Let’s say you want to buy a gallon of milk. All you know about the price is that at Pathmark it is $2.50 and at Waldbaum’s it is $4.75. You know that these are the highest and lowest known prices; but it is possible that it might be cheaper or more expensive at stores that were not tracked. You are going to the local grocer and do not know what the price is or how many others want it at that price.
How much do you expect to pay? Are you sure you will be able to get a gallon of milk? After all, maybe it is so cheap that it will be all gone.
Another problem; you have heard rumors that stores are selling milk in gallon bottles that are only half filled, or even less. And you can only go to your local grocer; you cannot shop around.
It seems beyond any doubt that you have no idea what the price will be, other than that there is a good chance, but not a certainty, that it will be somewhere between $2.58 and $4.75. And you do not know whether you will be able to get a gallon at your local store, unless you have some inside information. And you may buy a bottle, only to have it be half full or less.
Welcome to the fallacy that is average draft position.
Let’s take Troy Tulowitzki as an example. As of this writing he has an ADP of 27, according to MockDraftCentral.com, with a high of 18 and low of 37. That is not even a large spread; Dustin Pedroia is at 36, with a high of 21 and a low of 51.
If you have the wheel at 23 and 24 can you be certain of getting either player?
Like the gallon of milk, you know the high and low prices, but that gives you only some information on what the average cost might be. How that helps you in your individual draft is guess work. It is not a matter of analysis or logic; it is a pure stab in the dark.
In roto you have differing reports of the high, low and average cost of that gallon of milk. CBS, ESPN, MockDraftCentral, Couchmanagers.com and a zoo of other sites have differing values because they all have different users.
The fallacy is assuming that because you know something about the overall population of results you can make an inference about the specific. If you did 500 drafts with the same group, you might find that your results for an individual player approach the ADPs since you will have a larger sample. But this makes no sense when you are trying to prepare for one, two or three drafts against different, and in many cases unknown, owners.
Why do we do it?
Fantasy owners know they must have some process by which to make decisions. So, in the pursuit of a process in drafts, they devise the infrastructure of ADP analysis, even though it makes zero sense. From this flawed infrastructure they have a framework within which to make decisions that can be rationalized and justified.
The ADP analysis we typically see comes from an attempt by drafters to have the price certainty of auction players. Fantasy Baseball is, at its core, about economics. Auction players can use economics to make decisions. Drafters want to be able to do so, but they cannot in any reasonable understanding of a market. But that doesn’t stop us from trying.
The other reason why we do it is conformity. Every so-called “expert” discusses ADP, analyzes it to death and rates decisions as good or bad based on it. So who are we, the non-experts, to disagree?
Solomon Asch did an experiment called the Asch Experiment, and it explains why this happens. See the picture above, and then read on.
The test subjects for the Asch Conformity Experiments were brought into a group of supposed random people, but who were all Asch’s accomplices. The whole group was shown a card with one line on it. They were then shown a card with three lines and were asked to compare the one line to the other three and find the one that was the same length.
That seems easy enough.
However, Asch’s plants in the group argued strongly for the wrong line on the second card, even though in an unbiased atmosphere it would be clear that they were wrong. They were asked to give answers out loud, argue vociferously and come to unanimity.
And the unwitting test sap would go last.
In one-third to three-quarters (!) of the cases the unwitting subjects gave the wrong answer, and afterwards admitted doing so because of the others.
This is why we use ADP; like the poor sap in the Asch Experiment we are led to believe by experts that it is not only legitimate but our decisions will be rated as good or bad by them. And who wants to be ridiculed.
My view is that ADP is mostly worthless, as you can tell. It is beset on all sides with illogic and fallacy, and does not do what it purports to do, even within the framework it provides.
Latest posts by Patrick DiCaprio (Posts)
- Closer Identifier Algorithm Results June 23, 2014 - Jun 23, 2014
- 2014 Fantasy Baseball Pitch or Ditch for Sunday June 22, 2014 - Jun 21, 2014
- Closer Identifier Algorithm Results June 16, 2014 - Jun 16, 2014