Ranking The Rankers: Who Has The Most Accurate Projections?

Fantasy managers employ a variety of strategies for ranking and valuing players on draft day:

1) Use your gut. This drafter doesn't need to prepare and bring in fancy spreadsheets. He waits until it's his pick, stares at the available players, listens to his heart and chooses. Unfortunately, fantasy football can be so random and luck-based that sometimes, he wins. Most of the time, however, this guy comes in last place.


2) Magazines. This drafter stops by Safeway to get a 6-pack of brewdawgs on the way to the draft and also grabs a fantasy magazine to gain the edge over his opponents. Most of these magazines were written in May and printed in June, but say they're good though September. Notice how the fantasy magazine may list "Ben Tate" or "Glen Coffee" as a good Sleeper pick.

3) Online Rankings. This savvy owner looks at a variety of recently updated online rankings from Rotoworld, ESPN, Yahoo, FFToday, and FFToolbox to get a sense of how various experts project players to do. If his league's draftboard has Michael Crabtree at 80, but most experts project Crabtree to be around 50 overall, then Crabtree becomes a good player to target in the 6th round, or even a steal in the 7th Round. This drafter will gain the most from today's article.

4) Obsessive / Expert. Finally, the expert drafter watches every preseason game, notes every coach's announcement, and local news blurb about training camp. He takes note of all other expert online projections, but then makes his own projections for every player down to the yard. He then uses said projections to create tiers at each position and mesh all the players together into a Top 200 using the value based drafting formula. We encourage you to do this, but unless you have little to no life, we'll help you out with that.

So for most of you, you should check out OUR top 200 rankings and consult a few other sites. Various websites proclaim that they have the most accurate projections and cite some made-up-sounding award to prove it - i.e. "Winner of The Fantasy Excellence In Accuracy Trophy (awarded by our own staff)". So, we here at FFMVP did the dirty work for you to rank the rankers. We took the Top 100 Rankings from last year for 8 different fantasy sites and compared each player's ranking to how he actually ended up performing. For example, Andre Johnson was the 6th most valuable fantasy player by years end, but NFL.com had him ranked 13th, with a deviation of 7. We then averaged out that deviation for each player to determine which website's projections had the smallest deviation from the actual results. The results may surprise you; they certainly surprised us.

First, it was a little disheartening that the average site had an average deviation of about 25. That means for any given player, most sites were off by 25 spots in their projections. I suppose that's not TOO surprising considering Ray Rice finished 4th overall but most sites had him around 100th overall. So when your site has Steve Breaston as the 100th most valuable player and you take him there, remember that he'll probably finish as the 75th player or the 125th player (or the best player in fantasy or the worst - who the hell knows?!?). Fantasy Football is a cruel and wild rollercoaster of emotions. Do your research to prepare but remember that you could draft Ben Tate in the 7th round only to see him land on IR at the start of the preseason (which is exactly what happened to me).

More importantly, here's how the sites stacked up in their average deviation from actual results. #1 is the most accurate site:

1. FFtoolbox: 22.6
2. NFL.com: 23.8
3. CBS: 24.1
4. Rotoworld: 25.0
5. ESPN: 25.1
6. Yahoo: 25.6
7. FF Starters: 27.3
8. Rotter of FoxSports: 29.0

It's a small sample size and only 1 season, but still, I won't be putting much faith in Mr. Rotter from Fox this season. I expected Rotoworld to be the best because I trust their analysis the most, but these are the numbers.

Remember to check out OUR top 200 and rankings for each position above. We compile them by averaging all these updated expert projections, giving more weight to the historically more accurate sources, and then tweaking a little where we see fit.