So I Guess This Means the Sox Should DFA Manny...

When Jack and I (ok, it was mainly Jack) created Primarily Baseball a few months ago, the last thing I ever suspected is that we would evolve into a forum for critiquing Baseball Prospectus. My first post, in fact, relied almost entirely on work done at Baseball Prospectus. Joe Sheehan is probably my favorite baseball columnist, Nate Silver is one of the smartest sabermetricians I know of working today, and most of the other writers are very good too.

That said, this has to be the dumbest fucking thing ever written.

(Password necessary, I'm afraid.)

For those who can't access the article--or choose not to suffer Huckaby's meandering verbiage-- the basic point is that replacement level, as traditionally understood, has been set much too low by baseball analysts, at least for players on the "left end" of the defensive spectrum (i.e. 1B/DH's and bad fielding corner outfielders). Replacement level is supposed to be the level of a typical AAA player or backup, i.e. the sort of player that can be easily acquired for little to no cost. The notion plays a large value in many value metrics, which are built around the assumption that it is the contributions a player makes above what a replacement level player would have done that constitute that players value to his team. Typically, replacement level is set to be around 75-80% of league average performance.

Now, the thesis that replacement level performance is actually higher than sabermetricians have thus far assumed is not, on the face of it, an absurd claim. It may even be true. However, Huckaby gives absolutely support to this thesis. His argument, so far as I can tell, is the following: Jack Cust was a replacement level player last year, in the sense that the A's acquired him from the Padres AAA affiliate after injuries had decimated their roster. Thus, we can use Cust as a reasonable benchmark for replacement level. Cust, in case you forgot, immediately started raking upon joining the A's: he hit 6 home runs in his first 7 games, including a walk off blast off one Joe Borowski. His numbers regressed as the year wore on, but Cust still finished the year with an extremely impressive .256/.408/.504 line (as that suggests, he was an amazing Three True Outcomes hitter; over 58% of his plate appearances resulted in a walk, a strikeout or a home run.) Using Cust as the definition of replacement level, Huckaby posts a long list of big name players--including Mike "MVP" Lowell--who were below replacement level by this definition.

The problem with this argument seems blindingly obvious: just because Cust happened to be freely available talent--a term sometimes, perhaps incorrectly, considered synonymous with replacement level--does not imply that the calibre of his performance was somehow replacement level. Suggesting as much makes as much sense as saying that that Mike Piazza represents "63'rd draft round talent." Jack Cust drastically outperformed general expectations last year.

Amazingly, in his subsequent chat, Huckaby denies that Cust's performance last year was really all that surprising, saying "And did Cust *really* exceed reasonable expectations by all that much? I don't think so." Um, well, I can't say I remember exactly what people's expectations were, but I am pretty damn sure they were a hell of a lot lower. Look at it this way: Cust's Equivalent Average--one of BP's stats for evaluating general offensive perfomance--was seventh in the American League, just behind Jim Thome and ahead of Vlad Guerrero (Thome and Guerrero have better slash stats on the surface, but they both played in much better hitters parks than Cust). Guerrero and Thome, in case any one forgot, are Hall of Fame calibre hitters who had terrific years. Does anyone think that, if the expectations were that Cust would be anywhere close to that level, he would have begun the year rotting in AAA? Don't you think that the Padres--who are, I might add, one of the more sabermetrically astute franchises in baseball--would have promoted him if they thought he would be a better hitter than Vlad fucking god damn Guerrero? Hell, there is no way Billy Beane thought he was going be that good; if he did, why did he bother signing a 38 year old Mike Piazza in the off-season for vastly more money than Cust would receive despite the fact that Piazza hasn't been the hitter Cust was this year since 2002?

Cust was a desperation move on Beane's part that payed off big time. One could argue, of course, that Cust's stellar minor league numbers merited some team giving him a major league shot. But suggesting that Beane thought he was going to receive the quality of production that Cust offered--and, implicitly, that there are numerous minor leaguers out there who could provide something comparable to Cust's 2007 performance--is batshit insane.

That was way too much time to make a really obvious point.