White Noise
The ABC's and 1-2-3's of planning your draft strategy

I have seen articles on the alphabet soup of fantasy football -- initials such as AVT, VBD, CIA, CYA, ADT. And I have seen columns on ranking players which like it or not, has to be done in one form or another. - NFL Football -
For years, I was a proponent of producing player predictions based on a mathematical model. The problem was, and remains, that no model can accurately predict the performance of NFL players. Sounds like a simple enough concept, but many don't get it. They spend hours poring over stat sheets from the past five years in an effort to project what Michael Bennett is going to do in 2005. When the truth is that no one has a real clue what to expect from Bennett this year. Or any other player for that matter. - NFL Football -
So what I want to do with this column is put down some easy to follow instructions that should help you manage your draft in such a way that you come away feeling like you have an edge on every other team. To do this, we are going to use three of the above alphabet combinations.
AVT (Average Value Theory) - NFL Football -
The problem I have with just about everyone's projections is that they tend to overproject the mid- and lower-tier players.
Check out any projection and you will find about 20 players targeted for 1,000 rushing yards and 10 or more TDs. Last year, only eight players rushed for over 1,000 yards while scoring 10 or more TDs on the ground. The year before there were eight such players too. This causes a disproportionate value to be placed on lower-tier players and negatively affects the drafts of anyone following those projections. The simplest way to overcome this is to use historical results to project slotted performances. - NFL Football -
Sounds complicated doesn't it? Well, it's not really. What you need to do is find a program that will allow you to input your scoring system and then produce a list of the top 100 performers at each position, with their scores, for each of the past three years. Average those numbers to come up with a projection for the coming year. - NFL Football -
Sideline Sports has a program called Fantasy League Manager that I use for this purpose. I paste those results into a spreadsheet that combines and averages the totals for me. With this, I can tell you that I expect the 18th-ranked running back to score 199 fantasy points in 2005. I can also tell you that between 2001 and 2004, the 18th-slotted back scored 180, 219, 197 and 200 points. Will this year's 18th-ranked back score exactly 199 points? Probably not, but he should be close enough that I don't over, or undervalue him. - NFL Football -
One thing to note: This system is inherently weak in predicting career seasons by players. There is no system in the world that is going to predict a 4,600-yard, 49-TD season like Peyton Manning had last year. That means the top rated players might actually outperform their projections by quite a bit. Last year, both Manning and Daunte Culpepper outperformed their slotted projections by about 90 points. But getting more than you expect is usually a very good thing.
Now all I have to do is rank the players. And for me, that is much easier than it was in the past. Because in the past I would spend days formulating projections of stats for all the players I might even think about drafting. Now all I have to do is rank the player. Do I think LaDainian Tomlinson will be better than Priest Holmes? Will Lee Suggs outperform Cedric Benson? I don't worry about the statistics -- which I can't accurately project anyway -- I just worry about who I think will have the better season. And I can change my mind whenever I want without worrying about what changing one player's projections will do to another's.

Greg Kellogg, Special to SI.com