None of us are immune to observation bias. That applies to the observation of data as well as the observation of actions. None of us is immune to preconceived notions, and, at the moment, the notion that the Phillies' front office doesn't know what the hell it is quite prevalent amongst observers of the sport. But the essence of rational thought is to follow the evidence wherever it leads, and the evidence has led me to believe that Ruben Amaro Jr.'s handling of the Cole Hamels situation is not nearly as deluded as it is popularly believed to be, and that the processes that allege to prove that notion are actually shaped by it.
In our search for an accurate representation of reality, the admission of what we do not know is often as important as the inclusion of what we do. Any attempt to quantify Hamels' real value on the trade market not only requires us to formulate objective measures of individual performance and market behavior, but to acknowledge that we have no way of knowing how our objective measures compare to the ones developed and employed by organizations that devote significantly more resources to such processes than we lay people have the time and means to devote. There is a reason the Rays make their employees sign non-disclosure agreements. If the values we project using the public domain were as accurate as we sometimes portray them to be, there would be far less incentive to build and maintain such a private domain as expansive as the one that has developed in baseball over the past decade. In fact, the level of secrecy insisted upon by that private domain suggests that there is a significant gap.
I bring all of this up w/r/t Hamels' trade value because of a metric upon which most of the valuations in the public domain are based. Websites like Fangraphs and Beyond the Boxscore feature have produced a tremendous amount of research and literature in an attempt to determine the market value of a Win Share. It's fascinating stuff that has dramatically enhanced the ability of those of us in the public domain to analyze the sport with a critical eye. But our fascination with it might also blind us to a better way.
Case in point: One of the public domain's most extensive studies dedicated to determining the market value of a Win Above Replacement was authored by Lewie Pollis, who happens to be working for the Phillies' fledgling analytics department this season. Using all of the available data, Pollis estimated that the Cost Per Win during the 2013 offseason was roughly $7 million dollars. This was an increase of $2 million dollars over the previous estimate employed by most analysts. But what if a member of a major league front office simply told us the Cost per Win figure that their crew of professionals employ? And what if that number was a substantially higher number? On the one hand, we have evidence of Pollis' methodology. On the other hand, do we need any stronger evidence of accurate methodology than what we get from our knowledge that a major league front office produced it? Our ultimate decision rests on two questions:
1) Do we have reason to believe the methodology of the front office in question is inferior to the methodology we see in the public domain?
2) Do we trust that the front office is telling us the truth w/r/t the number?
Well, in an interview with Fox Sports Cleveland in October of 2012, Indians GM Mark Shapiro said that a win in free agency costs $9 million. Shapiro and the Indians are generally believed to have one of the more advanced analytics department. Let's say we give the benefit of the doubt to the Indians. Now, do we trust Shapiro? There is some evidence that we should, at least with regard to a top-of-the-rotation starter. Jon Lester's six-year, $155 million contract averaged out to $9.4 million per WAR if you use his average annual output over the previous three years as calculated by Baseball Reference and then project it over the next six years. Zack Greinke's 2012 extension averaged out to about $8.75 million per WAR.
Now, we're pretty sure the Indians' WAR metric is different from Baseball-Reference's, which is different from the one used by FanGraphs. And we're pretty sure the Indians' methods of projecting future wins over the life of a contract is more complex than taking the three-year average and multiplying it by X years. The point isn't that Lester's contract proves wins are worth $9 million, nor does it rule out the possibility that Shapiro was simply throwing out a ballpark number that is a simplified version of the actual inputs they use in these kinds of analyses. Even a minor difference in the calculation of past WAR and future WAR can dramatically impact the dollar per war figure. But I think we have enough information to say that we should at least consider the potential that there are teams who are interested in Hamels who might value each of his WAR closer to $9 million than $7 million.
Let's say we use $9 million instead of the $7 million that sites like Fangraphs use in projecting Hamels' value. Suddenly, the upper bound of Hamels' value in this Fangraphs projection jumps from a net of $23.5 million to a net of $30.2 million over the life of his deal. Fangraphs eyeballs the worth of a top prospect like Blake Swihart at between $35 million and $50 million. The Yoan Moncada contract suggests it is probably closer to the top end of the spectrum. But even at $50 million, the Phillies only need to eat $5 million per year to hit that mark.
Look, I would be very surprised if the Phillies landed a Top 20 hitting prospect. Those guys simply don't get moved in most of these deals. The Phillies didn't give one up for Roy Halladay, and they didn't give one up or get one in the Cliff Lee deals. But I also disagree with the notion that the Phillies should just be happy to rid themselves of Hamels contract and land a Top 50 pitching prospect in return. That's not to say that they won't end up having to accept something along those lines, but it is to say that the data isn't nearly conclusive enough to suggest that they are making a mistake in holding out with the hope of something more.