BP Comment Quick Links


March 3, 2010 Checking the NumbersPower SappedAn anomaly is defined as a strange discrepancy or deviation from an established trend or baseline. In baseball terms, they occur when teams or players defy expectations—in either direction—to the extent that it proves difficult to offer confident explanations. Take, for instance, the 57 home runs on Luis Gonzalez’s resume under the 2001 heading: Gonzo was a solid hitter, but 57 dingers? Tallies like that are generally reserved for the cloutiest of cloutsmen, and not a player perennially ranging from 1725 homers. On the opposite end of this spectrum stands David Wright, the Mets' third baseman with all the makings of a Hall of Fame career, but who experienced an anomaly for the ages last season in the power department. While many have worked diligently to unearth the cause of his decline in power output, I am more interested in the historical ramifications—has a player with as established of a power baseline ever experienced a similar dropoff? And, if so, is there enough of a precedent to help fuel or modify expectations moving forward? Wright hit 10 home runs last year in what was, for all intents and purposes, a full season of playing time. That's right, 10! After averaging 29 dingers per season from 200508, Wright’s bopping of just 10 long balls would be akin to Ichiro batting .264, Roy Halladay posting 4.47 ERA, or the Royals acquiring a batter whose OBP exceeds .319. It just isn’t bloody likely, and it can drive fans batty searching for causes. Was it the new Citi Field and its dimensions (not likely, as park effects would account for a dropoff of, at most, three or four homers)? A change in swing mechanics or in hitting philosophy? A lingering injury? Or, perhaps, a decision to mail it in until Omar Minaya signed more backup catchers? Whatever the cause, it was particularly peculiar that he sustained talent in the midst of being sapped for power, as Wright hit .307/.390, a BA and OBP virtually identical to his PECOTA forecast, but with a .447 SLG that more closely resembled what one might expect from teammates Daniel Murphy or Fernando Tatis. Determining if Wright has any company in this regard does not involve too much heavy lifting, and the method I employed is more back of the envelope than anything else as well. I started by isolating any season from 18732009 in which a hitter amassed 300 or more atbats. Next, I calculated the average tally and standard deviation of home runs hit, broken down by league and year. For instance, qualifying hitters in the 1982 AL averaged 14.92 blasts with a standard deviation of 9.79; it could be approximated that about twothirds of these players ranged from 524 home runs. The next step was to compute the zscore for each player season, which is simply the number of standard deviations from the mean a data point strays. Reverting to our handydandy 1982 AL example, Tom Brunansky’s 20 home runs for the Twins turned out to be 0.52 standard deviations from the mean. From there, I manipulated the results to show a fiveyear span in each row, since Wright’s power dropoff occurred in the fifth year of a similarly long span of seasons. A grand total of 6,267 fiveyear spans emerged, but in order to determine if Wright has any peers, I needed to find a commonality amongst his personal numbers. The table below shows Wright’s pertinent data from 200509:
Looking at the zscore column, the commonality is that his home run total in each season either met or exceeded the average by 0.65 standard deviations. Leafing through the data for players who similarly bopped for the first four years returned 821 spans, but here comes the kicker: When I search for players meeting the aforementioned benchmarks but who fell to 0.60 or more standard deviations below the mean in the fifth season, a grand total of five rows are returned. Five! Over the last 150 years or so, there have literally only been a handful of players to experience a power dropoff from a previously established and high baseline of hitting home runs. The Oceanic Five:
What initially stands out is Wright was only 26 years old last season while the other cast members were 31 years of age or older. Another fantastically curious factoid is that the other four players barely surpassed the atbats minimum in that fifth year, ranging from 322340, while Wright surpassed the 500 mark. Both of these are points in favor of Wright’s anomalous season being about as rare as rare can be in this sport. If I loosen the fifth year parameter to be less than or equal to 0.40 standard deviations below the mean, the sample barely grows to 14 fiveyear spans. No matter how I choose to slice the numbers, it is evident that Wright’s season is of bizarroworld status. Even with the additions to the sample, Wright still turns out to be the youngest by a fairly wide margin, which suggests that what happened to him last year is unprecedented in baseball history. No other player has been as productive from a home run standpoint at such a young age, for a fouryear stretch, who suddenly turned into Juan Pierre (exaggeratory, sue me!) from a power perspective, while playing full seasons. Now, sure, I may have been a bit liberal with my querying criteria, but if I loosen it over the first four years, the largest sample I can produce while satisfying the integrity of an attempt to find comparables is 21 spans. Among those 21 spans, only one nonWright player was around the same age in his bizarro power season: Dusty Baker—yeah, him—who averaged around 19 HR/year from 197275 as a member of the Braves before bopping a grand total of four home runs in his first season in Dodger blue. Baker only played in 112 games in that 1976 season, but that does not explain the entirety of his dramatic decrease in home runs. Maybe he threw his arm out. Luckily for Wright, Baker quickly regained his power, averaging 23 homers per year for the next four seasons. Regardless of what I set as the searching criteria, examining what "similar" players did in the next year or two, which would equate to the sixth and seventh years in a span, proves futile, as most did not even amass 300 or more atbats after their power plummet, and only a couple went on to once again bash. The samples are so small, however, that their successes (or lack thereof) should in no way drive what to expect from Wright moving forward. PECOTA largely agrees, and a fun exercise is to compare his actual 2010 projection to what his 2009 PECOTA card forecasted for the 2010 season. PECOTA is very bullish on Wright moving forward and appears to be dismissing his 2009 power output, thinking it to be immaterial to what Wright is capable of moving forward. The antiBrady Anderson in every sense of the term, there are few reasons to expect Wright to continue last year’s putrid power output. Something definitely happened last season to cause such a vast decrease, but searching through the annals of history reveals very, very few comparables from even a broad perspective, indicating that his 2009 campaign was not only anomalous, but literally unprecedented. The actual 2010 projection pegs Wright’s weighted mean at .313/.415/.541, compared to the 2009 projection of 2010 which had him at .305/.405/.543. Simply put,In all likelihood, fans a decade from now will look back at his 10homer season as a blip on an otherwise fantastic career consisting of many more 30plus dinger years, and in a few weeks, when live regularseason baseball once again occupies our time and Wright looks like the player of years past, few will even think about what happened last season.
Eric Seidman is an author of Baseball Prospectus. 39 comments have been left for this article.

What a great article! I'd love one of you to look at the other players last year who were about the PECOTA 90% projection or below the 10% projection. I wonder if there are patterns?
Above, not about. I wish you could edit your messages here!