Adjusted VORP: A First Experiment
Keith Woolner
One of the objections I've gotten to the method I use to compute VORP
in my weekly posting is that I use only one league's worth of players
to set the positional average. If there is an imbalance between
the leagues at a position (e.g. the AL has all the good shortstops
in the majors), that will make one set of players look exceptionally
strong, and the other exceptionally weak. If the goal is to model
the difficulty or scarcity of a position by the differences in offensive
production, then we should use all available positional data to make
that adjustment.
I have expressed concerns in the past that simply combining the joint
production of both leagues (i.e. looking at both AL and NL shortstops
together) would produce a problem when the league's offensive levels
are different. An adjustment
Well, the comments I got prompted me to look at the '98 data, and I
found that the league offensive levels aren't all that different this year,
especially after accounting for the DH:
LG AVG OBP SLG OPS
-- ----- ------ ------ -----
A .271 .334 .429 .763 (DH's removed)
N .269 .336 .425 .761 (pitchers removed)
With that objection removed for this season (I still have concerns
about simply combining the leagues in historical VORP comparisons),
I decided to try computing VORP using a positional average
computed from both leagues worth of data. The positional averages
are as follows: (using the same caveats that apply in the VORP
reports themselves)
POS AVG OBP SLG
--- ---- ---- ----
1b .284 .361 .485
2b .270 .333 .392
3b .269 .336 .430
c .257 .315 .399
cf .273 .338 .427
dh .275 .355 .464
lf .271 .341 .444
p .147 .185 .184
rf .285 .350 .477
ss .266 .321 .386
Below, I list the top offensive players 20 in each league. Note that
a component of VORP is still the overall league offensive level (not
just the positional offensive level), and I am still using separate
figures for the AL (including DH) and NL (including pitchers), so an
NL player will still come in a bit lower simply because NL teams
get less run production out of the rest of their lineup, and VORP
compares to a league average team.
NAME TM LG POS PA AVG/ OBP/ SLG SB CS MLVr PMLVr VORPr VORP
ABelle CHS A lf 674 .331/.408/.656 6 3 .513 .490 .621 99.8
ARodriguez SEA A ss 731 .310/.353/.560 46 13 .241 .332 .471 85.2
NGarciaparra BOS A ss 637 .323/.358/.584 12 6 .271 .362 .493 74.5
KGriffeyJr SEA A cf 709 .284/.361/.610 20 5 .290 .293 .428 73.4
BWilliams NYY A cf 565 .341/.425/.580 13 9 .405 .408 .536 68.9
DJeter NYY A ss 675 .324/.381/.481 27 6 .189 .280 .417 68.6
EMartinez SEA A dh 662 .322/.431/.567 1 1 .374 .299 .429 66.3
JOfferman KSC A 2b 683 .311/.400/.435 43 12 .148 .207 .346 61.2
MVaughn BOS A 1b 670 .337/.397/.591 0 0 .351 .231 .362 57.6
RDurham CHS A 2b 692 .285/.360/.452 35 9 .116 .176 .313 55.6
MRamirez CLE A rf 647 .294/.377/.599 5 3 .319 .226 .357 53.8
JuGonzalez TEX A rf 644 .318/.366/.630 2 1 .314 .222 .354 53.1
IRodriguez TEX A c 601 .320/.353/.510 7 0 .137 .228 .361 52.2
JEdmonds ANA A cf 653 .307/.368/.507 7 5 .180 .183 .312 47.6
RPalmeiro BAL A 1b 677 .297/.377/.570 11 6 .283 .163 .294 46.7
CDelgado TOR A 1b 603 .292/.378/.592 3 0 .297 .177 .310 45.3
JThome CLE A 1b 529 .293/.412/.584 1 0 .348 .228 .360 44.7
TSalmon ANA A dh 549 .303/.417/.538 0 1 .292 .216 .346 44.5
TWalker MIN A 2b 566 .315/.371/.473 19 7 .130 .189 .322 44.4
EDavis BAL A rf 488 .328/.387/.587 7 6 .356 .263 .392 44.1
TFernandez TOR A 2b 521 .324/.380/.464 12 8 .163 .222 .350 42.6
NAME TM LG POS PA AVG/ OBP/ SLG SB CS MLVr PMLVr VORPr VORP
MMcGwire STL N 1b 667 .297/.465/.746 1 0 .633 .472 .596 93.2
BBonds SFG N lf 677 .304/.436/.612 27 12 .457 .387 .511 81.2
CBiggio HOU N 2b 683 .324/.385/.499 47 8 .318 .326 .463 80.7
MAlou HOU N lf 644 .314/.398/.587 11 3 .440 .370 .496 75.4
MPiazza NYM N c 441 .348/.417/.607 1 0 .500 .538 .661 70.2
ChiJones ATL N 3b 679 .311/.406/.547 15 6 .339 .296 .420 69.7
BLarkin CIN N ss 608 .311/.400/.509 25 3 .271 .310 .442 68.5
JOlerud NYM N 1b 644 .353/.447/.553 2 2 .472 .311 .433 66.5
SSosa CHC N rf 706 .310/.381/.654 17 9 .404 .267 .392 64.7
GVaughn SDP N lf 641 .272/.360/.602 11 4 .334 .264 .389 60.1
JBagwell HOU N 1b 622 .306/.423/.559 18 5 .427 .266 .393 58.7
RLankford STL N cf 618 .291/.390/.538 26 4 .289 .244 .376 58.5
JKendall PIT N c 583 .327/.386/.474 26 5 .193 .230 .361 56.1
JKent SFG N 2b 569 .296/.355/.557 9 4 .274 .283 .406 54.0
SRolen PHI N 3b 666 .286/.381/.527 13 6 .234 .191 .314 50.1
VGuerrero MON N rf 660 .323/.365/.580 11 9 .312 .176 .297 47.3
JLopez ATL N c 505 .284/.325/.538 5 3 .188 .225 .348 42.2
AJones ATL N cf 609 .278/.325/.527 27 4 .169 .123 .256 42.1
AGalarraga ATL N 1b 601 .302/.373/.587 7 6 .335 .174 .295 41.8
GSheffield LAD N rf 370 .316/.443/.535 18 6 .443 .307 .437 41.2
LWalker COL N rf 511 .365/.442/.633 14 4 .334 .198 .326 41.1
Pitcher VORP is not affected by this change (although you could compute
an offensive VORP for pitcher batting and combine that with his pitching
VORP in much the same way that you'd add Fielding Runs to VORP to get
a more complete measure of a player's value).
Comments on the adjustment are welcomed. I'm still working on
VORP for players who appeared with multiple teams, but I wanted
to crank out some results while the IBA balloting is still going on.
--
Keith Woolner kwoo...@best.com http://www.best.com/~kwoolner
Baseball Engineering Library, http://www.stathead.com
List Owner, Red Sox mailing list, http://www.best.com/~kwoolner/redsox/list
Newsgroup Moderator, rec.sport.baseball.analysis
Check out the latest VORP ratings, updated weely at http://www.stathead.com!