Sabermetricians spend countless hours projecting the performance of individual players and by and large do a remarkable job. Yet every year, there are flukes—or perhaps it is more correct to simply say unexpected performances—that nobody sees coming. Who knew that Magglio Ordonez would have a legitimate case for AL MVP at the break? Or that Dmitri Young would not only hit acceptably but end up on the NL all-star team? This is not an indictment of projection systems; it is simply an observation that every year baseball brings us surprises. It’s part of why baseball is the greatest game in the world.
Now that we are halfway through the season, it is a good time to take a look at some performances that are out of line with what we might have guessed last winter. To do this, we’ll compare actual numbers with projected numbers from The Hardball Times 2007 Season Preview.
Batting average (>250 PA) Player Actual Projected Diff Magglio Ordonez .370 .290 +.080 Dmitri Young .339 .261 +.078 Reggie Willits .324 .250 +.074 BJ Upton .320 .248 +.072 Kevin Youkilis .329 .262 +.067 Brendan Harris .305 .239 +.066 Orlando Cabrera .329 .268 +.061 Hanley Ramirez .319 .259 +.060 Corey Hart .311 .252 +.059 Jorge Posada .325 .267 +.057 ... Carlos Quentin .210 .265 -.055 Adam Kennedy .213 .269 -.056 Jason Kendall .224 .283 -.059 Jermaine Dye .220 .279 -.059 Morgan Ensberg .207 .270 -.063 Nick Punto .205 .270 -.065 Jay Gibbons .207 .273 -.066 Andruw Jones .199 .273 -.074 Julio Lugo .192 .278 -.086 Dioner Navarro .181 .288 -.107
Ordonez and Young are both obliterating their projected batting averages. In Ordonez’s case, his batting average on balls in play is at an insane .390, out of line with both his established career norm (as astutely observed by Dayn Perry) and what would be expected from his batted ball distribution (the distribution of batted balls among line drives, ground balls, pop ups and fly outs). He is not the only one.
Almost all of the names at the top and bottom of the lists show a batting average on balls in play either out of line with established levels performance or batted ball distribution (or both). When this occurs, there is a strong possibility of a correction. Only Hanley Ramirez has a strong claim to his lofty batting average, although one might also make rather weak arguments for Brendan Harris and Corey Hart. Traditionalists hate hearing it, but a large component of batting average is chance.
What is alarming, however, is that many of the hitters at the bottom of the list have very poor balls in play distributions—in particular, they are not hitting line drives. Hitting line drives is very important for hitters; liners are more likely to become hits than other types of balls in play. That is troubling for a young hitter like Dioner Navarro, puzzling for an established hitter like Julio Lugo, and downright sad for a declining veteran like Jason Kendall.
Isolated slugging (>250 PA) Player Actual Projected Diff Carlos Pena .320 .164 +.156 Jack Cust .324 .183 +.141 John Buck .288 .163 +.125 Prince Fielder .326 .217 +.109 BJ Upton .225 .124 +.101 Carlos Guillen .251 .153 +.098 Matt Stairs .259 .165 +.094 Alex Rodriguez .340 .246 +.094 Curtis Granderson .265 .173 +.092 Victor Martinez .225 .142 +.083 Magglio Ordonez .243 .161 +.082 ... David Ortiz .233 .318 -.085 Jacque Jones .100 .189 -.089 Jim Edmonds .155 .245 -.090 Manny Ramirez .184 .276 -.092 Travis Hafner .191 .287 -.096 Derrek Lee .154 .251 -.097 Scott Rolen .117 .217 -.100 Albert Pujols .215 .321 -.106 Frank Thomas .195 .308 -.113 Nomar Garciaparra .060 .177 -.117 Note: Isolated power = SLG - AVG = (TB-H)/AB
Change in ability level, or fluke? When expectations don’t match reality, one needs to ask whether there has been a concomitant change in ability. If you buy into the theory that power is often the last tool to develop, then this question is particularly apt when discussing unexpected power spikes of young hitters. Guys like Carlos Pena, Jack Cust, and Victor Martinez may just be hitting mid-career power spikes—all were highly touted for their power potential as minor leaguers and all are in their late twenties. But younger guys like Prince Fielder and B.J. Upton and even Curtis Granderson may be showing development as power hitters. For example, Hit Tracker categorizes home runs as “no doubt,” “plenty,” or “just enough” based on the distance the ball travels. There is nothing strange about Fielder’s distribution of home run distances; if anything, a greater fraction of his home runs have had plenty of distance this year than last year. There is nothing fluky about this year’s power surge, and the uptick in power may signal the arrival of an elite power hitter.
On the flip side, there are some established sluggers who have seen a marked decrease in their power output. Yes, offense is down. ISO is down 11 points in the AL, for example. But the gentlemen populating the bottom of the list have underperformed their projections by much more than that. As before, we must ask whether there has been a change in ability, whether due to age-related decline (Frank Thomas?) or injury (Scott Rolen?). It may even be an indication of a player hiding an injury (Travis Hafner?). Or could just be a plain old fluke. In David Ortiz‘s case, there has been speculation that he is consciously trying to beat the defensive overshift usually employed against him by hitting to the opposite field. If that is true, then the defensive overshift is working; opposing teams would gladly trade a handful of singles for a few bombs.
Remember when Nomar Garciaparra and Alex Rodriguez were part of the shortstop trinity in the early aughts? While both have moved off of shortstop, Rodriguez is on pace to threaten the AL single-season home run mark while Garciaparra has hit two home runs all year.
Walks per plate appearance (>250 PA) Player Actual Projected Diff Jose Reyes 11.8% 5.6% +6.1% Pat Burrell 20.1% 14.1% +6.1% Jim Thome 22.5% 16.9% +5.6% Ryan Howard 17.6% 12.5% +5.1% Magglio Ordonez 12.5% 7.5% +5.0% Nick Punto 12.8% 7.9% +4.9% Grady Sizemore 13.3% 8.6% +4.7% Chad Tracy 12.0% 7.6% +4.4% Orlando Hudson 12.3% 8.0% +4.3% Kevin Millar 14.4% 10.3% +4.1% ... Bobby Crosby 5.6% 9.0% -3.4% Dave Ross 6.3% 10.0% -3.7% Ivan Rodriguez 1.7% 5.5% -3.8% Jason Kendall 4.1% 8.0% -3.9% Aubrey Huff 4.7% 8.7% -4.0% Chris B Young 5.0% 8.9% -4.0% Adam Dunn 12.2% 16.3% -4.1% Bobby Abreu 12.1% 16.6% -4.5% Jim Edmonds 9.3% 14.4% -5.2% Carlos Delgado 7.4% 13.0% -5.6%
What does it mean when a hitter starts walking up a storm? For young hitters, it is very promising. Jose Reyes and Grady Sizemore are already superstars at a young age, and they are taking walks like never before. The increase in plate selectivity is a very good sign, both for their OBP and their ability to wait for a good pitch and smack it. (The numbers here include intentional walks, so we haven’t properly projected how scared managers would be of Jose Reyes and Ryan Howard, whose intentional walks account for 21% and 42% of their free passes, respectively.)
However, for older hitters, a large increase in walk rate can be an ominous sign. Morgan Ensberg, for example, walked like crazy last year as his batting average plummetted. The speculation was that, for whatever reason, he no longer trusted his swing to generate hits and preferred to trot to first base. That’s a good strategy but one with a limited half-life—when pitchers catch on, they will force the hitter to put the ball in play by pounding the strike zone more consistently. And then what happens? Morgan Ensberg 2007 happens: his walk rate has dropped from 21% to 12% and his overall batting line has suffered. So what of Pat Burrell and Jim Thome? Both are seeing marked decreases in their isolated power and may be good bets to decline next year.
This phenomenon might be at play for Bobby Abreu, who saw an uptick in his walk rate last year and has struggled mightily this year. In general, hitters who are not as scary was they once were are likely to draw fewer walks. Aubrey Huff, Jim Edmonds, and Carlos Delgado aren’t the hitters they once were. Pitchers are going to challenge them. And nobody is afraid to throw Jason Kendall strikes; what is he going to do, bloop one to the right side?
Strikeouts per plate appearance (>250 PA) Player Actual Projected Diff BJ Upton 30.1% 22.5% +7.6% Carl Crawford 19.3% 13.2% +6.1% Jay Gibbons 19.4% 13.4% +6.1% Juan Uribe 22.6% 16.5% +6.0% Craig Monroe 25.3% 19.5% +5.9% Jack Cust 32.9% 27.1% +5.8% Craig Biggio 19.8% 14.3% +5.5% Dave Ross 30.3% 25.2% +5.1% Andruw Jones 24.1% 19.2% +4.9% Adrian Gonzalez 21.9% 17.6% +4.3% ... Corey Patterson 14.5% 20.8% -6.2% Chris B Young 17.5% 23.7% -6.2% Kevin Youkilis 11.7% 18.0% -6.4% Dmitri Young 14.0% 20.6% -6.6% Pedro Feliz 11.1% 17.7% -6.6% Adam Kennedy 8.2% 15.1% -6.9% Austin Kearns 16.1% 23.1% -7.0% Mike Napoli 23.3% 30.6% -7.3% Andre Ethier 10.6% 18.8% -8.2% Jose Bautista 15.4% 23.7% -8.3%
I am not sure how much to comment on the leaders and trailers for strikeouts. Strikeouts aren’t necessarily a bad thing, although a chronic inability to make contact can torpedo a young hitter’s career. Perhaps B.J. Upton is having trouble adjusting to breaking balls (there’s a task for MLB Enhanced Gameday!). An older player may be seeing a decrease in bat speed. A mid-career player like Andruw Jones…who knows?
Questions about contact ability have followed Corey Patterson and Chris Young. Now that where they are making more contact than expected, both are having pretty poor seasons. On the other hand, Mike Napoli is also striking out less than expected and he is having a very good year. Is it in spite of or because of the additional contact? Which way does the causality arrow point? Strikeouts are a funny thing, and it is difficult to say what it means for these hitters.
Before we leave the subject, it is worth mentioning Jack Cust. Cust is a cartoon character; rarely does a hitter take the “take-and-rake” approach to such an extreme. It could be the bottom of the ninth in game seven of the World Series with the winning run on third base and one out, and Cust would gladly take strike three on the outside corner rather than attempt to swing. Hey, whatever works for him: he also appears on the list of hitters overperforming their projected isolated slugging.
Regression to the mean is a powerful thing. Odds are, dollars to doughnuts, that these lists will be populated by different hitters if we prepare them using second half statistics of this year. Players who perform toward the extreme tend toward the center over time. Keep an eye on the guys on these lists: if they continue to over—or underperform in these component stats, it may be time to start thinking that they are showing legitimate changes in abilities.