Sigmund Bloom and I have a lot of conversations throughout the course of a week. One vibrant, on-going topic is the state of fantasy football rankings. There's value to a linear set of rankings spanning 1-250 that are based on statistical modeling. It's the safest way to create a blueprint for a draft.
There are also notable gaps with this format. Rankings are harder for the user to see the draft with a broad perspective that tiers offer. Not that you can't look at various possibilities at one time with a vertical list, but tiers make the decision-making tree of long-term cause and effect easier to see.
Most rankings based on statistical projections don't incorporate the value of upside. Dorial Green-Beckham and Eddie Royal might be within five picks of the other, according to an ADP list or ranking, but the potential for Green-Beckham to produce fantasy-dominant scores with his big-play ability is much higher than Royal. What about Stevan Ridley versus Antone Smith? Both have a similar range in ADP, but Smith has a far lower chance of becoming a 15-touch-per-game starter than a healthy Ridley.
Then there's confidence factor. A fantasy owner or analyst may have DeAndre Hopkins among his top-15 receivers based on projections, but when asked about Hopkins they may also tell you that they are so concerned about Houston's quarterback situation and the Arian Foster injury that they pass on the Texans' receiver in every draft. Or, like me, you might like Falcons tight end Jacob Tamme's potential so much (even if you can't truck with the idea of projecting production that, like me, places him as a starter on your tight end board) that his ADP of 242 factors into your overall draft strategy.
Straight up and down rankings lack a lot of valuable context. It also creates a lot of layers of conventional thought that create barriers to taking players that you really want on your team. Some barriers are healthy, because the logical thought baked into the rankings process helps you realize that some of your views are too impulsive or risky for the good of building a team.
But there are unintentional barriers baked into the conventional thought process of linear rankings that discourage you from recognizing the value of strategies that don't fit into a traditional ranking process. Bloom and I have been discussing these issues within the context of tiers vs. rankings and using my tiers article with the safe, balanced, and high-risk builds as a starting point. The conversational evolved to a point where Bloom came up with an idea about creating a Desirability Ranking.
Something about typical fantasy rankings has been bothering me lately. By listing all relevant players in order of draft desirability, we still don't really have anything actionable until we add a denominator of ADP. If I tell you a player is my #5 running back, that doesn't tell you whether I would recommend drafting him or not. When you add the dimension of ADP, it becomes clear right away who my rankings recommend targeting and who my rankings recommend avoiding. This leads to an experiment idea of rethinking rankings. Instead of ranking players by draft desirability (or projected end of year finish), why not rank players by desirability of drafting them at ADP? This table will show you the rough ADP of a player, and where I would consider drafting the player. This ranking should help when prioritizing targets pre-draft and understanding the next best outcomes when targeted players go off the board ahead of you. Any and all feedback welcome!
Bloom's ADP Desirability Rankings are simple. It involves the player, his ADP, and where you would pick the player to ensure you got him. Bloom uses these three pieces of information to create a ranking. In this scenario, Andre Ellington might have an ADP range of the 4th round, but you value him enough to take in the 3rd and it makes Ellington the third-most desirable running back on your draft board.
As you can see, this is not a ranking format for the conservative, "I need to see hard numbers" drafter. It's an experiment. I've told Bloom it's like trying to cut through all the traditional movements of old-school Fantasy Kung Fu and create our own version of Fantasy Jeet Kune Do.
I like this idea, but I wanted to see if I could use statistical projections for my traditional rankings as the backdrop? Could I convert my traditional rankings into ADP Desirability Rankings?
This article is the result of my efforts. It may not be the best way to do it, but it's taking another step forward in exploring this big room for the first time. I'll share the process of creating the rankings, rankings at each position, and three teams I selected by using these rankings.
Converting Traditional Rankings to ADP Desirability Rankings
The highlights:
- I updated my traditional rankings and added my ranking order to the right of each player's name.
- 12-team PPR league
- Starting lineup of 1QB, 2-3 RBs, 3-4 WRS, 1-2 TEs, 1 DEF, 1K
- I added the ADP of each player to the left of each player's name.
- I created a four-year average value VBD for each position (QB1, QB2, QB3, etc.).
- Four years of positional rankings (2011-2014) with accompanying fantasy points.
- Compiled the average fantasy points for each spot.
- QB1, QB2, QB3...QB20, QB21, QB22, etc.
- RB1, RB2, RB3...RB20, RB21, RB22, etc.
- WR1, WR2, WR3...WR20, WR21, WR22, etc.
- TE1, TE2, TE3...TE20, TE21, TE22, etc.
- Generated the average (VBD) value for that four-year average for each of those spots.
- This average value is my projected value for my 2015 rankings and not the VBD I projected to originally create the rankings.
The first column is the current ADP based on Footballguys' compiled tracking of ADP. The third column is the average value based on four years of season-ending performance for players at that final ranking. In other words, the top-ranked player averaged 246.06 VBD, the seventh-ranked player averaged 158.21 VBD, the third-ranked player averaged 204 VBD, etc.
The fourth column is the same average value as described above, but matching my rankings for these players. Marshawn Lynch has an ADP of seven, which matches an average VBD of 158.21. I value Lynch No.4 overall and the average VBD for that spot is 199.65.
The next step is creating a column that shows the average value difference between the ADP's VBD and my rankings.
The final step is one where I generate a number that reflects my view of ADP Desirability:
The sum of my rankings average value and the change column (ADP average value + My rankings average value=ADP Desirability).
This desirability doesn't incorporate the value of upside as much as I'd like to see in the later rounds, because I'm still basing a lot of what I'm doing on the traditional statistical projection model. It's something to continue tweaking, and like Bloom, I'm open to suggestions. I'm posting it because I want to advance the conversation.
The Rankings
QB ADP Desirability Rankings
The good? I think this method of creating the rankings adds more weight to early picks that you value. When Bloom and I discussed his tight end ADP Desirability Rankings, he believes that Rob Gronkowski or Travis Kelce feel a little too low.
I agree. If you properly identify the value of an early-round player you should be rewarded for it. Upside Down Drafting is an inside-out view of this idea: Punish the majority of fantasy owners who fail to pick the right running backs during the first two rounds and benefit from their folly.
My rankings add value to Roethlisberger, Brady, Eli Manning, Rivers, and Cutler as potential picks to build around. Remember, this ranking isn't based on how many points they'll earn. I believe this information gives you a road map to determine the hot spots in your draft to pick a QB.
This method could supply an added boost of confidence to those who believe in taking a stud like Luck and Rodgers. At the same time, you'll see that taking one of the top two studs is far from a "must" when you examine the context of all the positioning rankings side by side.
The bad? It's easy to forget that this isnt a ranking based on season stats and the position of Blake Bortles and Brian Hoyer can be jarring to one's sensibilities. I have more to examine, but I believe that this Desirability value that I've created provides a natural cut-off point between positive and negative.
I see the Bortles score as a place where you know there are 10 starters worth considering at various points at the draft before the Jaguars' second-year starter. Every point beyond Bortles signals an increasingly less desirable place to select a passer. It doesn't mean you should avoid the quarterbacks with negative desirability numbers, but their ADP relative to my ranking of them isn't as desirable as the other options with positive values.
I'd still consider Russell Wilson, Tony Romo, Drew Brees and others on this list, but if my projections vs. value rating has accuracy this year, building around the first 10 players should offer me more bang for my buck.
Note that Tom Brady's value is top-five. I adjusted Brady's VBD based his rest of the season value. If you think of Brady in these terms, he becomes a viable player that you strategically build a team around.
RB ADP Desirability Rankings
A longer list of players illustrates where confidence in a player's value above his ADP can have a huge payoff--especially when you value a 4th-round pick as a 2nd-round producer (Andre Ellington) or a 7th-round pick as a 4th-round value (Doug Martin). Desirability rankings encourage you to act on your player values and build around them.
The format also discourages you from taking players because "that's where they're ranked, and I'd feel stupid if I didn't." Eddie Lacy is a good example. I like Lacy, but my desirability for his value is 11th versus ADP of 5. If there are early-round backs (Marshawn Lynch) or receivers with a greater D score and similar ADP, I'm probably passing on Lacy.
But let's not go crazy here. There's no letter of the law with this format. If I'm drafting second overall and Adrian Peterson is available, I'm not passing on the Vikings back because he's fourth on my RB D-list. His D Value is 195.30, which is a closer to my RB3 C.J. Anderson than my RB5 Andre Ellington.
My top WR's D Value is 221.4 (Antonio Brown) and the No.2 and No.3 receivers have D Values of 189.55 (Jordy Nelson) and 189.43 (Dez Bryant). None of my tight ends or QBs have D Values this high. While I will consider Antonio Brown vs. Adrian Peterson--and in many cases go that route in PPR leagues--I may find that I like the gap between Peterson and the rest of the backs because I feel I can find enough receivers with strong D Values on my board to support taking Peterson over Brown.
D Values also help owners spot different strategic options to build a team. It's clear that Andre Ellington, Justin Forsett, Joseph Randle, Doug Martin and Arian Foster present enough value to me at their ADPs that I should try to build around some of these players where I can. A Desirability list like this emphasizes that opportunity whereas a traditional draft list can unintentionally encourage you to take a passive "let them fall to you" mentality where you're not building around those core values.
WR Desirability Rankings
You begin to see what I mean about constructing around values when you see how much I actually like A.J. Green, Nelson Agholor, Amari Cooper, Brandon Marshall, DeAndre Hopkins and Keenan Allen relative to their ADP. [Author's note: rankings reflect Jordy Nelson's value pre-injury]There's a confidence factor that Agholor with an ADP of 77 should out-perform his value. While I'd find it a complete waste of resources to take Agholor in the first five rounds, why not build around Agholor as a foundation to your plan and him within 10-12 spots of his current ADP?
Marvin Jones is my No.14 D-Value receiver. His ADP is 172. It doesn't mean I value Jones as the No.14 fantasy scorer at his position, but I value him so much more than his ADP that it tells me that he's a must-grab in the later rounds. It also tells me that if I have a close call between a WR4-WR5 and another position, knowing that Jones will be a must-grab later will help me break the tie in favor of the other position.
Tight End Desirability Rankings
The TE D-Values validates what many see as common sense with drafting this position. If you don't take one of the top four option by ADP (Gronkowski, Graham, Kelce and Olsen), the better values are available during the second half of the draft. Jacob Tamme's presence with a D-Score slighty above Olsen's is a great discussion point.
If you're an extreme risk-taker, you could wait until the very end to select Tamme and ride with him as your TE1. I have the Falcons TE ranked 9th overall and it bumps his D-Value to fourth on this list. Even I find this a nutty risk to take.
It's probably wiser to consider the likes of Delanie Walker, Heath Miller, Kyle Rudoph, Eric Ebron, Jared Cook or Jason Witten earlier in the draft and wait until the final rounds to add Tamme. Then you at least have a backup plan if Tamme doesn't pan out. I often take Tamme as my backup plan to one of the four top options, knowing I can take depth at other positions until the very end.
Using D-Values To Build Teams (3 mock-Builds)
I used my rankings to draft teams from the 1, 7, and 12-spots in 20-round drafts. Overall, I'm pleased with the results.
Team 1.01: The receiver choices appear risky in the first half of the draft, but I like enough of the options in the back half to supplement my wideout depth chart that it's worthwhile. Hopkins, Cooper and Agholor have fine upside, even if the QBs throwing to them are real question marks. My confidence in Brown, Johnson and Royal is high enough that even if two of my three starters falter, I think this second trio can keep me competitive if the rest of my team pans out. Funchess and Britt offer big-game upside and potential for greater consistency than their ADP's indicate.
Not a team for the faint of heart, I'd easily roll with it. the squad is an expression of my confidence in the players I have valued above more proven names. You have to be willing to lose big to win big. And if you're worried about looking bad for a season in front of your competition then you're not ready to win big.
Team 1.07: I love the balance of this team. Lynch, Ellington and Foster could prove dominant if Foster returns to form by Week 6. I built in temporary depth with Williams and Woodhead thanks to the suspension of Le'Veon Bell and the slow progress of Melvin Gordon adapting to his passing game responsibilities. If Ellington and Foster can't stay healthy and Lynch remains the anchor of the Seahawks' fantasy producers, I still have a foundation of a good team thanks to depth elsewhere.
Green, Allen, Cooper, and Brown have the talent and situation to each earn top-15 production at their position. Taking Eli Manning and Brady with consecutive picks gives me trade bait or depth at a bargain. Creating a team with these rankings wasn't difficult. I looked across the rankings as if they were tiers, examined the ADP as the matching point for my round, and weighed a potential choice versus the options that would be available for the next 3-4 rounds.
Team 1.12: It's another squad where I walk the talk. Anderson and Nelson aren't difficult picks. For me, neither is Ellington. Randle, Bryant and Agholor with consecutive picks is devoid of any long-term proven talent in its prime. It gives the roster a high-risk mood, but when I view each pick on an individual basis, they all make sense. Many of my selections during the back stretch of the draft are similar to the other teams.
I'll be trying out this method in a league or two this year. Stay tuned.
Next Week: Crimes Against Fantasy. The Waldman Commission examines longstanding crimes against fantasy prospects and profiles potential trouble spots for 2015.
-png)