NOT MANY GUYS WHO BATTED .249 over their major-league careers still have name recognition a dozen years after they leave baseball.
For that, Bill Pecota can thank Nate Silver.
Silver, a 27-year-old Chicagoan, has devised a system that projects how a player will do in the future by comparing him to past players of similar skills and productivity. It has become a useful tool — or a wonderful toy, depending on your outlook — for major-league team officials, player agents, baseball fans and fantasy league participants.
Silver named his system the Player Empirical Comparison and Optimization Test Algorithm — PECOTA for short, a tip of the cap to the obscure former Kansas City Royals infielder. So every year, around this time, when the numbers for the next season come out, baseball enthusiasts of all types are talking PECOTA. Or trying to.
“It’s incredibly accurate and incredibly confusing. It amazes me,” says Jayson Stark, senior baseball writer for ESPN.com. “I don’t know how someone can sit back in his chair at home and invent something like this.”
But that’s what Silver did.
“It was just something I did on my own,” he says. “I was pretty confident in it. It was an accurate system, and I thought I could sell it. … What it is is a more interesting way to project what a player will do.”
When Silver explains his system, he talks about categories familiar to even casual fans — batting average, walk rate, strikeout rate, height, weight, batters faced per game. Things get fuzzy when we get into areas such as attrition rate, break-out rate, equivalent runs per plate appearance (EqR/PA for short) and Weighted Mean forecast.
“I’m afraid to actually use PECOTA in a column,” Stark says, “because I don’t know if it’s reached the level where you can casually drop it into baseball literature without the average fan saying, ‘What is he talking about?’ And you can’t explain it.
“Another really interesting stat I’m familiar with is VORP. Value Over Replacement Player [in a nutshell: the number of runs contributed beyond what a replacement-level player would contribute to a team]. You can explain that to people. But PECOTA, how’d you explain that to people between commas?”
Gary Huckabay, a consultant for several major-league teams, understands it; “Nate isolated five components how a player provides value . . . then PECOTA looks through history to find similar players who have similar performance shapes at the same age. It takes a look at all those guys in the past who have been similar, then takes a look at what their careers did after a certain point.”
Silver says 40 or 50 categories go into the final analysis. Some can be gleaned from box scores. Others are a little more arcane. “Ground ball to fly ball ratio,” he says. “We look at minor-league stats. We even look at Japanese League stats. Cuban and Korean League stats.”
A 2000 graduate of the University of Chicago with a degree in economics, Silver caught the baseball bug when he was 6, growing up in East Lansing, Mich. It was 1984, the year the Detroit Tigers won the World Series. The Tigers became his team and baseball his sport. And if there’s anything that goes hand in glove with baseball, it’s numbers, another of Silver’s childhood interests. (“It’s always more interesting to apply it to batting averages than algebra class.”)
After college, he took his love of numbers to accounting giant KPMG, where he was an economics consultant for four years.
In his spare time, he began putting PECOTA together because he wasn’t satisfied with other performance projection systems. In 2002, he pitched PECOTA to Baseball Prospectus, an Internet company dedicated to a better understanding of baseball.
Huckabay, who had founded Baseball Prospectus in 1994, liked Silver’s system:
“He and I kind of have similar backgrounds. We’re both KPMG alumni and have an analytical bent. I had had a performance forecasting system called Vladimir, and Nate had an idea to do something similar but with more depth of information. And so Nate started doing the PECOTA as the cornerstone for the fantasy baseball product we were going to offer.”
Information aplenty
Baseball Prospectus has been offering PECOTA for the last three years (a premium BP account runs $39.95 a year and gives a subscriber PECOTA as well as a wealth of articles; there’s a $19.95 subscription that’s heavy on stats and without the articles, aimed at fantasy league participants). And PECOTA is no longer just a Nate Silver production.
“I might be `the PECOTA guy,’ but it very much is a team effort,” he says of the BP staff. “We all do it. It’s my baby, but it takes a village to run a PECOTA.”
And everyone from fantasy-leaguers to major-leaguers benefit.
“I think we have among our subscribers 60 subscriptions for major-league baseball teams,” says Silver, who left his KPMG gig to become the executive vice president at BP. “Maybe that’s an intern. Or it might be a general manager or owner. But I will say we have a good relationship with 25 of the 30 clubs.”
The White Sox are believers. According to spokesman Scott Reifert, PECOTA is one of several systems the team uses. He says the Sox’s baseball operations people consider it “one of the top couple of projection systems out there.” A Cubs spokesman said that the club uses a projection system, but declined to be specific.
On a smaller scale is Anil Kashyap. A professor of economics and finance at the University of Chicago Graduate School of Business, he consults it for his fantasy league team.
As an example of how Silver’s system works, Kashyap points to last year’s PECOTA for slugger Jim Thome, who was traded by Philadelphia to the White Sox in November. It compared Thome to retired sluggers Willie McCovey, Mark McGwire, Harmon Killebrew, Boog Powell and Reggie Jackson.
“That’s the kind of career path Thome is on,” Kashyap says. “But [Silver] also tells you which season for those players looks the most like the season Thome was supposed to have [in 2005]. And for four of the five people on the list, [PECOTA] saw a significant deterioration in playing time that next year. They got injured, just like Thome did [he played in only 59 games last season because of elbow and back injuries].”
What is PECOTA thinking for 2006? Silver sees a big upside for pitchers B.J. Ryan, A.J. Burnett and Kyle Farnsworth, and Tampa Bay outfielder Aubrey Huff, who has been mentioned in numerous trade rumors. He’s not as excited about Thome (“when a guy hits his 30s and starts getting injured, they keep getting injured”) or Johnny Damon, who recently left the Boston Red Sox and signed a four-year, $52 million contract with the New York Yankees (“He’ll miss Fenway”). PECOTA sees Damon hitting .290 with 13 homers and 68 RBIs next season, and then declining slowly over the next five years.
A guessing game
Still, any projection system is just that, a projection.
“[Silver] is always advertising the `deadly accurate’ PECOTA system,” Kashyap says. “It’s a little bit ironic because he knows it’s so difficult to predict any one person’s performance. This may be pretty good, but the margin of error on any one forecast is huge. But, you know, he’s making [1,600] of them . . . You average it out and the thing typically works pretty well.”
Now if he can just get the fan in the stands clued in. That day may come, Stark says.
“If it continues to be as accurate and indispensable a tool as it is, people will understand it,” he says. “We’ll reach a time when they’ll list a player’s stats, and it’ll have average, home runs, runs scored . . . and PECOTA.”
———-
bhageman@tribune.com
