Long time no see. Sorry for the absence, it has been a trying summer. That said, something recently posted on the NFBC forums got my attention, so much so that I felt inclined to address it in this space.
The original theme was the poster’s lament that he did not go with his gut, feeling that Curtis Granderson would have a very good season and instead allowed himself to be swayed by the so-called experts, who were not as optimistic. A few more general comments followed, with the standard “there are no experts” mantra. The thread in question can be read HERE.
My objective is not to discuss the notion of fantasy baseball experts - been there, done that. My objective is not to express ire that a group of which I am a member is being taken to task. Load me up with sodium pentothal or snare me with Wonder Woman’s Lasso of Truth and you will get an earful, much of it will not be complimentary to some of my fantasy baseball brethren. No, today I wish to address something that is particularly frustrating to me, and that is the expectations from a person perceived to be, advertised as or is self proclaimed as an expert.
Basically, I feel much of what is expected from an analyst is at best unfair, and at worst, wrong. The job of an analyst is not to tell you what is definitely going to happen but rather what is most likely to happen. The problem is, what is most likely to happen is not always what happens. I know this analogy is not perfect, but it works. I have a pair of dice, what is your guess of the total after I roll them? The most likely outcome is seven. The proper answer, based on the available data is seven. If you guess anything other than seven, you are going with your gut. The job of the fantasy analyst is to look at the data and tell you the equivalent of rolling a seven. Here is where the analogy breaks down. Probability says that even though seven is the most likely result, it only occurs 16.67 percent of the time. This is not the time or place to discuss it, but the minimal expectation when it comes to player projection is 70%, which is the accepted accuracy of the Marcels, a weighted average system so simple that even a money can do it.
The point is, the job of the analyst is to absorb the data and decide what is most likely to happen. Different analysts use different data. Some may use scouting data, others a computer spreadsheet. Some data is open to interpretation; other has an accepted means of application. As a numerish disciple, I will focus on the spreadsheet genre.
With the caveat that not every analyst is as schooled in numerish (see Sodium Pentothal comment), a reputable analyst crunches the appropriate numbers and offers what probability dictates in the most likely outcome. Again, not everyone crunches the proper numbers or draws the proper conclusion, but I digress. The analyst should produce a completely objective expectation, representing the most likely outcome. Bringing it back to the 70% success rate, if there are 100 players with the exact same history, they should each have the same projection, with 70 of them being “right” at the end of the season.
Let us now segue into the notion of a gut call. Unless your name is Special Agent Leroy Jethro Gibbs, my personal contention is gut calls are not wise. Do I make them? Of course I do, everyone does. We would not be human otherwise. But, the fact remains, it is best to avoid them. Here is the way I look at it. If I use my gut, there is usually a reason. Whatever the reasoning was, I should rightfully apply it to all similar players. Smith had a great September; my gut says he will carry that over to next season. Proper projection theory would study the probability that a great September is carried over and every player that had a great September would be given a bump, otherwise the decision is subjective, replete with bias. Again, everyone, including me is guilty of this so who am I to judge? What I am judging, so to speak, is what is expected of an analyst.
Perhaps my pet peeve is the complaint that so and so never goes out on a limb. Hey, that’s a GOOD thing as it shows they understand projection theory. That said, the better analysts can certainly point out what it would take for the player to end up on the good side of that mysterious 30% missed. To take myself to task, here is my spring profile of Granderson, the impetus of this conversation:
“Granderson's fly ball rate is increasing which is good for power but bad for batting average, especially when he strikes out as much as he does. If you can cover the average, Granderson will give you homers. But beware; he has missed some games two of the past three seasons.”
For the record, I had him hitting .264 with 29 homers in 550 at bats.
Maybe I hinted at the power, but I completely swung and missed at the necessary event that could result in a break out and that is an improvement versus southpaws. Of course, this is exactly what happened, though there are strong signs there is some good fortune involved as his contact rate against lefties is still poor. When he has made contact, he has done more with it, which is not all good luck, but that is a story for another day.
My point is, saying Granderson would hit .264 with 29 homers in 550 at bats was the equivalent of guessing seven. Based on his history, that was the proper expectation. In retrospect, I wish I mentioned the lefty thing, but to upgrade the projection would have been akin to guessing ten or eleven, and there was just not tangible evidence to suggest that is the likely outcome. Was it a possible outcome? Absolutely.
Tying everything together and using Granderson as the specific example but this should be more of a global comment, I do not feel it is fair to take a so-called expert to task for “missing” on Granderson. He was part of the 30%, he rolled boxcars. If you drafted Granderson, feeling this was his break out year, congratulations. But, I ask you, did you draft others for the same reason, whatever that was? If not, was your Granderson call “right” or “lucky”?