April 27, 2011
Prospectus Hit and Run
Return of the Up-the-Middlemen
Earlier this month, I examined the timeworn adage that a ballclub must be strong up the middle—at catcher, second base, shortstop, and center field—to win. What I found was that while the aforementioned positions are the most demanding ones defensively—they're the four at the right end of Bill James' defensive spectrum, which runs 1B-LF-RF-3B-CF-2B-SS-C—the collective defensive abilities of those players had a much lower correlation with team winning percentage than their collective offensive abilities. In other words, having good hitters up the middle is far more vital to the winning effort than having good fielders.
On some level, this shouldn't have been all that surprising. The spread between good and bad hitters is far wider than that between good and bad fielders; teams simply don't suffer inept fielding at key positions for very long, but they'll let the Neifi Perezes and Cesar Izturises bleed their offenses to death year after year. Furthermore, as the spectrum tells us, it's much tougher to find above-average hitters to fill the premium defensive slots, and teams are at a considerable advantage when they do.
As one reader pointed out, missing from that examination was a basis for comparison. Which of the corner positions—(first base, third base, left field, and right field), or infielders and outfielders—are the collective performances of those groupings better indicators of team success? I sent our data elves back into the mineshaft to find out.
To examine the importance of this principle, we totaled the offensive and defensive contributions of each team's starters and backups at the various position clusters using Batting Runs Above Average and Fielding Runs Above Average, doing so for all teams from 1995 to the present. We then measured how well each component correlates with winning percentage.
There's a fair bit to digest in that little grid. As reported before, the level of offense provided by the middlemen is a much better indicator of team success than their level of defense. Furthermore, the middlemen’s offensive contribution is a slightly better indicator of team success than the offense provided by the cornermen, and once you factor in defense—where there's absolutely no correlation between the work done by the latter group and winning—it's the middlemen in a landslide.
Considering just the infielders (not including catchers) and the outfielders, it's the former with a consistent edge on the offensive, defensive, and combined fronts, though the correlation of their contributions with winning isn't as strong as it is with the middlemen. Combining the contributions of complementary clusters (infield and outfield, or corner and middle) produces much higher correlations with winning percentage, though again, defense tends to come out in the wash.
Leather's apparent lack of importance probably has something to do with the way we're measuring it here. BRAA is relative to all hitters, while FRAA is relative to the players at individual positions whose spread of performance may vary. To a lesser extent, these findings may have something to do with the way FRAA—new FRAA, with its margin for error regressed for the purposes of WARP—is able to measure a single season's worth of defensive data, even among a group of players. It's worth noting that for the same period, both raw and Park-Adjusted Defensive Efficiency have a far more substantial .41 correlation with winning percentage.
Moving from the generalities to the specifics, I previously found that 18 of the top 25 wild card-era teams in up-the-middle RAA have made the postseason, while just three of the top 25 finished below .500. (I wrote two for the latter instance, but the table clearly indicated three; score that E-6). Underscoring the importance of offense, 14 of the top 25 teams actually had negative FRAA contributions from their middlemen. Meanwhile, none of the bottom 25 made the postseason (none of the bottom 59, actually), and just one (the 2008 Astros, starring Brad Ausmus, Kaz Matsui, Miguel Tejada, and Michael Bourn) finished above .500. By comparison, 14 of the top 25 teams in corner RAA made the postseason, with four of the top 25 finishing below .500; at the other end of the spectrum, one of the bottom 25 (the 1996 Dodgers, starring Eric Karros, Mike Blowers, Todd Hollandsworth, and Raul Mondesi) made the postseason, and just four finished above .500.
Meanwhile, 16 of the top 25 teams in infield RAA made the postseason, again with just four sub-.500 entrants; two of the bottom 25 (the aforementioned Dodgers, with the immortal Delino DeShields/Greg Gagne double-play combo, and the 2000 Yankees, burdened by the fielding woes of Chuck Knoblauch and the hitting woes of his replacements) made the postseason, with only three rising above .500. For outfield RAA, 15 of the top 25 made the postseason, but oddly enough, none have finished below .500; meanwhile three of the bottom 25 made the postseason, with six finishing above .500 and one more right at that mark. In each instance, the 25-team sample represents just a hair above five percent of the total group of 474 teams; these are the cream of the crop, and the crust of the crud.
Four tables of 25 teams apiece is more data than I care to dump into a single article with any meaningful discussion, so I'll limit each group to the top 10.
There's a considerable amount of repetition within the groups, not to mention overlap between them, due to cores of star players remaining together for multiple seasons. Foremost are the Yankees, who as previously noted claim four of the top 10 spots among the up-the-middle group via Jorge Posada, Derek Jeter, Bernie Williams, and a revolving cast of second basemen; the 1998 Yankees outfield of Williams, Paul O'Neill, and the Chad Curtis/Tim Raines/Shane Spencer hydra crack that group's rankings as well. The Alex Rodriguez/Ken Griffey Jr. Mariners make the middle list twice, with Griffey, Jay Buhner, Rich Amaral, and Mark Whitten pushing the 1996 team onto the outfield list as well. The Albert Pujols/Scott Rolen Cardinals make the corner and infield lists twice apiece, with two more Pujols-led entries gracing each list; with solid seasons from Tony Womack and Edgar Renteria, the 2004 team's infield tops that field by a whopping 25 runs, the widest margin of any of these lists. Back-to-back Jason Giambi/Miguel Tejada/Eric Chavez A's teams make the infield list. Two Angels teams featuring Garrett Anderson and Tim Salmon top the outfield list, the first with Anderson in left and Jim Edmonds in center, the second with Darin Erstad in left during his 240-hit season and Anderson in center. The latter team also tops the corner list thanks to monster seasons from Mo Vaughn and Troy Glaus, but even with all that offense and defense, they finished in third place at 82-80 thanks to a staff ERA of 5.00. Those were the days.
Sometimes the continued excellence of one star is enough to drive multiple appearances. No less than five consecutive Barry Bonds-era Giants teams make the outfield list, with the 2004 version, which got a great year from J.T. Snow, cracking the corner list as well. Chipper Jones-led Braves teams make the infield list (2008) and the outfield list (2002), the latter during his relatively brief foray into left field, where he accompanied Andruw Jones and Gary Sheffield in the pasture. Andruw helps the Braves top the up-the-middle rankings, with a cast featuring Javy Lopez, Marcus Giles, and Rafael Furcal.
Among the strays of note, the 2007 edition of the Chase Utley/Jimmy Rollins Phillies appear on both the middle and infield lists, the former with Carlos Ruiz and Shane Victorino, the latter with Ryan Howard and the Wes Helms/Greg Dobbs/Abraham Nunez triumvirate at third base. The 1996 Orioles, featuring the double-play combo of Hall of Famers Roberto Alomar and Cal Ripken, make the middle list with Chris Hoiles and Brady Anderson, the latter during his 50-homer season. The 1996 Indians featured Julio Franco, Jim Thome, Albert Belle, and Manny Ramirez at the corners, with the latter two flanking Kenny Lofton to reach the outfield list as well. The 1999 Mets make the corner list via John Olerud, Robin Ventura, Rickey Henderson, Benny Agbayani, and Roger Cedeno. Rolen reappears on the corner list with last year's Reds, with Joey Votto, Jonny Gomes, and Jay Bruce.
In any event, this exercise reinforces the notion that offensive strength up the middle is a key building block for a winning team, more important than getting good defense from those positions, or good batsmanship from the corner positions That's not to say defense should be ignored, but its value is often overstated because the spread between good and bad is far less. The difference between the best up-the-middle clubs of the wild-card era and the worst is 504 runs (208 to -296) on the offensive side, but just 89 runs (44 to -45) on the defensive side; to a much less extreme degree, the same holds true across the other clusters. Aspiring general managers, when you build your teams, find the best hitters who can field the ball adequately, and go from there, as you'll have a leg up on those who pledge allegiance to the leather above all else.