Winning the Flex; aka: Score Points, What’s Your Plan?

Shawn’s Siegele’s Zero RB article written last year may not have been the first anti-RB-early article ever written, but it certainly crystallized an idea and made it understandable in a way that people appreciate.

Some part of Shawn’s article argues against the idea of Value Based Drafting that most people understand, while Shawn has also separately advocated a theory of fantasy football that stresses winning the flex.

Those articles, along with contributions on our site which promoted a TE-TE draft strategy, sent me into my database to look at the historical data and see which draft plan history prefers.

The analysis that follows makes three assumptions that I’ll lay out so you can judge their prudence. They are:

  1. Most leagues now have flexible enough settings that if you wanted to, you could just pretend as if there’s no difference between RB, WR, and TE. You can start two RBs in most leagues, but you can also start two TEs in most leagues if you want to. Of course you could also start three RBs in most leagues, but you would only do it if you had a third RB that could outscore the second TE you have. And then the same discussion applies to WR as well. You need to start two but you could also start three if you had enough high scorers.
  2. In light of that, you could almost forget about positions and draft as if they’re all one. You’ll just draft points and then you’ll figure out later who to start.
  3. Average Draft Position is a reasonable approximation of where players will finish the season in scoring.

Based on those three assumptions I looked at the average points scored for each ranking at each position for the past three seasons. So I calculated the average points scored for RB1, TE1, WR1, RB2, etc. Pretty much every positional spot.

Then I applied those averages to current ADP in order to get a dummy projection for each player. So LeSean McCoy is currently going as RB1 in ADP and he gets the dummy projection of the average of the RB1 finishers for 2011-2013.

When I match ADP to those averages as described above, I get a projection table like this:

RB 1 367.50 McCoy, LeSean PHI RB
WR 1 342.73 Johnson, Calvin DET WR
WR 2 328.10 Thomas, Demaryius DEN WR
RB 2 327.83 Charles, Jamaal KCC RB
RB 3 315.53 Peterson, Adrian MIN RB
WR 3 300.07 Bryant, Dez DAL WR
WR 4 296.97 Green, A.J. CIN WR
RB 4 295.43 Forte, Matt CHI RB
WR 5 292.70 Jones, Julio ATL WR
TE 1 290.53 Graham, Jimmy NOS TE

That table is sorted based on the average point finish for each positional rank. So Jimmy Graham is being drafted as TE1, and he gets the average points of a TE1 finisher over the past three seasons.

Then you can actually take that information and compare it to today’s ADP in order to get relative values.

1 RB 367.50 McCoy, LeSean PHI RB 1 1 0
1 WR 342.73 Johnson, Calvin DET WR 5 2 3
2 WR 328.10 Thomas, Demaryius DEN WR 6 3 3
2 RB 327.83 Charles, Jamaal KCC RB 2 4 -2
3 RB 315.53 Peterson, Adrian MIN RB 3 5 -2
3 WR 300.07 Bryant, Dez DAL WR 9 6 3
4 WR 296.97 Green, A.J. CIN WR 10 7 3
4 RB 295.43 Forte, Matt CHI RB 4 8 -4
5 WR 292.70 Jones, Julio ATL WR 11 9 2
1 TE 290.53 Graham, Jimmy NOS TE 7 10 -3
6 WR 288.80 Marshall, Brandon CHI WR 13 11 2
7 WR 283.63 Nelson, Jordy GBP WR 17 12 5
5 RB 274.20 Lacy, Eddie GBP RB 8 13 -5
8 WR 272.70 Brown, Antonio PIT WR 18 14 4
9 WR 265.40 Jeffery, Alshon CHI WR 19 15 4
10 WR 263.70 Cobb, Randall GBP WR 23 16 7
11 WR 260.67 Allen, Keenan SDC WR 28 17 11
6 RB 256.10 Murray, DeMarco DAL RB 12 18 -6

It might help a little if I explain how to read the table. Calvin Johnson is being drafted as the fifth overall player off the board. But the top WR at the end of the season has scored on average 342 fantasy points, which would be 2nd overall among the flex eligible positions. So this suggests that he’s being under-drafted by three spots (VALUE). Then Matt Forte is going off the board on average at four overall. But the RB4 at the end of the season averages the eighth most fantasy points among the flex eligible positions. So he’s being over-drafted by about four spots.

Then you can do that for every position of ADP, which you can check out in this Google doc.

However, this information might be easier visualized in graph form. Here’s a graph which shows the VALUE for each positional ADP spot based on current ADP.

PPR Scoring – Positional ADP vs. VALUE


The red line on the graph represents neutral VALUE. Dots above the line mean that based on current ADP, that player is being drafted later than would be justified based on the scoring of the position rank that the player is assigned to. You can actually see that after RB1, every single RB is negative. You can also see that after TE1, TE2 and TE3, all of the TE spots are positive. Meanwhile, WRs are undervalued at each stop along the positional ranking scale. In the WR15 range they have a lot of overlooked value compared to ADP.

If you’re with me, then you’re getting to the logical conclusion of this exercise, which is that following this method would mean you wouldn’t ever draft RB because there would always be a more undervalued WR or TE to draft. This is of course pretty close to what Shawn Siegele has actually done to win high stakes leagues.

The other takeaway from this exercise is that the TE-TE start promoted in a few articles on this site might be preferable to RB-RB, but it isn’t on par with WR-WR. If you picked Jimmy Graham and Rob Gronkowski at their current ADP you would have –3 VALUE for Graham and then neutral value for Gronk. Going with RBs at those spots by picking Eddie Lacy and Arian Foster would yield a total of –13 in value. But picking Dez Bryant and Randall Cobb would be +10. Again, these are just based on historical averages for the positional finishes. I wouldn’t actually draft Randall Cobb if Jordy Nelson is healthy. I’m just talking expected value based on historical averages. The one thing that would change this exercise is if the TE premium scoring is added in.

TE Premium Scoring – Positional ADP vs. VALUE


With TE premium scoring you can see that the top TEs are positive values. Most WRs remain relative positive values as well, although the WR30 range looks closer to par.

There are two things that are probably worth addressing at this point. First, there’s the issue of my Positional Ranking matched to ADP methodology. I’m using end of season ranks, and beginning of season ADP. Should that matter? It only matters if drafters have an increased ability to predict one position, like RB, over another. So if drafters nail their RB picks, while WR is more of a crapshoot, then it would actually make sense that they overvalue RBs at the expense of WRs.

The other thing that’s worth addressing at this point is that you can’t just not pick RBs. I mean you could, but it doesn’t seem like a great plan. Shawn Siegele’s response has been to pursue an anti-fragile plan where his team benefits from chaos. That seems like as good of an idea as any. But somewhat related to the same issue is that it’s perhaps the case that the shape of RB usage dictates the draft strategy that we’ve all always used (RB early and often). Early running backs are picked because those are the ones that will see usage. Most teams have only one RB on the field at any given time, while they employ lots of WRs, so waiting on WR doesn’t carry the same penalty. This is actually an issue that you can investigate with the Snake Draft Planner.

The Snake Draft Planner will optimize a team for you. You can pre-select and blacklist players, but the Snake Draft Planner will suggest an optimal draft plan based on scoring settings and your draft slot. If I input PPR scoring with just one flex and only ask for a starting lineup, this is the lineup it gives me picking out of the 5 spot:

Round Overall Conf. Based ADP Player POS Proj. Pts
1 5 6 Demaryius Thomas WR 315.16
2 20 21 Alshon Jeffery WR 269.05
3 29 33 Shane Vereen RB 224.99
4 44 44 Michael Crabtree WR 241.61
5 53 56 Ray Rice RB 216.14
6 68 77 Greg Olsen TE 189.87
7 77 90 Colin Kaepernick QB 311.79

If I change to TE premium scoring with a double flex it changes the lineup as follows:

Round Overall Conf. Based ADP Player POS Proj. Pts
1 5 6 Jimmy Graham TE 326.98
2 20 26 Julius Thomas TE 273.48
3 29 35 Andre Johnson WR 255.51
4 44 44 Michael Crabtree WR 241.61
5 53 56 Ray Rice RB 216.14
6 68 77 Greg Olsen TE 226.32
7 77 90 Colin Kaepernick QB 311.79
8 92 106 Knowshon Moreno RB 163.10

I would always want to mix the suggestions of the Snake Draft Planner1 with my own thoughts on player targets (the lineups above are just for illustration purposes). So don’t take the results of the Planner as gospel. But it’s a great way to get suggestions as to maximizing your draft. It’s also great for trying to come up with a “shape” for your draft. In the cases above I think the Planner realizes that the early RBs are perhaps overvalued in a PPR format, but then also deals with the reality that you need some RBs, so it suggests mid-round RB values.


I have three primary takeaways from this exercise:

  1. If you play in a league with roster flexibility and PPR scoring, you really shouldn’t be afraid of hammering WR early. In a double flex plus three WR league if other drafters want to choose RBs then you can probably really punish them.
  2. Even though drafters have somewhat accounted for recent scoring trend changes, there might still be room for exploitation and RBs might still be overvalued.
  3. TE-TE is an interesting plan but probably works best in a TE premium format. At least that’s the case when I look at the very most favorable TE stretch of the last 14 years. Is it possible that either Gronk or Graham could end up outproducing the competing WR options in their areas of the draft? Sure, anything is possible. It’s also possible that the injuries that allegedly prevented Jimmy Graham from dominating in 2012 could manifest again. Or that the foot injury that allegedly led to about a 40% drop in his fantasy scoring at the end of 2013 could re-appear. I would argue that current ADP is aggressive as it relates to the relative value of Graham and Gronk and probably properly prices in the risk that each offer (primarily the risk related to Gronk playing a partial or injury hampered season). But if anyone disagrees with me on that, there’s certainly room for reasonable disagreement. Based on historical averages, there’s not actually a way to go TE-TE without overvaluing one of the TEs. You’ll either take Graham before he’s a value (TE1 is 10th in flex scoring over the past three years) or you’ll select Gronk before he’s a value (TE2 is 21st in flex scoring over the past three years).

Because I can anticipate some questions related to my definition of “historical” I’ll say that in addition to running this exercise with 2011-2013 data, I also tried other seasonal splits like 2000-2013 and 2012-2013. The only thing that really changed values very much was using 2000-2013, in which case RBs are more appropriately valued by current ADP. The big RB seasons of the last decade end up impacting things. If you think football trends are headed back in that direction then this article probably won’t help you very much.

Subscribe for a constant stream of league-beating articles available only with a Premium Pass.

  1. which uses Fantasy Pros projections  (back)
By RotoViz Staff | @rotoviz | Archive

No Comment

Leave a reply