Agnes, don't you think that these two sentences contradict themselves just a tad?
And everytime you mention an Anderson or Hidalgo alls you are doing is citing the exceptions to the rule, and the rule is most players do not break out in their rookie seasons. In most cases, it takes them a few seasons to actually start producing up to their capabilities, and once they do they normally do not revert back to their struggling early seasons. Other than Pujols, that is the norm. Besides, what would you say if Beltre played in the minors until he cracked through last season at age 25 and put up those types of numbers?
So I guess you are backing James in saying that come draft day I should be taking Pickering and MacPherson ahead of this list of FA's since you say these MLE's are just as accurate as any other projection method? Looking at some of his projections I will have to say that you are right as long as you are using his projections in comparison.[/quote]
No, I don't think it's a contradiction. It's a question of how much
you weight past data. If you weight last year at 100% and prior years at 0%, then you expect Beltre to post an OPS over 1.000. If you weight last year at 0% and the two prior years, you project about a .720 OPS.
James' estimate turns out to be very close to what you get if you use .5/.33/.17. So, my statement is that you definitely use past data and you definitely weight last year more than any other year. But, you don't weight it at 100%. So, the real debate (in terms of weighting the most recent year) is whether it's 50% or 60% or 70% or higher.
In terms of minor leagues and rookies, I would be saying the exact same thing about Beltre as I did about Posednik last spring...when a guy's rookie season is way above what his MLEs were as a minor leaguer, it's unlikely to be sustained.
My argument has nothing to do with rookie seasons or player psychology. It simply has to do with number patterns. When you see a string of numbers like Beltre (or Anderson or Hidalgo), you'll win a lot of money by not placing too much weight on the aberrent number. You weight it, but don't overweight it.
Now, assuming that James is recommending you take Pickering or McPherson over Beltre takes his projections far beyond what I think he would say is appropriate. His projections are not made with fantasy players in mind. They don't take into account likely playing time, for example. Like most SABR guys, James makes his predictions assuming the player gets to play.
Furthermore, his predictions don't translate into fantasy value because they don't value position shortages, lineup, certain fantasy stats, etc.
So, don't leap to the conclusion that James suggests you pick Pickering. He's predicting what Pickering will do, if given the opportunity.