Revisiting ’03: Defense wins championships?

facebooktwitterreddit

Welcome back to Revisiting ’03, a series in which we look at the 2003 World Series champion Florida Marlins through my new sabermetric lenses. Today, I wanted to talk about the often ballyhooed cliche of “defense wins championships.”

As of late, defense has gotten more and more important in the eyes of executives who run this game. But back in 2003, defense was perhaps still unappreciated, or at least appreciated in only a qualitative sense. While there were metrics like UZR and Probabilistic Model of Range (PMR) available back in those days, I think few teams considered numbers on defense as important. The all-knowing “scouting eye” was the sole determinant to how good a defender performed.

Now, this is not to say that scouting is bad. ESPN’s Jerry Crasnick had a great piece yesterday on how scouting and metrics should combine to give a proper evaluation of a player’s defense. I’m all for it. But like I said, to ignore one or the other would be wrong, in my opinion. With that in mind, we’re going to look at the perception as viewed by the eyes and the reality as defined by the stats, knowing full well that both are not totally right.

The Perception

The perception of the 2003 Marlins were that they were a pretty good defensive team. Reputations abound among the everyday players on that club. It started, of course, at catcher, with the ever talented Ivan Rodriguez. Pudge was multi-time Gold Glove winner and almost a sure-fire bet to win every season wherever he played, though he did not earn a GG in either 2002 or 2003. Moving down the position spectrum by number, you get quite a few other strong reputations. Derrek Lee was well known as one of the best scoopers at first base in the big leagues. Luis Castillo and Alex Gonzalez were both widely lauded in South Florida, though neither had been awarded at that point. Mike Lowell came up in the minors with a good reputation as well.

As strong as the infield supposedly was, it was balanced slightly by a weaker outfield. Juan Pierre was considered rangy thanks to his speed, but none of the remaining outfielders among Juan Encarnacion, Todd Hollandsworth, Miguel Cabrera, and Jeff Conine were considered good gloves in the corners.

The (Statistical) Reality

With defensive metrics, it’s always a little difficult to tell how good individual players are, but at the team level a lot of the research has shown that defense is pretty decently quantified. The Hardball Times, for example, has quantified a Plus/Minus type system for team fielding that goes above and beyond what you would get from simple Defensive Efficiency Ratio (DER). While DER measures simply a rate of outs made per ball in play (essentially reverse BABIP), THT’s metrics account for difficulty of the balls in play, adjusting the ratios accordingly.

But we won’t get into that statistic (maybe for another day). For this look, we’re going to show two different numbers: team bUZR as measured by FanGraphs, and team TotalZone as measured by Sean Smith and shown at Baseball-Reference. I won’t be weighing anything, I’m just going to show the numbers involved, but keep in mind that, of the two totals, I would trust bUZR more than TotalZone just due to the greater detail in the UZR dataset (taken from Baseball Info Solutions) as opposed to that of TotalZone’s dataset (taken from Retrosheet).

The following are all Marlins players that recorded at least 50 defensive games (as measured by UZR). The bottom is the team total, including all players that played less than the number of games mentioned. For Pudge, I used his TotalZone numbers to approximate catcher defense.

PlayerPosDefensive GamesUZRTotalZone
Ivan RodriguezC126-2.0-2.0
Derrek Lee1B1510.41.0
Luis Castillo2B1318.314.7
Alex GonzalezSS1256.00.9
Mike Lowell3B1500.14.2
Juan PierreCF1508.1-11.8
Juan EncarnacionRF161-12.0-13.5
Miguel Cabrera3B/LF86-0.9-3.1
Todd HollandsworthLF/RF60-9.2-2.9
Total-9.3-15.1

You can see from the table that the Marlins came out as a little below average according to the metrics. If the team really came out somewhere between nine and 15 runs below average, that works out to around 0.006 and 0.01 runs below average per inning, or about one run every 11 to 18 eighteen games. Not a lot at all, but it sure makes you wonder about the focus on defense that has been the goal of the franchise the last year or so, and what they actually believe defense is.

Looking at the individual numbers, you can see one major discrepancy in the evaluation of Juan Pierre. TotalZone really dislikes him, rating him at -12 runs for the season, while UZR has him as well above average at +8 runs. The difference is entirely in their range measurements, as Pierre recorded almost 14 runs above average with his range via UZR, while the range aspect of TotalZone rated him as around -8 runs. Both metrics agree that Pierre’s arm is awful, ranging between four and six runs below average. If I took a weighted average between these two evaluations (weighing UZR more because of the better dataset), you get something like three runs above average, which definitely would not surprise me.

The big winner on defense for the Marlins was Luis Castillo, which comes as no surprise to fans of the Fish. It’s interesting to note, however, that up until that season Castillo had not rated well with either metric. We only have UZR data dating back to 2002, but in the 2002 season Castillo posted a nondescript three runs below average. TotalZone had Castillo as a below average second baseman up until that monster 2003 season. Since 2003, however, both UZR and TotalZone have Castillo as being excellent, which was indeed followed quickly by some Gold Glove awards. In fact, from 2003-2005, Castillo posted three well above average seasons with the glove, and in each year he earned a Gold Glove award playing for the Marlins. Sometimes the voters do get it right.

The biggest loser on the Marlins defensively was Juan Encarnacion, and this should not surprise anyone either. What surprised me the most about Encarnacion wasn’t the range data, as even with good tools like Encarnacion’s, players can and often do lack range because of bad fundamentals. Rather, I was quite surprised that his arm rated so poorly. The Fans Scouting Reports for Encarnacion in 2004 and further on rate his arm quite well, so it was surprising to see him put up somewhere between four and five runs below average in that department.

I think the biggest surprise for me might have been Derrek Lee’s defensive numbers. At the time, I thought defense at first base was pretty important, but I’ve come to realize that, due to the nature of balls hit to first, there is not a whole lot of difference between the best and worst players at first. Lee always impressed with his stretches and scoops, and that data is not involved in UZR or TotalZone, but looking at MGL’s use of WOWY to determine scoop value, we did not seem to be missing out on much. Lee did not rank highly in this, and I’m inclined to say that this is one area where the numbers can definitely help clear up how much value an action on defense has.

Hopefully, you now have some better context on how well the Marlins played defense in 2003. The truth lies somewhere between what our eyes said (good) and what the numbers said (below average, but not by a whole lot). I think I can live with that.