The Futility Infielder

The epic saga of the post-Jeff Nelson Yankee bullpen is one which I've been rehashing here for three seasons. By now, I could write an opera. Nelson's departure after the 2000 World Series created a void -- reliable righty setup reliever -- that the three-time World Champions have struggled to fill ever since. They tried faking their way through it in 2001 (Todd Williams, Carlos Almanzar, Briah Boehringer, Jay Witasick?), and then, admitting their mistake, threw huge dollars at Steve Karsay (4 years, $22 million) for 2002. But after last winter's shabby treatment of mainstay lefty Mike Stanton (fifteen minutes to accept a two-year, $4.6 million contract) and their willingness to set the similarly reliable Ramiro Mendoza free, the Yanks' karmic chickens began coming home to roost. The fragile Karsay turned up lame in the spring of 2003, a victim of shoulder tendinitis, and Mariano Rivera began the year on the DL with a groin strain.

One of the Yanks' major shortcomings over the Brian Cashman era has been a general ignorance of the secondary talent market. With the Yanks, every solution is an expensive one, a name brand player whose signing is carefully constructed to make a splash -- a directive straight from the throne of mad King George. So the Yanks threw their Stanton money at Chris Hammond, rewarding him for his first good season in seven years. They traded Orlando Hernandez for Pricey Proven Reliever Antonio Osuna. So it was something of a surprise to find that in this case, the Yanks also had a backup plan. Juan Acevedo, fresh off a 28-save, 2.65 ERA season as the Detroit Tigers closer, had been signed to a minor-league deal worth just under $1 million. With Rivera on the DL, the Acevedo signing looked brilliant. That would change soon enough.

I'll spare you the further stomach-turning drama of the Yanks' 2003 pen. The people who arrived over the winter fell out of favor, and some of them were replaced by more people who fell out of favor even more quickly. Finally some more people arrived, and the Yanks took them to the playoffs. The bodies of Acevedo, Osuna, Jason Anderson, Randy Choate, Al Reyes, Dan Miceli, Jesse Orosco, and Armando Benitez were strewn along the highway, while not-particularly-better bodies of Felix Heredia, Gabe White, and a not-as-good ol' Jeff Nelson went to the postseason. Those of us watching considered taking turns gouging each other's eyes out, as it was considerably less painful than what transpired when the bullpen door swung open.

In all, exactly three relievers topped 50 innings while wearing the pinstripes: Rivera, Hammond, and Osuna. Sterling Hitchcock, who spent the first half of the season auditioning as some fringe contender's stopgap rotation solution, just missed with 49.2 innings. From the rest of the bunch, only Acevedo topped 25 innings, and the aforementioned trio of Heredia, White and Nelson totalled only 45 innings. Since such small sample sizes don't yield fruitful evaluations, I'll examine the full season stats of these pitchers, dispensing with the bystanders, innocent or not (conveniently, many of these are free agents and will be dealt with accordingly). I'll use most of the same stats I did with starters, including DIPS ERA (dERA), but ignore the Won-Loss records. I'll also include Adjusted Runs Prevented (ARP), a useful metric from Baseball Prospectus which I'll explain below. Here's how the 2003 Yankee relievers performed:

Player        IP    ERA   ARP   K/9    WHIP   K/W    HR/9   BABIP  dERA

M Rivera     70.7  1.66  18.1   8.02   1.00   6.30   0.38   .294   2.57   

C Hammond*   63.0  2.86   9.9   6.43   1.21   4.09   0.71   .302   3.49

A Osuna      50.7  3.73  -0.6   8.35   1.54   2.35   0.53   .344   3.29  



S Hitchcock *88.0  4.72  -2.1   6.98   1.40   2.13   1.44   .287   4.71  

F Heredia*   87.0  2.69  12.4   4.66   1.23   1.36   1.03   .233   4.61

J Nelson     55.3  3.74  -8.4  11.06   1.36   2.83   0.65   .336   3.30

G White*     46.7  4.05   4.0   5.59   1.11   3.63   1.35   .257   4.14 

J Acevedo    38.3  6.57 -10.9   6.57   1.83   1.56   1.41   .343   4.72

Because relievers often come and go with runners on base, traditional ERAs tend to over- or understate their performance based on that of their neighbors -- similar to a hitter having more or fewer RBI based on who he's batting behind. ARP uses a run expectancy matrix to give a more accurate measure of a pitcher's responsibility; a positive number means a reliever did a good job at cleaning up other people's messes, a negative number means you may want his fingerprints on file in case of arson. I like the stat a lot because it adds some shading to actual ERAs; for example, Jeff Nelson was at -8.4, making his 3.74 ERA all the less impressive, while Chris Hammond was at 9.9, a pretty decent number to go along with his 2.86 ERA. But since high or low BABIP rates are reflected in ARP as well as in actual ERA, I'm going to exclude it from my larger analysis below.

As a whole, the Yankee bullpen had the same ERA as the starters (4.02). Mariano Rivera was, of course, the class of the bullpen, striking out batters, showing amazing control, and keeping the ball in the park. When healthy, he's still as good as it gets. Hammond didn't repeat his amazing 0.95 ERA/1 HR season, but he showed excellent control and pitched pretty well; unfortunately, he lost Torre's trust and sat idle in October while Jeff Weaver yielded a World Series-turning gopher ball. Osuna got little help from the Yankee defense and walked a lot of batters, but was otherwise pretty decent. He too lost Torre's trust late in the season and didn't make any of the October rosters. Hitchcock won his prized spot with the Cardinals about eleven minutes before they fell out of the NL Central race. Heredia and White were especially helped by their defenses (in Cincinnati as well as New York), which enabled them to get by without striking out a lot of batters; Heredia's control was nothing special, but White's was pretty good. Nelson fooled a lot of batters with his wicked slider, but he was particularly bad in handling other people's problems, and had control problems once coming over from Seattle. Acevedo pitched like a guy auditioning to man the boiler room of Hell.

Rivera and Hammond are both signed for 2004, while everybody else above is a free agent. This leaves the Yankees with an opportunity to overhaul the pen -- characteristically, in expensive fashion. As I write this, the Yanks have already made their moves, but I'll nevertheless examine them in the context of the free-agent class, shown here with their innings rounded to the nearest whole:

Player        IP    ERA   K/9    WHIP   K/W    HR/9   BABIP  dERA

L Hawkins     77   1.86   8.73   1.09   5.00   0.47   .302   2.57  signed CHI

T Gordon      74   3.16  11.07   1.19   2.94   0.49   .294   2.69  signed NYY

T Adams       68   2.65   6.75   1.34   2.22   0.13   .324   2.95

J Nelson      55   3.74  11.06   1.36   2.83   0.65   .336   3.01

P Quantrill   77   1.75   5.12   0.98   2.93   0.23   .260   3.07  signed NYY

D Plesac*     33   2.70   9.99   1.20   3.36   0.81   .292   3.18  retired

A Osuna       51   3.73   8.35   1.54   2.35   0.53   .344   3.29  

E Guardado*   65   2.89   8.27   0.98   4.29   0.96   .240   3.30  signed SEA

T Worrell     78   2.87   7.47   1.30   2.32   0.57   .291   3.33  signed PHI

D Holmes      42   4.29   9.86   1.38   4.18   1.07   .356   3.33

A Rhodes*     54   4.17   8.00   1.31   2.67   0.67   .312   3.36

K Foulke      87   2.08   9.14   0.89   4.40   1.04   .221   3.40

J Tavarez     84   3.66   4.20   1.22   1.44   0.11   .266   3.45

C Leskanic    53   2.22   8.54   1.27   1.72   0.34   .267   3.50  signed KC

U Urbina      77   2.81   9.12   1.13   2.52   0.94   .241   3.64

D Veres       33   4.68   7.16   1.26   5.20   1.10   .320   3.68

G Lloyd*      48   5.29   4.72   1.72   1.79   0.38   .365   3.68

M Timlin      84   3.55   6.99   1.03   7.22   1.18   .263   3.70  signed BOS

R Beck        35   1.78   8.15   1.02   2.91   1.02   .228   3.77  signed SD

T Martin*     51   3.53   9.00   1.18   2.13   1.06   .236   3.82  signed LA

S Hasegawa    73   1.48   3.95   1.10   1.78   0.62   .250   3.93  signed SEA

C Fox         43   3.12   9.55   1.52   1.48   0.62   .274   3.98  signed FLA

J Grimsley    75   5.16   6.96   1.65   1.61   0.72   .340   4.08  signed KC

K Ligtenberg  59   3.34   7.13   1.25   3.36   1.37   .291   4.09  signed TOR

S Sullivan    64   3.66   7.88   1.25   1.75   0.84   .239   4.11  signed KC

G White*      47   4.05   5.59   1.11   3.63   1.35   .257   4.14

A Alfonseca   66   5.83   6.92   1.55   1.89   0.95   .330   4.15

D Burba       43   3.53   7.27   1.41   1.84   1.04   .285   4.30

C Eldred      67   3.74   8.96   1.38   2.16   1.20   .291   4.33  signed STL

K Mercker*    55   1.95   7.81   1.41   1.50   0.98   .256   4.35

T Jones       69   7.08   7.73   1.81   1.90   1.31   .369   4.40

R Rincon*     55   3.25   6.51   1.39   1.25   0.65   .253   4.46

T Van Poppel  48   5.59   6.33   1.37   2.27   1.49   .281   4.49  signed CIN

M DeJean      83   4.68   7.73   1.51   1.82   1.42   .304   4.51

S Sparks     107   4.88   4.54   1.41   1.46   1.09   .286   4.60

F Heredia*    87   2.69   4.66   1.23   1.36   1.03   .233   4.61  signed NYY

R White       67   5.78   7.25   1.42   2.57   1.75   .303   4.64

J Acevedo     38   6.57   6.57   1.83   1.56   1.41   .343   4.72  signed PIT-m

D Miceli      70   3.20   7.42   1.19   2.32   1.66   .235   4.75  signed HOU

S Kline*      64   3.82   4.38   1.35   1.03   0.71   .249   4.80  signed STL

J Orosco*     34   7.68   7.68   1.82   1.38   1.06   .339   4.80  signed ARI

H Carrasco    38   4.93   6.34   1.57   1.35   1.17   .292   4.84

M Williams    63   6.14   5.57   1.70   0.95   0.71   .290   4.86

D Hermanson   69   4.06   5.11   1.37   1.63   1.18   .282   4.89  signed SF

S Reed        63   3.27   5.54   1.34   1.50   1.28   .267   5.04  signed COL

T Wendell     64   3.38   3.80   1.28   0.96   0.84   .233   5.06

J Wright      56   7.35   7.99   1.90   1.61   1.44   .381   5.07  signed ATL

J Franco*     34   2.62   4.19   1.40   1.23   1.31   .265   5.38

M Myers*      36   5.70   5.20   1.62   1.00   0.99   .281   5.40

M Guthrie*    43   2.74   5.06   1.45   1.09   1.27   .258   5.42

A Levine      71   2.79   3.80   1.35   1.03   1.14   .250   5.45

R Hernandez   60   4.35   6.75   1.73   1.05   1.50   .282   5.73

T Mulh'nd*    99   4.91   3.82   1.56   1.14   1.55   .292   5.87

J Fassero*    79   5.68   6.37   1.64   1.62   1.97   .309   5.87

Average       61   3.87   6.82   1.36   1.93   0.99   .285   4.30

Before we get too heavily into the discussion here, I'll again remind that many of these pitchers don't have a whole lot of innings to go on. One thing Voros McCracken showed was that even at low numbers of batters faced, DIPS ERAs correlate better with the following season's ERAs than the actual ERAs do. The reason for this isn't magic; the defense-independent stat line is built by regressing a pitcher's rate stats towards the league averages, which is where this group will head. If I had more time, I'd consider these pitchers' stats over the last two seasons, but since things have already shaken down quite a bit, I'm going to have to let that one go.

Anyway, the Yanks have done pretty well along these lines, signing the #2 and #5 pitchers on the board in Tom "Flash" Gordon and Paul Quantrill. The 36-year-old Gordon, who got a 2-year/$7.25 million deal, has the high K-rate, good control, and low homer tendencies that make DIPS salivate (he was the top-rated righty in my 2002 free-agent reliever analysis). His experience as a closer gives the Yanks a fallback if Mariano Rivera needs some rest. He's got a reputation for fragility -- Tommy John surgery in 1999, a torn shoulder muscle in 2002 -- but was healthy in 2003, the first time since 1998 that he topped 50 innings. It's still mindblowing to think that this guy won 17 games as a starter for the Royals back in 1989, on a team that included George Brett, Frank White, Willie Wilson, Bob Boone, and Bo Jackson. A bit pricey, but not a bad signing by the Yanks.

The 35-year-old Quantrill, who the Yanks got for 2 years/$6.8 mil, is coming off an outstanding season for the Dodgers. He's been a good reliever for a long time, with a career 126 ERA+ (in other words 26% better than league average). And he's durable, averaging 85 games a year for the past three seasons and 75 over the past seven; his 89 games last season are tied for 8th on the all-time list, and he's got three of the top 17 spots up there. His strikeout rate is nothing special, but his control is good and he gives up very few homers (only 3 in the past two seasons). This guy is a rock, and though pricey, he's another smart signing.

Heredia, 28, turned down an option of $1.7 million to sign a 2-year, $3.8 million deal with the Yanks. As I said above, he was helped a lot by his defense, with that .233 BABIP, something that might be of concern with the Yanks' middle infield woes. But what's strange is the way Heredia's statistical pattern has evolved. From 1997-99, his K rate was above 8.0, and it was above 7.0 in 2000 and 2001. In those years he was basically a flyball pitcher, with grounder/flyball ratios just under 1.0. But his K rate has dropped to below 5.0, and he's become a groundball-dependent pitcher (1.56 G/F ratio). His number of pitches per batter has fallen as well, from about 3.8-4.0 in the high-K years to 3.44 last season. All of this would tend to indicate a conscious choice in preserving his arm and trusting his stuff, and it seems to be working; his ERAs in 2002 and 2003 have been the lowest of his career by far. If he can improve his control while continuing this groundball evolution, he might be the Ramiro Mendoza Liner the Yanks need out of the pen.

The Yanks are rumored to be close to re-signing White, a 32-year-old lefty flyball pitcher. As I said above, his control is good. But his strikeout rates have fallen without the clear pattern that Heredia exhibits; his pitches per batter and G/F ratios have remained pretty stable. A bit above average, especially with the Death Valley of Yankee Stadium's left centerfield, but really nothing special.

Jeff Nelson looks very good in the above analysis, thanks mostly to that awesome strikeout rate. Under normal circumstances, it wouldn't have been a surprise to see the team re-sign him. But in addition to his maddening late-season inability to get ahead of hitters, and despite all of his past glory in the Bronx, Nelson's a guy whose welcome has worn thin 'round these parts. Over the course of the season, he creates more than his share of headaches, even if they pale next to those of a David Wells. Credit the Yanks for moving on rather than being fooled by his jewelry. Still, somebody out there will get a pretty good reliever.

With Rivera, Karsay, Hammond, Gordon, Quantrill, Heredia, and presumably White, the Yanks will not only have the most expensive bullpen in the history of Western civilization, they have too many pitchers. White's signing would presumably come at the expense of Hammond, whose squawking about New York's tough crowds probably did more for his late-season disappearance than his pitching; he allowed only two runs in 21.1 innings in August/September, yet pitched only two innings in October. Grrrr... Hamond's a better pitcher than White, and the only way this truly makes sense is for the meager dollars it will save (my guess is White will approach but not top $2 mil). That's hardly worth the bother on a $200 million payroll.

Time prevents me from making more of this stuff right now (am I writing like I've got a plane to catch? very well, then, I have a plane to catch). Looking around, it's surprising to see that Proven Closer Eddie Guardado got more money than former teammate Latroy Hawkins, the top-rated reliever here. Guardado looked good thanks to a low BABIP rate (.240) while Hawkins was lights-out even with a .302 BABIP. The Hawkins signing (3 years, $11 mil) -- along that of Shigeosi Hasegawa and those made by the Yanks -- helps to demonstrate that though the overall market for player salaries seems to be depressed, the top-flight middle relievers are still getting paid. Their relatively low salaries ($3-4 million per year) don't compare to those of position players or closers, so mistakes aren't as costly, and if the shit hits the fan there's always a cheap replacment out there somewhere.