expert fantasy leagues are full of idiots.
The guy running a fantasy league I'm in recently complained that Giambi, Ensberg and Berkman have all had down years, implying that there was no way that he could have anticipated this, that he's a victim of profoundly bad luck, a jinx, yada-yada, and I pointed out that slow sluggers in their thirties often decline rapidly and referred to PECOTA's high Collapse and Attrition rates for all three.
Here's one response I was given:
PECOTA's 2007 prediction for Giambi should have taken into consideration it's own qualifiers. In no way can PECOTA predict a performance for a baseball player and then throw in elements such as collapse rate and attrition rate (whatever they are) as being a variable influence or indication.
First, PECOTA sells itself as having developed an algorithm that comes up these answers.
Second, Algorithms can only work in with a finite set of data/data points. That puts PECOTA in a shamans position.
Third, as soon as PECOTA has been proven wrong with just a single event, it makes all of it's other predictions... suspect.
To apply mathematics to any process, the population has to be defined and if a new variable is introduced, the new entrant then becomes an influence and therefore the "algorithm" is required to be re-processed. Example: if a minor leaguer were brought up the majors and was not initially considered in the study, then that element was an unanticipated influence...
PECOTA may be correct on many predictions (I have no idea) but it would not be due it's "algorithm". PERCOTA may as well predict what the weather will be in Phoenix at 3pm on November 28, 2007.
That wasn't the commissioners' response, that was another smart-guy's response. The Commish's response was a little more intelligible:
Why don't you tell PECOTA to suck my balls.
It's easy to forget running in certain savvier circles that most people haven't advanced beyond the "RBIs is good" mode of thinking, or that Ludditism still abounds.