The problem with mock drafts, as i see it, is that there isn't really any rigor applied to them; just some guy picking who he thinks a team needs. But what are those needs? And how does that impact the process?
Well, I decided to look into it myself. What I did here is take PFF player grades and create snap-weighted unit grades for each team: sum of grade x snaps. I then normalized the scores into percentiles for each team:
TEAM | QB | RB | WR | OT | TE | OG | C | CB | S | LB | IDL | Edge |
ARZ | 0.96 | 0.35 | 0.97 | 0.50 | 0.53 | 0.52 | 0.24 | 0.95 | 0.30 | 0.37 | 0.69 | 0.29 |
ATL | 0.74 | 0.51 | 0.32 | 0.86 | 0.32 | 0.76 | 0.47 | 0.42 | 0.22 | 0.26 | 0.34 | 0.22 |
BAL | 0.65 | 0.13 | 0.70 | 0.74 | 0.52 | 0.37 | 0.52 | 0.03 | 0.92 | 0.33 | 0.58 | 0.31 |
BUF | 0.53 | 0.54 | 0.26 | 0.50 | 0.27 | 0.50 | 0.76 | 0.64 | 0.31 | 0.34 | 0.36 | 0.28 |
CAR | 0.89 | 0.31 | 0.42 | 0.66 | 0.93 | 0.91 | 0.84 | 0.69 | 0.69 | 0.92 | 0.29 | 0.79 |
CHI | 0.36 | 0.77 | 0.43 | 0.28 | 0.70 | 0.11 | 0.98 | 0.32 | 0.28 | 0.11 | 0.26 | 0.32 |
CIN | 0.70 | 0.82 | 0.62 | 0.57 | 0.82 | 0.83 | 0.21 | 0.60 | 0.74 | 0.75 | 0.62 | 0.17 |
CLE | 0.13 | 0.58 | 0.19 | 0.92 | 0.43 | 0.42 | 0.70 | 0.49 | 0.02 | 0.51 | 0.62 | 0.11 |
DAL | 0.01 | 0.04 | 0.19 | 0.80 | 0.54 | 0.59 | 0.82 | 0.01 | 0.66 | 0.78 | 0.07 | 0.37 |
DEN | 0.41 | 0.49 | 0.79 | 0.13 | 0.92 | 0.73 | 0.74 | 0.85 | 0.84 | 0.76 | 0.97 | 0.84 |
DET | 0.33 | 0.37 | 0.71 | 0.21 | 0.07 | 0.20 | 0.61 | 0.53 | 0.36 | 0.82 | 0.09 | 0.49 |
GB | 0.91 | 0.47 | 0.95 | 0.89 | 0.26 | 0.97 | 0.57 | 0.90 | 0.97 | 0.19 | 0.32 | 0.92 |
HST | 0.32 | 0.60 | 0.91 | 0.95 | 0.20 | 0.33 | 0.74 | 0.65 | 0.57 | 0.16 | 0.81 | 0.55 |
IND | 0.13 | 0.22 | 0.62 | 0.09 | 0.58 | 0.99 | 0.09 | 0.13 | 0.39 | 0.51 | 0.73 | 0.52 |
JAX | 0.30 | 0.74 | 0.61 | 0.18 | 0.55 | 0.21 | 0.60 | 0.62 | 0.08 | 0.88 | 0.20 | 0.46 |
KC | 0.55 | 0.77 | 0.40 | 0.12 | 0.67 | 0.55 | 0.89 | 0.98 | 0.76 | 0.56 | 0.91 | 0.78 |
MIA | 0.35 | 0.47 | 0.59 | 0.26 | 0.20 | 0.03 | 0.57 | 0.47 | 0.61 | 0.76 | 0.33 | 1.00 |
MIN | 0.52 | 0.59 | 0.13 | 0.11 | 0.70 | 0.53 | 0.83 | 0.81 | 0.36 | 0.84 | 0.34 | 0.83 |
NE | 0.95 | 0.85 | 0.28 | 0.29 | 0.98 | 0.58 | 0.72 | 0.80 | 0.98 | 0.89 | 0.27 | 0.83 |
NO | 0.88 | 0.85 | 0.58 | 0.95 | 0.79 | 0.34 | 0.74 | 0.28 | 0.31 | 0.33 | 0.05 | 0.27 |
NYG | 0.33 | 0.39 | 0.66 | 0.57 | 0.15 | 0.14 | 0.78 | 0.68 | 0.19 | 0.47 | 0.18 | 0.23 |
NYJ | 0.18 | 0.86 | 0.95 | 0.08 | 0.00 | 0.54 | 0.62 | 0.31 | 0.30 | 0.27 | 0.89 | 0.04 |
OAK | 0.59 | 0.05 | 0.28 | 0.95 | 0.35 | 0.30 | 0.50 | 0.30 | 0.22 | 0.18 | 0.63 | 0.96 |
PHI | 0.75 | 0.35 | 0.09 | 0.62 | 0.77 | 0.46 | 0.81 | 0.18 | 0.99 | 0.21 | 0.97 | 0.44 |
PIT | 0.87 | 0.99 | 0.93 | 0.62 | 0.65 | 0.91 | 0.16 | 0.71 | 0.80 | 0.22 | 0.87 | 0.62 |
SD | 0.70 | 0.55 | 0.35 | 0.47 | 0.67 | 0.14 | 0.06 | 0.30 | 0.07 | 0.04 | 0.10 | 0.37 |
SEA | 0.89 | 0.80 | 0.86 | 0.31 | 0.68 | 0.14 | 0.39 | 0.27 | 0.61 | 1.00 | 0.05 | 0.94 |
SF | 0.07 | 0.13 | 0.09 | 0.16 | 0.17 | 0.78 | 0.01 | 0.80 | 0.32 | 0.29 | 0.84 | 0.19 |
STL | 0.06 | 0.07 | 0.03 | 0.11 | 0.11 | 0.21 | 0.22 | 0.84 | 0.48 | 0.74 | 0.80 | 0.18 |
TB | 0.41 | 0.97 | 0.40 | 0.16 | 0.26 | 0.73 | 0.21 | 0.06 | 0.28 | 0.53 | 0.24 | 0.08 |
TEN | 0.10 | 0.08 | 0.02 | 0.72 | 0.95 | 0.05 | 0.17 | 0.10 | 0.46 | 0.40 | 0.81 | 0.31 |
WAS | 0.58 | 0.17 | 0.62 | 0.92 | 0.41 | 0.90 | 0.01 | 0.44 | 0.31 | 0.07 | 0.70 | 0.65 |
Couple caveats here: first, these are normalized percentiles, so a 0.99975 is in fact a whole standard deviation better than a 0.955 - this works fine for our purposes. Second, major injuries aren't factored in -- e.g. Kelvin Benjamin coming back, Jordy Nelson coming back, Tony Romo, etc. There aren't that many of these so its not a major factor...but the need to be taken into consideration.
Mock Draft
So, let's do a mock draft based on this. What I'd like to do is a combination of best player available and needs-based drafting strategies. First, in order to wash out some of the extreme scores, I'll just use the square root of each of these - that way, a 0.01 is just a 0.1. Then, I'll just divide draft grades by this score - I'm using NFL.com for this exercise. Let's start with TEN.
Player | POS | GRADE | TEN |
Laremy Tunsil | OT | 7.7 | 10.7 |
Myles Jack | ILB | 7.5 | 18.8 |
DeForest Buckner | DE | 7.3 | 23.4 |
Ezekiel Elliott | RB | 7.2 | 90.1 |
Jaylon Smith | ILB | 7.1 | 17.8 |
Jalen Ramsey | CB | 6.8 | 69.0 |
Joey Bosa | DE | 6.7 | 21.4 |
Jared Goff | QB | 6.7 | 70.3 |
Andrew Billings | NT | 6.7 | 8.3 |
Vernon Hargreaves | CB | 6.7 | 68.0 |
Ronnie Stanley | OT | 6.7 | 9.3 |
A'Shawn Robinson | DT | 6.5 | 8.0 |
Leonard Floyd | OLB | 6.5 | 20.8 |
Reggie Ragland | ILB | 6.5 | 16.3 |
Hunter Henry | TE | 6.5 | 6.8 |
Paxton Lynch | QB | 6.4 | 67.2 |
Cody Whitehair | OG | 6.4 | 141.6 |
Jarran Reed | DT | 6.4 | 7.9 |
Jack Conklin | OT | 6.4 | 8.9 |
Kevin Dodd | DE | 6.3 | 20.2 |
Robert Nkemdiche | DT | 6.3 | 7.8 |
Derrick Henry | RB | 6.3 | 78.9 |
Laquon Treadwell | WR | 6.2 | 256.5 |
Kyler Fackrell | OLB | 6.2 | 19.8 |
Mackensie Alexander | CB | 6.2 | 62.9 |
As is shown, nothing makes a ton of sense for TEN at #1 - so they probably take a discount to trade down. Taylor Lewan receives fairly excellent grades from PFF, and was drafted high - so all these people picking Tunsil there in drafts are probably way off. They have a franchise left tackle. CB won't go #1, nor will a RB. Maybe Treadwell is #1 overall pick materiel after the combine. They're invested heavily in edge positions lately -- maybe Buckner as a tackle. Lets go with that. The rest of the first 37 picks is as follows:
Pick | Team | PLAYER | Grade | Pick Fit | ||
1 | TEN | DeForest Buckner | DE | Oregon | 7.3 | 13.1 |
2 | CLE | Jared Goff | QB | California | 6.7 | 18.9 |
3 | SD | Myles Jack | ILB | UCLA | 7.5 | 35.9 |
4 | DAL | Jalen Ramsey | CB | Florida St. | 6.8 | 59.2 |
5 | JAX | Laremy Tunsil | OT | Mississippi | 7.7 | 18.2 |
6 | BAL | Vernon Hargreaves | CB | Florida | 6.7 | 38.9 |
7 | SF | Laquon Treadwell | WR | Mississippi | 6.2 | 20.8 |
8 | MIA | Ezekiel Elliott | RB | Ohio St. | 7.2 | 10.5 |
9 | TB | Joey Bosa | DE | Ohio St. | 6.7 | 24.0 |
10 | NYG | Andrew Billings | NT | Baylor | 6.7 | 15.8 |
11 | CHI | Jaylon Smith | ILB | Notre Dame | 7.1 | 21.7 |
12 | NO | A'Shawn Robinson | DT | Alabama | 6.5 | 28.1 |
13 | PHI | Reggie Ragland | ILB | Alabama | 6.5 | 14.2 |
14 | OAK | Derrick Henry | RB | Alabama | 6.3 | 29.0 |
15 | STL | Ronnie Stanley | OT | Notre Dame | 6.7 | 20.5 |
16 | DET | Jarran Reed | DT | Alabama | 6.4 | 21.9 |
17 | ATL | Leonard Floyd | OLB | Georgia | 6.5 | 13.8 |
18 | IND | Jack Conklin | OT | Michigan St. | 6.4 | 21.4 |
19 | BUF | Kevin Dodd | DE | Clemson | 6.3 | 11.8 |
20 | NYJ | Hunter Henry | TE | Arkansas | 6.5 | 131.9 |
21 | WAS | Scooby Wright III | ILB | Arizona | 6.1 | 23.1 |
22 | HST | Paxton Lynch | QB | Memphis | 6.4 | 11.2 |
23 | MIN | Robert Nkemdiche | DT | Mississippi | 6.3 | 10.8 |
24 | CIN | Shaq Lawson | DE | Clemson | 6.2 | 14.8 |
25 | PIT | Kyler Fackrell | OLB | Utah St. | 6.2 | 7.9 |
26 | SEA | Cody Whitehair | OG | Kansas St. | 6.4 | 17.0 |
27 | GB | Vernon Butler | DT | Louisiana Tech | 6.1 | 10.7 |
28 | KC | Le'Raven Clark | OT | Texas Tech | 6.0 | 17.4 |
29 | ARZ | Emmanuel Ogbah | DE | Oklahoma St. | 6.1 | 11.4 |
30 | DEN | Carson Wentz | QB | North Dakota St. | 6.1 | 9.5 |
31 | CAR | Mackensie Alexander | CB | Clemson | 6.2 | 7.5 |
32 | CLE | Corey Coleman | WR | Baylor | 6.1 | 14.0 |
33 | TEN | Josh Doctson | WR | TCU | 5.8 | 37.3 |
34 | DAL | Connor Cook | QB | Michigan St. | 5.9 | 60.6 |
35 | SD | Jonathan Bullard | DT | Florida | 6.0 | 19.0 |
36 | BAL | Jordan Jenkins | OLB | Georgia | 5.9 | 10.6 |
37 | SF | Ryan Kelly | C | Alabama | 5.8 | 48.5 |
I had teams pass on RB Elliott because RB just doesn't make sense yet -- although after the combine that could change. CLE had a near-tie between Bosa and Goff. For SF, Elliott came up as the top score, but since Hyde is coming back, that will change -- but it wouldn't be out of the question to take him! near tie between Lynch and Treadwell. I don't know if Lynch would be a fit - and I don't think Chip is going to commit to a top QB pick this year - he wants to give Kap a chance.
Note: The lower the pick fit, the more likely it is that team is looking to trade down. These scores will change quite a bit after free agency.
Appreciate thoughts and feedback