, where I rank with guys now, or in history?" he says. The Angels' Hunter takes his defense as seriously as you'd expect from someone who says, "To catch a ball, I'd commit suicide." And to most of the baseball world, for that matter, it's conventional wisdom that he's one of the best fielders, period. Try to prod Hunter into naming another American League centerfielder in his class defensively, and he just shakes his head. The Indians' Grady Sizemore, perhaps? "Nah," Hunter says. "He's got some work to do. He takes bad jumps." The Tigers' Curtis Granderson? "He's up there, but he has to learn to take better routes," Hunter says. "You shouldn't dive as much as Grandy dives." Hunter's replacement in Minnesota last season, Carlos Gomez? "Dude is quick, but he also goes from points A to B to C to D when he should be going A to B," says Hunter. "And he's too aggressive with his throwing. Just look at his errors. [Gomez had eight.] I had none. Zero." He flashes a smile and shrugs. Case closed, as far as he's concerned.

And few would disagree. Certainly not the managers and coaches who last fall voted him to an eighth straight Gold Glove award. (Only five outfielders in history have won more.) Certainly not the Angels, who last winter lured him off the free-agent market with a $90 million contract, as much for his defensive reputation as his bat. And certainly not the fans, players, scouts and other baseball cognoscenti who favor traditional fielding statistics—errors, fielding percentage (Hunter's was a perfect 1.000) and putouts (his 350 were fifth among AL outfielders)—and watch the familiar sight of the 33-year-old gliding gracefully over the grass and, on occasion, scaling the outfield wall to rob a batter of a home run. To the eye, there is nothing to indicate that Hunter is anything but what he thinks he is: an elite centerfielder, the best in the American League, possibly one of the best ever to play the position.

But here's a flash for Hunter: Comparing players' defensive skills is no longer as scientific as sizing up Best Supporting Actress performances before the Oscars. In his first Bill James Abstract, in 1977, the oracle of statistical analysis lamented the inability to quantify defensive success with anything other than such antiquated statistics as errors and fielding percentage. It has taken three decades, but the mystery of defensive analysis, perhaps the last frontier in the statistical ether, has been cracked by sabermetricians who have devoted 15, 20 years to the cause. The clunkily named metrics that have emerged within the last five years may sound like topics at a symposium for mechanical engineers—Probablistic Model of Range, Defensive Regression Analysis, Special Aggregate Fielding Evaluation, Ultimate Zone Rating—but not only have they become accepted by analysts like James as accurate tools, they have also infiltrated the daily vernacular of front offices.

But major league clubhouses? Not so much. "The Probablistic Model of who?" asks Hunter, after he's told where he stands by measure of the metrics. Not only does he rate below Sizemore, below Granderson and below Gomez, Hunter was also regarded across the board as a merely average fielder, and in many instances, below average—which is the case according to Bill James's disciple John Dewan, author of The Fielding Bible and creator of the Plus/Minus Runs Saved metric, which six years ago did rank Hunter as the league's top centerfielder. Now Dewan's numbers show that Hunter has been steadily slowing down, and that last season he made plays on five fewer balls than an average centerfielder would be expected to make, costing the Angels four runs. The best defensive centerfielder in the AL? Dewan's calculations says it's Gomez, who tracked down 14 more balls than the average centerfielder and saved Hunter's old team 16 runs. "If I've lost a step, I'm still better than the average person," Hunter huffs. "When I need a walker, I'll go to rightfield and be the best rightfielder in the game."

In addition to telling us that Hunter is an average defensive outfielder, the metrics also suggest that the Yankees should relocate Derek Jeter, long rated by the metrics as the worst-fielding shortstop in baseball, to the outfield and that the Cubs' Alfonso Soriano, because of his strong arm, has saved more runs than any other leftfielder since moving from second base in 2006. In a poststeroid era, when scoring and home run totals have fallen as fast as the NASDAQ and speed and defense are becoming as important and as appreciated as they were during the Whiteyball days in St. Louis two decades ago, these metrics are becoming essential tools for winning organizations.

"There are still teams stuck in the Dark Ages," says one American League general manager, "but the secret's getting out. Defensive metrics have almost caught up to the offensive side. Some people would say they didn't think they'd see this day. But the revolution's here."

IN 1982 John Dewan was an actuary living in Chicago when a coworker handed him a copy of the Bill James Abstract. Dewan, a die-hard White Sox fan who grew up playing the baseball simulation board game Strat-O-Matic, was instantly hooked. Two years later he was sitting at his kitchen table reading one of James's articles about creating an organization of volunteers who would record detailed play-by-play information not found in the box scores of every major league game. "I put the book down, and I went to the phone and called directory assistance in Lawrence, Kansas," says Dewan. "I got Bill James's assistant on the line and signed up immediately."

Dewan became the director of the organization, Project Scoresheet, a year later while continuing to hold down his day job. Soon after quitting his actuary job in '87 to devote himself full-time to statistical analysis, he invented a metric called Zone Rating, in which he took play-by-play data and calculated the percentage of balls fielded by a player in his defensive zone, as well as balls outside of his zone. (By these metrics, Jeter always ranks among the lowest shortstops because he doesn't get to many balls outside a shortstop's zone.) Ten years later companies such as STATS Inc. and Baseball Info Solutions (which Dewan cofounded in 2002 with Steve Moyer) began hiring armies of new college grads who collectively would watch every game and keep a detailed log of what happened to every batted ball: what kind of pitch was hit, where the ball was hit, how hard it was hit, who fielded it and how it was or wasn't turned into an out.

THE DATA has given Dewan and other analysts the power to compute, with great precision, a player's ability to turn batted balls into outs. In 2003, in a forum on the website Baseball Think Factory, a professional poker player living in Las Vegas named Michtel Licthman introduced, in a 6,800-word primer, a metric that he called Ultimate Zone Rating (UZR). Lichtman was crunching numbers with data he purchased from STATS Inc.—paying nearly $10,000 for it annually—and, like Dewan, was measuring the runs saved or lost by every fielder compared to the league average at his position. His UZR model was similar to the plus/minus system that Dewan had come up with, but with more parameters for each batted ball; among them, the ballpark, whether the pitcher and batter were left- or righthanded, and the ground ball and fly ball tendencies of the pitcher.

Gradually baseball people outside the sabermetrics community began to take notice. During spring training in '04, Dewan was giving a presentation to the White Sox' front office in the team cafeteria when manager Ozzie Guillen and his players wandered in to have lunch, their game that day having been rained out. Dewan noticed that Guillen would occasionally glance over at the presentation. Eventually he walked up to Dewan and started flipping through his statistical samples. "If they had this s--- when I was playing," the manager announced to the room, "I would have been the best f------ shortstop who ever lived."

Continue Story
1 2 3