February 19, 2005

Starting Pitching, 2005.  1983?  1993?

Not to pick on Bob Foltman again, but I think we should get him away from Ozzie’s Optimism-flavored Kool-Aid.  In today’s Chicago Tribune, Foltman has the audacity to compare the starting pitching of the 2005 Sox to a couple of very famous, division-winning stables: “The White Sox have assembled a rotation that could be among the best in baseball and rival the team’s playoff starters of 1983 and 1993.”

Bobby!  Spit it out!

At least former GM Roland Hemond tempered his comments: “There are some comparisons.” 

True -- they have two arms each and throw from 60 feet, 6 inches.

If the 2005 starting pitchers perform like the ‘83 bunch, we’ll need to track down Mike Squires to climb the foul pole and hang the 2005 Division Championship Pennant.  And if they perform like the studs of ’93. . . well, who gets to be World Series Parade Marshall?  Minnie Minoso or Harold Baines? 

Look, there’s optimism and there’s optimism, but Foltman’s going out on a long limb here.  Does he remember just how good the 1983 Sox were?  Dotson, Hoyt, Bannister. . . even I remember sitting next to my father on the first-base side, watching Bannister rack up K after K against the hapless Mariners on the way to a two-hit shutout on a fine September night in 1983.   I was seven years old, but even with even young memories I know there’s no Floyd Franklin Bannister on the 2005 Sox.  And there’s certainly no Rich Dotson or Dewey LaMarr Hoyt. 

How can I make such statements?  After all, Bannister was only 16-10 in that fabled season of 1983 – surely that will be matched by a few of the Sox starters this year, right?

In our sabermetrics course, we talk about the limitations of traditional baseball statistics.  Wins and Losses can be misleading because they depend on a pitcher’s offensive support.  ERA can be distorted by a pitcher’s surroundings – the defense behind him, the park he pitches in, and the run-scoring environment of the league and era.  It would be nice to filter out all that noise to create a metric that says, “How does this pitcher do relative to some random guy that we could call out of AAA for a spot start?”  We’ve got such a stat – promulgated by Baseball Prospectus – called VORP for pitchers.

VORP stands for Value over Replacement Player.  It’s a measure of runs prevented above the amount a “replacement” pitcher would allow, given an equal number of innings.  When you think of replacement-level, think of young Danny Wright on the 2001 Sox. Over a dozen starts he posted a 5-3 record with a 5.70 ERA.  Gave up 78 hits and 39 walks in 66 IP.  An unremarkable performance for a team that finished in 3rd place in a weak division.  The Sox tried many starting pitchers that year:  Buehrle, Biddle, Baldwin, Wells, Wells, Glover, Garland, Wright, Lowe. . . and for all the good Wright did them, they may as well have left him in AAA.

For a more recent cohort of replacement-level pitchers, peruse the 2004 stats of guys like Brett Myers, Cory Lidle, or Jose Contreras (his NY numbers).  Below pedestrian. 

But what does a GOOD pitcher look like?  Last year’s leaders were Johan Santana, Curt Schilling, and Randy Johnson, with 89, 73, and 69 runs saved, respectively.  The leaders in 2003 were Tim Hudson, Esteban Loaiza, and Jason Schmidt (79, 78, and 75).  And 2002 was led by Derek Blowe, Randy again, and Barry Zito (81, 81, and 73).  To convert these totals into approximate runs, just divide by 10:  bringing Loaiza to spring training in 2003 was equivalent to the Sox 'buying' 8 wins.   It was a great pickup.


With that context, let’s look at the staff for the 1993 Sox, who won the AL West by 8 games:

Alex Fernandez            65.7
Jack McDowell            62.8
Wilson Alvarez             56.9
Jason Bere                   32.7
Kirk McCaskill             2.7
TOTAL VORP:           220


The top of the rotation doesn’t show any Unitesque values, but when you consider that Fernandez, McDowell and Alvarez ranked in the Top 8 of the AL, it’s clear that these guys were studs.  Again, for perspective, 220 VORP was the total of Maddux, Smoltz, Glavine, and Avery for the Braves in 1996.

What about the ’83 club?

Rich Dotson                 55.4
Floyd Bannister            45.4
LaMarr Hoyt                45.1
Britt Burns                    27.6
Jerry Koosman             7.9
TOTAL VORP: 181

Not as impressive, but enough to give the Sox a Division Title by 20 games over Kansas City.

And finally, let’s look at the 2005 club.  I chose VORP for this exercise because it summarizes pitcher performance with one simple number.  But more importantly, it can be used as the dependent variable in regression models predicting future performance.  The projections for the 2005 staff, using Baseball Prospectus’ PECOTA model:

Mark Buehrle            35.1
Salsa Garcia              30.5
Duque Hernandez      24.1
Jon Garland               19.7
Jose Contreras           17.1
PROJ. VORP            126.5


No Hoyts, no Dotsons, no Fernandez or Alvarez.  Barely a Jason Bere among the group.

Those guys aren’t 1983.  And they’ll never be 1993.

There’s much more to be said about this exercise; I’ll write more about VORP, regression, projections, and the strengths and limitations thereof in the next few weeks.


Home