2024 NFL Mock Draft: Kent Weyrauch’s “The Model” Picks (4.0)

This mock will be slightly outside the norm for mock drafts you see during the NFL Draft cycle. Earlier this year, I set out to compile pre-draft visits for NFL prospects so that I could create visual representations of teams and prospects. During that, I realized that draft visits contained some predictive power regarding NFL Draft position. One thing led to another and soon enough I found myself creating a machine learning model using several well-known inputs, in addition to my draft visit data.

This mock will be mostly model-driven, with slight shifts due to team fit. There will also be no trades. Enjoy!

Kent Weyrauch has been in the top 10% of experts for the FantasyPros Mock Draft Accuracy Contest over the last three years.

“The Model” Explainer

This model was built using a Quantile Random Forest methodology. The quick and dirty way to explain why this model works well for the NFL Draft is that the draft is chaotic. Usually by round three, the expected draft position of most prospects goes out the window and NFL teams make (seemingly) unexpected picks and trades.

This mock will be mostly model-driven, with slight shifts due to team fit. There will also be no trades. Enjoy!

Kent Weyrauch has been in the top 10% of experts for the FantasyPros Mock Draft Accuracy Contest over the last three years.

“The Model” Explainer

Quantile Random Forest models capture rapidly expanding variance pretty well by scaling the quantiles as they progress through the data. It’s also very receptive to log-form data, which is how I transformed the pick number for the model. Think like the Fitzgerald-Spielberger NFL Draft Trade Value Chart, which makes the difference between pick 1 and pick 2 larger than the difference between pick 2 and 3, etc.

To train the model, I used draft visit data from Walter Football and Pro Football Network, Lance Zierlein’s prospect grades, NFL NextGen Stats’ grades, age, and consensus big board data from Mock Draft Database — which combines mock draft and big board data in a very clean way. With this, I found the following feature importance given these variables:

(Note to those who saw the original tweet, the model has been changed since its inception and may continue to change as needed)

This structure makes a lot of sense from a face-validity perspective. This model essentially has three arms of determination:

Wisdom of the crowds through consensus mock draft and big board data
Player evaluation through film (Zierlein) and production/analytics (NextGen) grading
Team/player sentiment via draft visit data

2023 NFL Draft Results

Tie that all together and boom, we’re cooking up a model, baby! I used this with 2016-2022 data to “predict” the 2023 draft. This was the result:

Legend:

Green names are first-round projections that went in the first round
Red names are first-round projections that fell out of the first round
Blue names are the players that went in the 1st round that were not first-round projections
The “ridges” show the distribution of likely draft position
Dots show their actual draft position

(Note: the original tweeted version of this graph contained errors and was deleted.)

To look at this another way, here is the actual pick order with our model laid over it.

Without even lifting a finger our model has correctly guessed five player-to-spot predictions and correctly identified 27/31 first-round picks, which would be a respectable number given we aren’t trying to identify player-to-team matches. Granted, the mock draft data accounts for this to a decent degree.

2024 NFL Draft Predictions

Below are the predicted ranges for the 2024 NFL Draft prospects:

My final mock draft will include this information combined with several other non-quantifiable factors. For now, let’s use this data to create a mock draft powered by these rankings.

2024 NFL Mock Draft (Round 1)

PICK	Team	Player	Pos	College
1	Chicago Bears	Caleb Williams	QB	USC
2	Washington Commanders	Drake Maye	QB	North Carolina
3	New England Patriots	Jayden Daniels	QB	LSU
4	Arizona Cardinals	Marvin Harrison Jr.	WR	Ohio State
5	Los Angeles Chargers	Malik Nabers	WR	LSU
6	New York Giants	Rome Odunze	WR	Washington
7	Tennessee Titans	Joe Alt	OT	Notre Dame
8	Atlanta Falcons	Terrion Arnold	CB	Alabama
9	Chicago Bears	Dallas Turner	EDGE	Alabama
10	New York Jets	Brock Bowers	TE	Georgia
11	Minnesota Vikings	J.J. McCarthy	QB	Michigan
12	Denver Broncos	Quinyon Mitchell	CB	Toledo
13	Las Vegas Raiders	Taliese Fuaga	OT	Oregon State
14	New Orleans Saints	Olumuyiwa Fashanu	OT	Penn State
15	Indianapolis Colts	Jared Verse	EDGE	Florida State
16	Seattle Seahawks	Byron Murphy II	DT	Texas
17	Jacksonville Jaguars	Cooper DeJean	CB	Iowa
18	Cincinnati Bengals	JC Latham	OT	Alabama
19	Los Angeles Rams	Laiatu Latu	EDGE	UCLA
20	Pittsburgh Steelers	Jackson Powers-Johnson	C	Oregon
21	Miami Dolphins	Johnny Newton	DT	Illinois
22	Philadelphia Eagles	Kool-Aid McKinstry	CB	Alabama
23	Minnesota Vikings	Nate Wiggins	CB	Clemson
24	Dallas Cowboys	Graham Barton	OT	Duke
25	Green Bay Packers	Troy Fautanu	OT	Washington
26	Tampa Bay Buccaneers	Chop Robinson	EDGE	Penn State
27	Arizona Cardinals	Amarius Mims	OT	Georgia
28	Buffalo Bills	Adonai Mitchell	WR	Texas
29	Detroit Lions	Brian Thomas Jr.	WR	LSU
30	Baltimore Ravens	Tyler Guyton	OT	Oklahoma
31	San Francisco 49ers	Jordan Morgan	OT	Arizona
32	Kansas City Chiefs	Xavier Worthy	WR	Texas