Florida Marlins

On Prediction and Other Self-Defeating Propositions

Several people suggested that my predictions for Florida last year were either tongue-in-cheek or based on the simple device of taking the opposite viewpoint from the burgeoning neo-sabe groupthink. Neither suggestion was true, but sometimes it's best to let people live with their illusions.

The truth: despite being the only person outside of the Marlins organization to suggest that the team could be a contender in 2003, I did not do a very good job of prediction. I didn't foresee the loss of A.J. Burnett, and I didn't anticipate the meltdown of most of their bullpen early in the season. I did not suggest that Dontrelle Willis or Miguel Cabrera would be able to step in as solid major league performers, and I was unable to divine that a managerial change would have such a galvanizing effect on a young team that had been languishing in their development.

Despite getting so many details wrong, my prediction was correct---which is like having a projection system where the individual results are so all over the place that the aggregate error rate actually looks pretty good. It is possible in this world to take something too seriously while simultaneously not taking it seriously enough; the fact that such is not only possible, but likely, and that this is not paradoxical at all, but simply the nature of things, highlights the paradoxical circularity of "enlightenment."

And with those doleful thoughts out of the way, let's move on to...

What Happened Last Year In A Nutshell

The Marlins succeeded in 2003 because:

Four major mid-season additions contributed: Willis, Cabrera, Ugueth Urbina, Chad Fox.

Juan Pierre did not stink up the joint, as the neo-sabe cabal had decreed.

The Fish got solid offense and excellent defense up the middle (the left side of the defensive spectrum: C-CF-SS-2B).

The starting rotation, while not spectacular, was both consistent and deep, especially in the second half of the season. Mark Redman, a mostly unheralded acquisition, turned out to well-suited to Pro Player Stadium (2.88 ERA there as opposed to 4.27 on the road), and was especially effective in June and July, when the Marlins began to jell.

The bullpen was retooled over the course of two months from an eyesore into an asset.

The team offense finally became solid enough to play competitively on the road (more details below).

The Four-Year Home-Road Pythagorean Data

One of the main reasons I had continually downgraded the Marlins' chances to contend prior to 2003 was due to their sub-par road performance. As the home-road Pythagorean vs. actual data for 2000-03 shows, Florida developed a sizable home-road performance split in 2001, which actually got worse in 2002. Pythagorean differentials show that the team's 79 wins in 2000 was about five games too high, and that the expectations aroused by their jump were premature.

Yr      W   L  WPct   RS   RA   PWP   difW    rg   lrg   r/l
2000   36  44  .450  372  421  .444   .014  4.65  4.93  0.94
2001   30  52  .366  363  419  .435  -.158  4.43  4.70  0.94
2002   33  48  .407  330  424  .387   .051  4.07  4.37  0.93
2003   38  43  .469  372  399  .468   .002  4.59  4.51  1.02

(Acronyms: RS---Runs Scored; RA---Runs Allowed; PWP---Pythagorean WPct; difW---WPct minus PWP (expressed as a percentage); rg---runs/game; lrg---league runs/game; r/l---runs relative to league r/g, road teams)

The key in 2003 was that the Marlins were able to generate a road offense that was actually above league-average, something that they had managed to do only twice before (in 1995, and again in their other World Series year, 1997).

Yr      W   L  WPct   RS   RA   PWP  difW    rg   lrg   r/l    pf  r/l+  r/l+ ha
2000   43  38  .531  359  376  .479  .109  4.43  5.07  0.87  0.93  0.94     0.94
2001   46  34  .575  379  325  .570  .009  4.74  4.70  1.01  0.90  1.12     1.03
2002   46  35  .568  369  339  .539  .054  4.56  4.53  1.01  0.94  1.07     1.00
2003   53  28  .654  379  293  .616  .062  4.68  4.72  0.99  0.87  1.14     1.08

(Acronyms: same as above plus pf---park factor; r/l+---runs relative to league adjusted for park, home games; r/l+ ha: overall runs relative to league, home and away)

At least as important, however, was the additional improvement in run prevention at home, which was about twice as much (-46) as on the road (-25). In the context of their ballpark, the Marlins had a solid little offense; that, combined with the improvement from their pitching, made them a very difficult team to beat in Pro Player Stadium. Their overall adjusted offense (r/l+ h/a) grades out as close to the Blue Jays, who scored a lot more runs, but did so in a park that significantly favored hitters.

At Last, 2004

So what about this year? Some claimed that the Marlins had a mini-fire sale after their World Series upset over the Evil Empire. Five key players departed: Redman, Urbina, Derrek Lee, Ivan Rodriguez, and Juan Encarnacion. However, the Marlins either directly addressed these losses or had existing depth at the affected position; how well they have replaced these players will remain to be seen, but their situation in 2004 bears little or no resemblance to what things were like in 1998.

Redman will eventually be replaced by Burnett, who has recovered more swiftly that projected from elbow surgery and is now expected back in the rotation by May or June. In the meantime, the Marlins will go with retread lefty Darren Oliver, a long-time fave of yours truly, especially back in BBBA-the-book days. Oliver has made me look as foolish as anyone over the years, so I know better than to predict anything astonishing from him; however, he has always pitched markedly better in the NL and Pro Player will easily be the most favorable home park he's ever had.

The eventual rotation---Burnett, Josh Beckett, Brad Penny, Willis, and Carl Pavano---should remain at least as effective as last year's starting corps, assuming that they can avoid injuries.

Urbina (and the other departed closer, Braden Looper) have been replaced by Armando Benitez, who has been susceptible to the gopher ball during most of his career. His new home park might help him some, but his inconsistent control and tendency to start slow are potential drawbacks. It should be noted that Urbina had similar issues(though lesser in scope) prior to landing in Miami, where he was sensational, so there's at least some chance that the Fish will find a way to maximize Benitez' value.

Lee will be replaced by a platoon of Hee Seop Choi (the player the Marlins received from the Cubs for Lee) and free agent Wil Cordero. Lee's performance on the road in '03 (.979 OPS) is going to be the hardest thing for the Marlins to replace, along with his big performance against lefties (1.062 OPS). Cordero has been pretty good in that role (.844 OPS vs. lefties since 2001), but that's not going to take up the slack. Choi has a better chance of matching or exceeding Lee's output against righties (.810 OPS), and this is especially crucial, as he is expected to get ~70% of the platoon's plate appearances.

Rodriguez spurned a strong offer to return to Miami and signed with the Tigers, which opens up the catching slot for Ramon Castro. who had a big season as a reserve in '03, going 9-for-26 as a pinch-hitter. It's a stretch to think that Castro can match Pudge's performance, but he could conceivably exceed his SLG (.474). If he does that, the Fish will be very satisfied.

Encarnacion, despite 94 RBI and a serviceable season in right field for Florida in 2003, was expendable due to the Marlins' re-signing of Mike Lowell and the emergence of Miguel Cabrera, who will be the everyday right fielder. He was especially effective in his rookie year against Florida's division rivals (5 HRs, 1.245 OPS vs. Atlanta, 1.051 OPS vs. the Phillies). The Marlins have to be hoping that such a performance will be duplicated in '04.

Cabrera mostly played left field last year, but moved to third in September while Lowell was out with a hand injury. Jeff Conine was acquired from Baltimore, and helped Florida down the stretch (and especially in the post-season). He'll be 38 in late June, though, and he could go in the tank at any time. The Marlins have finally gotten one-time neo-sabe heartthrob Abraham Nunez (three years older than first advertisted...) healthy, and the switch-hitter has had a very hot spring (.400+ BA, 9 HRs); they appear to have a reasonable alternative if Conine takes a dive.

The rest of the lineup is comprised of holdovers from the '03 squad: Luis Castillo at second, Alex Gonzalez at short, Lowell at third, and Juan Pierre in center. Few will dispute the established value of Castillo and Lowell, but there will be varying levels of suspicion directed at Pierre and Gonzalez. Here's the same OPS+ breakout employed in the Red Sox and Giants previews for the Marlins' lineup (including two key departees):

Player       Age  2003 OPS+  Peak OPS+  Career OPS+
(Rodriguez    32        124        152         114)
Castro        28        144        144           84
Redmond       33         65        111           91
(Lee          28        135        135         115)
Choi          25        101        101           91
Cordero       32         94        119           96
Castillo      28        109        110           92
Gonzalez      27        100        100           77
Lowell        30        132        132          111
Conine        37        113        132          111
Pierre        26         98         98           81
Cabrera       21        109        109          109

Castro, Gonzalez, Lowell and Pierre have the greatest distance between their '03 performance and their career levels; the overall sense one gets from this display is that the Marlins are going to have trouble scoring as many runs with this lineup as they did with last year's.

Additionally, there are few options up the middle should Castro, Castillo, Gonzalez and/or Pierre falter or get hurt. Mike Redmond can be a stopgap, but he's not going to bat higher than seventh. Neither Mike Mordecai nor Damion Easley are credible long-term replacements at short or second, and they're just barely viable as role players. Nunez looks to be a reasonable surrogate for Pierre, but he might end up as the everyday left fielder. Brian Banks, who was a solid bench player for the Marlins in '03, is off the roster for at least three months due to a spring training injury. Lenny Harris, the all-time pinch-hit leader, isn't slated to play much in the field. In short, the Fish aren't nearly as deep this year, and any injuries they sustain are likely to have a much greater impact than they did in '03.

Thus for the Fish to win 90+ games in '04, they're going to have to see their pitching step up another notch. Is that possible? Yes. Their projected Big Three---Burnett, Beckett and Penny---could still be the significant force that many had predicted (albeit prematurely). If they are humming by mid-season, there will be less pressure on Willis and Pavano, and this could turn into a 50-60 run improvement in runs allowed.

However, some of that improvement could be eroded by the bullpen. Behind Benitez and set-up man Fox are lefties Michael Tejera and Tommy Phelps and righty Blaine Neal. Tejera and Phelps are serviceable, but not likely to be much more than that; Neal lost a significant amount of velocity in '03 and wound up in the minors for most of the year. The Marlins are considering keeping rookie Justin Wayne as a reliever, where he's been impressive this spring, but that's due more to their misgivings about Neal and the current unavailability of Tim Spooneybarger, out for at least the first three months of '04 with an elbow injury.

That was precisely the problem that the Marlins faced when Jack McKeon took over as manager last year, however; he and GM Jerry Beinfest figured out how to stop the bleeding. The Marlins allowed 5.5 runs per game on the road in April-June 2003; from July-September they allowed only 4.3 runs per game. The Fish also tightened their run prevention at home, but also added offense as they stormed to a 29-10 home record over the last three months of the seaon.

As McKeon worked out the kinks in the Florida pen, the team went from a 13-14 mark in close games (those decided by two runs or less) through May 31 to a 29-14 in such contests the rest of the way (11-8 on the road, and 18-6 at home).

Such serendipity may be harder to achieve with less offensive depth this year, however. Here's the bottom line for the Marlins in their quest to avoid a return to bottom-feeding: to reel in a similar number of wins in '04, their pitching is going to have to take the Fish upstream.

2004 ZiPS Projections

Name               P    AVG   OBP   SPC    G   AB    R    H  2B  3B  HR  RBI   BB    K  SB  CS 
Lowell             3b  .279  .355  .488  145  549   76  153  37   0  26   95   60   82   3   2 
Cabrera            lf  .301  .364  .466  146  549   84  165  34   3  17   80   50  116   7   4 
Choi*              1b  .246  .367  .441  117  395   63   97  18   1  19   62   71  114   2   2 
Cordero            1b  .273  .351  .434  114  366   46  100  23   0  12   53   40   73   1   1 
Castillo#          2b  .308  .382  .379  150  607   91  187  19   6   4   57   67   71  30  19 
Banks#             lf  .254  .339  .443  114  350   52   89  26   2  12   50   42   80   3   4 
Castro             c   .252  .318  .466   88  266   43   67  18   0  13   45   24   55   1   1 
Conine             1b  .269  .326  .430  134  509   61  137  34   3  14   67   39   73   6   2 
Gonzalez           ss  .263  .320  .439  140  499   50  131  32   4  16   68   33   96   1   3 
Colangelo          lf  .260  .350  .375   96  296   38   77  18   2   4   31   38   69   3   2 
Pierre*            cf  .299  .354  .362  162  665   93  199  25   7   1   56   50   39  55  19 
Aguila             lf  .271  .334  .387  110  413   56  112  20   2   8   47   36   92   6   7 
Padgett*           lf  .249  .319  .404  122  446   56  111  25   1  14   59   41  120   2   3 
Redmond            c   .282  .338  .356   71  188   15   53  11   0   1   18   15   23   0   1 
Valdez             ss  .264  .337  .352  116  375   47   99  20   2   3   35   37   58  10   6 
Nunez#             lf  .232  .324  .386  101  370   50   86  16   1  13   48   47  110  14   9 
Wood               3b  .250  .303  .389  119  432   58  108  28   1  10   52   29   98   2   1 
Medrano            2b  .247  .351  .330  105  385   58   95  17   3   3   59   59   80  18  12 
Ambres             cf  .226  .327  .356  129  438   73   99  20   5   9   43   61   88   9   7 
Christenson        cf  .240  .317  .365  104  337   46   81  22   1   6   37   34   66   7   4 
Treanor            c   .229  .307  .359   87  284   33   65  14   1   7   32   29   47   5   0 
Easley             2b  .230  .305  .380  105  379   42   87  27   3   8   42   32   57   3   4 
Jorgensen          c   .223  .318  .337   60  193   23   43  10   0   4   21   25   50   1   0 
Wathan#            ss  .250  .301  .342  106  368   41   92  17   4   3   31   23   53  10   8 
Wilson             ss  .244  .288  .353  121  431   50  105  21   4   6   41   23   77   5   5 
Mordecai           ss  .230  .294  .333   84  165   16   38  12   1   1   15   14   33   2   2 
Hooper             2b  .247  .301  .290  124  465   60  115  13   2   1   34   32   64  14   9 
Harris*            3b  .214  .275  .279  101  154   13   33   5   1   1   11   12   18   2   1 
Williams           rf  .208  .255  .333  105  303   40   63  19   2   5   29   17   47   8   7 

Name                  ERA   W   L   G  GS     INN    H   ER  HR   BB    K 
Willis*              3.47  12   8  30  30   179.0  163   69  12   40  155 
Beckett              3.51   9   6  22  21   123.0  100   48  11   45  140 
Spooneybarger        3.69   2   2  56   0    61.0   53   25   3   26   55 
Olsen                3.74   6   5  20  20   118.0  115   49  10   28   89 
Burnett              3.92  11   9  27  26   172.0  148   75  12   77  170 
Penny                3.93  11  10  30  30   181.0  175   79  17   51  141 
Pavano               3.94  11   9  32  27   169.0  169   74  15   43  118 
Neal                 3.95   3   3  59   0    66.0   63   29   2   27   44 
Borland              4.00   4   4  55   0    72.0   62   32   4   35   67 
Benitez              4.06   4   3  68   0    71.0   57   32   9   34   75 
Fox                  4.06   3   3  49   0    51.0   38   23   4   32   65 
Bump                 4.10   7   6  32  16   112.0  116   51   7   33   61 
Phelps*              4.13   3   2  40   4    72.0   71   33   6   26   52 
Neu                  4.17   2   2  48   0    54.0   46   25   3   31   50 
Wayne                4.50   7   7  28  27   148.0  155   74  12   53   83 
Bowers*              4.62   3   4  38   5    76.0   67   39   8   42   76 
Tejera*              4.65   6   7  41  17   120.0  120   62  14   46   92 
Oliver*              4.71   9  12  29  28   153.0  164   80  14   60   92 
Fyhrie               4.77   6   8  28  19   132.0  133   70  12   60   88 
Gracesqui*           4.94   3   3  43   0    51.0   41   28   1   44   51 
Perisho*             4.97   4   5  50   3    67.0   73   37   7   25   40 
Small                5.09   6   8  27  16   106.0  120   60  16   30   56 
Flannery             5.34   3   6  65   0    64.0   72   38   1   25   41 
James                5.38   3   6  32  18   117.0  139   70  18   34   58 
Kent*                5.43   2   3  36   1    58.0   60   35   4   36   32 
Manning              5.93   3   7  21  16    91.0   95   60   9   64   55 
Florie               5.96   3   6  19  13    71.0   72   47  10   47   55 

ZiPS Projections are not playing time predictors; they project a player's performance given playing time in the majors in their accustomed role. As with all projections, specific knowledge of non-statistical attributes of the player should temper what the computer says. ZiPS Projections are computer projections and not generated by the author.