May 5, 2004
Lies, Damned Lies
Aging AlfonsoAlfonso Soriano (.321/.357/.472 after Monday night's victory over Tampa Bay).
Soriano has undergone a couple of changes since his last incarnation as the undisciplined Yankee second baseman whose terrible second-half and postseason campaigns were enough to trigger Bronx Jeers at his every at-bat. The switch from the neo-classical, interlocking N-Y to the tacky, scarlet T on his uniform breast is the most obvious, but Soriano has also changed batting order positions (Buck Showalter has him hitting third, instead of first, a role that he is considerably better suited for). He's also switched birthdays--or at least, birthyears. Turns out that A-Sore was born on the 7th of January, 1976, and not the same date in 1978, as he was previously listed.
John Hart and the Rangers knew full well about the change in birthdate before agreeing to acquire Soriano for Alex Rodriguez. Indeed, baseball teams--and baseball fans--have grown pretty well used to these sorts of surprises; before Soriano, there were only a few hundred other players whose reported birthdates were revealed to be incorrect. With a few exceptions like Bartolo Colon, however, most of those guys were marginal prospects in the lower minors, and not an established star like Soriano, for whom any change in expected performance could potentially cost his club the equivalent of millions of dollars in value.
The Rangers weren't concerned--but should they have been? Following is a comparison of Soriano's weighted mean PECOTA projections, before and after the new birthdate:
Alfonso Soriano: 2004 PECOTA Comparison AB BA OBP SLG EqA VORP Age 26 631 .305 .354 .550 .297 56.8 Age 28 625 .299 .345 .537 .292 52.1The difference boils down to around five points of VORP--that is, five runs, or about half a victory. Most of the difference stems from a decrease in his breakout rate--the 26-year-old Soriano was given a 14% chance to have a breakout campaign, compared to 8% for the 28-year-old version. It's a little bit less likely now that Soriano is going to emerge as the true, Sosaesque slugger that some people have confused him with.
At the same time, five runs is five runs; as discerning as PECOTA can be, the difference is well within sample size variance. When the news broke, I received a flurry of e-mails form Soriano's fans and bashers alike, all of whom seemed to place a great deal of importance on the age change. Several of those e-mails intimated that the change was particularly important since it meant that Soriano had already experienced his age-27 peak season, rather than having it to look forward to. Couple that with Soriano's poor second half, and the inevitable Juan Samuel comparisons, and the consensus seemed to be that Soriano's days as an All-Star caliber player were already behind him.
But in fact, the opposite is true. The fact that the change occurred around Soriano's peak means that it is in fact less important in the near term, than if a similar discrepancy had been revealed for a player who was a few years younger or a few years older. Remember that a 'typical' aging curve looks something like this:
The area between ages 25 and 28 is highlighted in red. What is noteworthy is that this is the flattest portion of the graph--that is, the point in a player's career at which his performance can be expected to change the least from season-to-season. While the typical player can expect to be a bit more productive at age 27 than at age 26, and a bit less productive at age 28 than at age 27, the differences are almost indiscernible.
The converse of this is that, the further outward we project Soriano's statistics, the more significant the age change becomes:
Alfonso Soriano: Five-Year Forecast Comparison (Wins above Replacement) 2004 2005 2006 2007 2008 Age 26 5.5 4.5 4.4 4.3 3.7 Age 28 4.9 4.7 3.8 3.6 2.5 Difference 0.6 -0.2 0.6 0.7 1.2Check out the first and the last numbers: the difference in Soriano's expected performance in 2008 is twice as great as it is in 2004. If we projected his statistics out even further, the gap between the two forecast lines would continue to expand.
This is also apparent from looking at Soriano's top five comparables:
Alfonso Soriano: Top Five PECOTA Comparables Rank Age 26 Age 28 1 Ernie Banks Kelly Gruber 2 Juan Samuel Max Alvis 3 George Bell Gene Freese 4 Sammy Sosa Juan Samuel 5 Andre Dawson Raul MondesiThe Age 26 list includes players like Sosa, Dawson, and Banks, who continued to be productive well into their 30s. The Age 28 list, by contrast, consists instead of players like Gruber who were productive at their peaks, but did not age particularly gracefully. It is fair to say that the age change radically reduces the chances that Soriano will put together a Hall of Fame-type career.
Whether or not that ought to concern the Rangers is a different matter. For all intents and purposes, the team need only be concerned with Soriano's performance through 2006--after that year, he'll become a free agent. The net difference in Soriano's expected performance in the three years between 2004 and 2006 is about one win, or so thinks PECOTA. Because the arbitration process tends to focus on past performance, rather than future, expected performance, it is true that the Rangers might not be getting quite as much bang for their buck--but the difference in expected value is small when compared to the magnitude of the sums exchanged in the A-Rod trade. (Once he becomes a free agent, any difference in Soriano's expected value should theoretically be arbitraged out by the bidding process, though it is probably the case that teams do not consider age as much as they ought to when bidding on free agents).
Every run counts, of course, as does every dollar. But the Rangers have bigger concerns on their hands--like whether Hank Blalock can start to hit lefties, Mark Teixiera can mash 30 homers in spite of his late start, and R.A. Dickey's mutant arm can maintain an ERA under 4.00. The guess here is that the team stays in contention all season--and that Soriano finds a new set of fans to torment with his late-season plate discipline misadventures.