Koji Uehara: the Cliff Lee of the bullpen?

Just about a week ago, the Red Sox announced that they had finalized terms with Koji Uehara on a one-year deal worth $4.25M. The 37-year-old Japanese reliever strained his lat muscle in mid-June of last season, limiting him to 36 innings, but while the quantity of his innings was affected the quality did not appear to be. Since entering the league in 2009, Uehara has been one of the best relievers in baseball at preventing the free pass, sporting a career walk rate of 1.23/9 IP and coming off season in which he allowed a miniscule 0.75 BB/9. Uehara’s no slouch at punching batters out, either, with a career rate of 9.82 K/9 despite averaging only 89 MPH on his fastball. Add it all up, and you end up with a pitcher who’s been among the top three relievers in K/BB for three years running, including a 14.33 last season that was tops among all pitchers in baseball.

It strikes me that Cliff Lee’s taken a similar route to success. Unlike essentially every ace in baseball, Lee developed into a top-of-the-rotation starter from relative obscurity. This is a fairly rare phenomenon, because in order to be an ace, you have to be able to dominate, and dominant stuff generally doesn’t go under the radar. There are dozens, maybe hundreds, of pitchers with “ace potential” who flame out or force expectations to be adjusted for every one that becomes a true staff ace. However, besides Lee, it’s nearly unheard of for a player to develop into an ace without scouts and analysts recognizing that potential during the player’s time as an amateur or in the minors (Canuckleballer R.A. Dickey excluded). Justin Verlander, for example, ranked eighth on Baseball America’s 2006 Top 100 Prospects. Felix Hernandez first appeared at the 30 spot in 2004, then moved up to BA’s second best prospect in baseball (trailing only Mauer) for 2005. Clayton Kershaw was 24th in 2007, then seventh in 2008. Sabathia was 57th in 2000, then seventh a year later. Halladay, Greinke, Cain… basically, name a true ace and you can bet that his prospect status at one point reflected his current standing.

Lee is the exception. The Expos selected him in the fourth round in the 2000 amateur draft out of Arkansas. In 2002, Lee was dealt to Cleveland in the ill-fated Bartolo Colon trade, as he and Grady Sizemore served as secondary trade chips with Brandon Phillips (BA’s 20th-best prospect that year) as the centerpiece of the deal. Lee threw 10.1 major league innings in 2002, and finally appeared on BA’s top 100 list at 30 following that season.

It’s not hard to understand why Lee flew (relatively) under the radar. While most aces dominate with eye-popping velocity, Lee’s fastball clocks in at an average velocity of just under 92 MPH. As a result, he rarely put up the strikeout totals generally associated with aces, at least until he showed up in Philadelphia (for the second time). Until his breakout 2011, Lee had never struck out more than 185 batters in a single season. However, Lee was clearly seen as an ace throughout the league before 2011, as he turned down at least two offers guaranteeing more total dollars to sign a $120M, five-year deal shortly after the 2010 season. While Lee’s strikeout rate didn’t suggest he deserved that kind of deal, his K/BB (among numerous other factors) convinced the Phillies otherwise. In 2008 and 2009, he finished in the top 10 among starters in K/BB rate. Then, in 2010, he put up an absolutely ridiculous 10.28 rate to lead all starters. While his strikeout rate of 7.84 K/9 wasn’t significantly better than the league average of 7.13, his 0.76 BB/9 rate blew away the competition. However, this formula doesn’t necessarily produce the results you might expect.

If you’re a big sabermetric buff, or a big fan of THT’s writers, you might have caught Glenn DuPaul’s excellent article on Beyond the Box Score, “Stop Using K/BB!” In it, Glenn presents a well-reasoned argument as to why it makes much more sense to use percentage-based stats, like K% and BB%, rather than ones like K/9 and BB/9 with innings pitched in the denominator. Instead of K/BB, Glenn argues, we should consider using K%-BB%, and supports that argument by showing that K%-BB% correlates more closely with run prevention than K/BB does. K/BB overrates the value of preventing walks, especially in the case of pitchers with extremely low walk rates (like Uehara and Lee).

Of course, this doesn’t make Lee a slouch. It’s possible that the Phillies knew he had bigger strikeout totals in him when they signed him, or Lee might have realized that in order to live up to his huge contract he’d need to deliver more punchouts than he had up to that point in that career, and that his new ballpark was not kind to pitchers who allowed bigger numbers of balls in play. Whatever the reason, Lee’s had the two biggest strikeout totals of his career by far in the last two seasons. His 445 punchouts are tied with King Felix for fourth in baseball over that span, trailing only Verlander, Kershaw, and the newly anointed ace of the Royals’ rotation, James Shields. Lee’s walk rate rose somewhat from his absurd 2.1% walk rate in 2010 (really, how could it not?), but he’s arguably a better pitcher because he’s struck out a quarter of the batters he’s faced, beating the league average by about 6 percent rather than finishing just above it as he had throughout his career before the contract.

With this in mind, I think the Red Sox paid a pretty reasonable price for Uehara. Most premier setup men who signed this winter got deals for multiple years and $6M+. Uehara’s in the next tier, a capable late-inning reliever but probably a guy you’d rather have pitching in the seventh than the eighth. His deal is the largest one-year pact signed this offseason for any reliever besides Mariano Rivera. Going by K/BB, Uehara’s among the best relievers in the game. However, investigating Uehara a little deeper and understanding the flaws in the statistic he dominates show why the Red Sox were right to pay him at a rate that reflects his 2012 K%-BB%, a still excellent but not elite 30.8%, much more so than his K/BB (14.33).

1 Comment
Newest Most Voted
Inline Feedbacks
View all comments
10 years ago

Canuckleballer may be the best word I’ve ever heard.