March 11, 2009

Checking the Numbers

Pedro's Tank

by Eric Seidman

Printer-friendly

Contact Author

When Pedro Martinez toed his home rubber to kick off the 1999 All-Star Game, fans across the globe knew they were in for a special performance. The game's best pitcher had breezed through the offense-heavy junior circuit in the first half, compiling a 2.10 ERA, 184/24 K:BB, and a Rafael Belliard-esque .213/.254/.292 opponent's line. On national television against some of the best hitters in the game, Pedro did not disappoint, fanning five of the six hitters he faced, and making each look foolish in the process. That image of Martinez is ingrained in our minds: a dominant and diminutive Dominican capable of shutting down and intimidating anyone who stepped into the box.

Ten years removed from that moment, the owner of arguably the two best-pitched seasons in major league history and a surefire Hall of Fame pitcher lacks employment, and has resorted to using the World Baseball Classic as an audition. After a disappointing 2008 season, does he have anything left in the tank?

Over the last few seasons, Pedro has fallen well short of expectations, a fact that he fully acknowledges. After consecutive seasons of 31-plus games started and 217 innings pitched in 2004-05, Martinez has averaged just 16 games started and 90 IP over the past three years, exhibiting a tendency for injuries that only exacerbates the inevitable performance decline of a pitcher in his mid-30s. Interestingly, despite extended trips to the disabled list due to rotator cuff and hamstring issues, as well as a slew of other injuries that hampered his performance on the field, Pedro's overall strikeout rate since 2006 has not suffered as substantially as some might expect. Here are some of his seasonal averages:

Seasons  GS    IP   K/9 BB/9 HR/9   NRA  DERA
2004-05  32  217.0  9.0  2.2  0.9  3.35  3.49
2006-08  16   90.0  8.5  2.8  1.3  5.03  5.57

Overall averages can be quite deceiving, however, because they hide year-to-year shifts. Pedro's 2006-08 strikeout rate may be 8.5 per nine innings, but the average is largely comprised of a 9.3 K/9 in 2006 and a career-worst 7.2 K/9 last season. Combine the relatively low strikeout rate with a 3.6 BB/9 mark that happened to be his highest since 1993, and it becomes a little easier to understand what went wrong last season. The underlying cause for these rate changes can be found in plate discipline metrics, which show a particularly disturbing trend since his last award-worthy season in 2005:

Season O-Swing O-Cont Z-Swing Z-Cont
 2005    23.3   46.9    64.8   83.1
 2006    24.9   52.4    66.7   83.7
 2007    27.8   44.8    68.8   82.4
 2008    27.3   67.5    59.8   85.4

League Averages: O-Swing=24.4%, O-Cont=59.9%, Z-Swing=66.3%, Z-Cont=88.1%

Ignore the 2007 season for a moment (Pedro made just five starts), and instead compare the years before and after. Pedro induced swings out of the zone at a higher clip last season, yet the contact on these swings soared to heights he had never before seen. Likewise, hitters proved more efficient on pitches thrown in the zone, as their zone contact rate increased despite a lower percentage of swings. All of this signals that Pedro had lost much of his ability to fool hitters, which led to fewer whiffs, more balls in play, and a heavier reliance on his defense, which went well for Martinez given the abilities of the Mets' defenders throughout his tenure.

Pedro did not make any significant changes in his pitch repertoire between these two seasons, and he held the fort in the velocity department as well. Unfortunately, there is no Pitch-f/x data with which to compare his 2008 numbers, but the 41 percent ground-ball frequency last season, coming on the heels of a steady decline from 42 percent to 31 percent since 2002, suggests that his movement may have fluctuated. He has a reputation for being able to dial up velocity when the situation merits more liveliness on the fastball, but Pedro's velocity remained very consistent last season, and as we discussed last week, the ability to add or subtract miles per hour from a fastball really means nothing in regards to overall production.

Despite reaping the benefits of solid defense and keeping much of his pitch data in check, Pedro's batting average on balls in play increased from .273 in 2006 to last season's mark of .327, a jump of 56 points thanks to a higher rate of line drives allowed. The standard statistical reaction to a discrepancy this vast is to suggest regression to the mean in the coming season, but if Pedro has lost part or most of his ability to fool hitters, regression may only rear its' head slightly, if at all, especially given his health history and that he's no longer a spring chicken.

PECOTA recognizes these factors, and only sees things getting worse, projecting Pedro for just a 6.1 K/9 and 4.7 EqERA in 110 innings of work. Pedro's recent track record makes him more of a liability than an asset, a perpetually injured player spitting in the face of his personal "best shape of my life" claims. Even with the substantial decline in controllable skills and performance metrics, Pedro's weighted mean projection calls for 1.9 WARP, almost the definition of a league-average pitcher, and equivalent to the weighted mean projections for both Jarrod Washburn and Freddy Garcia. Of course, this value is only expected to be present in a rather minimal amount of playing time, and teams in this economy are shying away from risky investments, unless the contracts feature low base salaries with lots of room for incentives, or non-roster invitations.

Pedro Martinez has always been a very proud person, so it should come as no surprise that in spite of declining abilities and health issues, he has repeatedly expressed an unwillingness to sign either of these types of contracts. Early in the offseason, he reportedly turned down a one-year, $7 million offer from an unnamed American League club that was perhaps convinced his veteran savvy and potential upside would be a worthwhile investment, but based on his weighted mean of 1.9 WARP, an appropriate salary for Martinez would be closer to $3.7 million, about half the money offered in the deal he apparently chose not to sign. Pedro has been linked to the Marlins and Pirates, both of whom have expressed limited interest, and he's currently engaged in a one-sided game of hard-to-get with the Mets, who already have enough fifth starters in camp.

The most interesting idea has been put forth by the Cardinals, who are pondering whether or not he could be an effective closer, a role Pedro seemed wary of but willing to consider. Closers have a reputation for being flamethrowers and strikeout machines, and while Trevor Hoffman has certainly succeeded with a Moyer-ball, his strikeout rates remain impressive, even coming in at 9.1 per nine innings last season. Pedro has lost velocity over the past four years, and he has also experienced a decline in whiffs per nine, meaning that adjustments would need to be made in order to avoid becoming a Dan Kolb type of closer.

Health permitting (an issue which looms much larger than the space those two words occupy), Pedro could fill out the back end of a rotation for any team still considering its options for fifth starters, as long as the expectations are along the same lines as those bestowed upon John Smoltz and Brad Penny, pitchers who are likely to contribute, but whose availability will not make or break a club's season. Where could Pedro land? Operating under the assumption that he wants to stay in the senior circuit, which teams currently need a fifth starter? Pedro would have been a better signing for the Braves than Tom Glavine, but Glavine is already in Atlanta's fold. The Rockies, with Greg Smith, Jason Marquis, and Jorge de la Rosa rounding out their rotation, could certainly benefit from a splash of Pedro, but color me skeptical that Martinez would be willing to pitch in Denver. The Marlins could platoon Pedro with Chris Volstad in the fifth spot, but the team does not seem to eager to make him an offer.

After probing the remaining National League teams, it seems that the one most in need of Pedro's services would be the Houston Astros, who will be relying on perhaps the worst starting pitcher in baseball in Brandon Backe, and also counting on the oft-injured Mike Hampton and Brian Moehler. Considering his projected performance, there are simply not many teams that would be vastly improving their rosters by adding Martinez, and if the former All-Star is only going to provide a marginal level of production over either prospects or freely available talent, why bother?

Then again, not all teams share this philosophy, and Pedro's celebrity status may very well come into play. Based on his work ethic, and on his proven ability to adjust in terms of repertoire and strategy as he's grown older, I'm comfortable with the assessment that Pedro Martinez is not exactly running on empty, but rather on a quarter-gallon in a vehicle with a faulty transmission. Some team is going to decide to take a chance and bring him aboard, but Martinez will likely need to get past his ego and accept a smaller role and an incentive-laden deal. When he does sign, the 13-year-old inside me who witnessed the dominance in that 1999 All-Star Game will be pulling for a career renaissance, but at this point, Pedro Martinez is a fifth starter with just a little bit left to contribute, whose past success may garner present and future opportunities even if the actual production fails to justify the playing time.

Eric Seidman is an author of Baseball Prospectus. 
Click here to see Eric's other articles. You can contact Eric by clicking here