Updated Sep 5, 2026

The RAD Ten.

The places our formula disagrees most with expert consensus. Superflex, half PPR.

Ranked by log value gap, the absolute value of ln(consensus rank) minus ln(our rank), rather than raw rank difference. A 25 spot gap at pick 2 is a much bigger claim than a 25 spot gap at pick 120, and raw difference cannot tell them apart.

The structural finding

Consensus top 24

11 QB, 6 RB, 7 WR

Our top 24

1 QB, 12 RB, 9 WR, 2 TE

Consensus top 36

15 QB, 11 RB, 9 WR, 1 TE

Our top 36

3 QB, 16 RB, 13 WR, 4 TE

This is not ten disagreements. It is one disagreement, priced ten times.

The fades

Where we pay less.

Where the market pays up for the arm.

Wolves

#28

Consensus

#2

−26

The gap between him and the quarterback you can take four rounds later is smaller than the gap between the back you give up and that back's replacement. Nothing here says he is bad. It says the price is a first round pick and the surplus is not.

02

Wolves

#48

Consensus

#4

−44

44 spots, and the biggest fade among the consensus top five. Our production input rewards realized weekly finishes rather than what a healthy season projects, and the opportunity input cannot separate him from the six arms behind him on volume.

03

Wolves

#12

Consensus

#1

−11

The smallest fade on the list and the only quarterback we hold anywhere near the top. He is the one arm whose floor is worth an early pick, because he is almost always top three even in the years he is not first. 12 is still a long way from 1.

04

Wolves

#31

Consensus

#3

−28

3 overall is the price of a settled top tier arm. He is a live dart, which is a real thing to own at a price that reflects the risk. 3 does not reflect the risk.

05

Wolves

#37

Consensus

#5

−32

From roughly QB7 through QB17 the weekly line goes flat, and the market cannot tell where inside that flat he lands. Pricing him 5 assumes it can.

Wolves

#50

Consensus

#13

−37

37 spots the other way. Consensus is buying the arm talent and the offense it is attached to. Our production input only credits realized weekly finishes, and there are not enough of them yet to price him inside the first 13 picks.

The raises

Where we pay more.

06

Wolves

#1

Consensus

#6

+5

The thesis in one player. A three down back who is also the goal line back is the rarest combination we track, and running back replacement level in half PPR is brutal in a way quarterback replacement level is not. Miss on him and you start RB25. Miss on the quarterback going at 1.02 and you start the quarterback going four rounds later, which costs about five points a week.

Wolves

#2

Consensus

#8

+6

Only 6 spots on paper, which is exactly why we rank by log value gap instead of raw difference. Moving a player from 8 to 2 is a far bigger claim than moving one from 60 to 90, and raw difference cannot see that. Same engine as Gibbs. A three down workload with the goal line work attached, measured against a running back replacement level that half PPR does nothing to soften.

09

Wolves

#11

Consensus

#41

+30

30 spots, and it is a scarcity read rather than a player read. We measure every player against replacement at his own position, and the drop from a top tier tight end to a streamer is the steepest cliff in the format. The market prices tight ends against the whole board. We price them against the tight ends you would otherwise actually be starting.

The one we'd most like broken

Burrow. If consensus is right about anything on this list, that is where we would put our money. Tell us why on X.