In Sam Walker’s book, Fantasy Land, he wrote stats are only able to predict the performance of ballplayers with about 60 percent accuracy, at least according to the people at NASA he hired for this analysis. Let me first qualify this by stating I don’t have the analysis and data to verify this and quantify my thoughts. NASA minds in combination with their computer power seem like a safe bet for information accuracy, but I would like to know the range factor for determining failure. Maybe I should attempt to contact Mr. Walker, but that doesn’t seem like something I’d do. Sixty percent seems awfully low. If I were to have that success rate for financial projections — a less volatile environment but not without its human factors and market variables — I would have been unemployable a decade ago. Back when I used to do projections for fantasy, before the minor leagues and writing consumed the time, I believe I had a much higher success rate, but again, have no proof without the range factor.
Projections can be handled in numerous ways, and thus not all projectors are equal. I could run each projection through three separate and unique formulas altered for baseball from my finance methods to arrive to a set of mean numbers, but that would be cost inefficient given I’ll need to run this for each of roughly 650 players at a minimum. And I’d need to check each one individually, and adjust one-third for being inaccurate due to historical sample size, as well as all the variables created by movement and changes to rosters etc.
Some projectors like to do this from the get-go, relying totally on instinct whereas others prefer to let the program do the prognosticating. I believe most of the fantasy professionals have to manually perform at least a percentage of them for the nature of the beast is competition against one another. I personally go back to my professional training and use methods that permit substantiating the claim, and then try and find an apt player comparison to justify my intuition choices.
With that knowledge one projector will certainly not be equal to another, and some will ultimately prove to be better over the long haul, producing the higher percentage of accuracy consistently. If you have the time and, or, money, I’d suggest picking up various projections and use an average, throwing out the high and low from each one, like Olympic judged events. Or, if you’re new to the notion of projections and don’t have the time, find the projector who best represents what your instincts are telling you and run with that.