The NFL draft is a complicated process. Teams value players differently on factors such as positional needs, talent evaluation and medical concerns. The possibility of teams trading picks adds a layer of complexity and intrigue as well. This draft simulation attempts to account for all of these factors, calculate the percent chance of teams selecting different players at each pick, and produce the most likely outcome for the first round of the draft.
Determining Player Value
One of the draft factors with the highest level of uncertainty is how a team stacks its big board. Scouting can be subjective and how teams rank a player’s value can vary widely from team to team. To account for this variability, the simulation used five different big boards and randomly assigned them to each team. Four of the big boards were from some of the top-performing mock draft websites (as ranked by The Huddle Report), while the fifth used the 33rd Team’s own big board.
Once a team was assigned a big board, the initial value of each player on their board was quantified using the Fitzgerald-Spielberger Trade Chart, which is similar to the famous Jimmy Johnson Trade Chart, but it instead quantifies the value of draft picks based on average second-contract earnings. For example, if a team had Trevor Lawrence, Penei Sewell and Kyle Pitts as the three top players on their boards, Lawrence would have an initial value of 3,000, Sewell 2,649 and Pitts 2,443. This valuation is an attempt to quantify the larger difference in player value early in the round.
This initial value, based on where a team has the player ranked, was then adjusted using positional value. The value of different positions was quantified by the Average Per Year (APY) Salary of the top 10 earning players at that position. The adjustment was based on how a position’s average salary compared to those of the top paid QBs, as QB is the highest-paid position. For example, the top 10 QB APY average is around $32 million, while the top 10 OT APY average is around $16 million, so the initial value of an OT on a team’s board would be adjusted down by 50%.
Next, team needs come into play. The needs of each team were determined using need tiers provided by DraftTek. Each player’s value to a team was adjusted by how much the team needs a player at their position. To increase the variability of results, team needs could randomly move up or down one tier with each simulation. This added randomness is to account for being unable to precisely determine how a team views its positions of greatest need.
Running the Simulation
The simulation was run 100 times for each draft pick, randomizing the draft boards and adjusting for positional value and team needs each time. The player who was selected the highest number of times through the simulation was the most likely draft pick at that position. Then, the player picked would be taken off the draft boards and the simulation would run again, going on to the next pick. This process was repeated for each of the 32 picks to determine the most likely outcome of the draft.
When it was a team’s turn to pick, they would select the player with the highest value on their board, unless another team successfully executed a trade. A trade would have the possibility of happening if another team had a significantly higher value on a player available than the team currently picking. The probability of the trade going through was based on the amount of first-round draft capital the team attempting to trade up had to offer.
Below is the draft outcome deemed most likely by the simulation. The highlighted teams successfully executed trades.
Not surprisingly, Trevor Lawrence was the first overall pick 100 times out of 100, and Zach Wilson to the Jets was the second-most common pick. The five picks with the highest probabilities based on the model were:
- Trevor Lawrence at No. 1: 100%
- Zach Wilson at No. 2: 85%
- Kyle Pitts at No. 6: 81%
- Justin Fields at No. 3: 72%
- Jaylen Waddle at No. 9: 64%
The trades in this simulation are especially intriguing, as two of the three have QB-needy teams trading up into the top 7 to select QBs. Given the amount of draft capital the 49ers gave up to trade into the third spot before the draft, it would be interesting to see what Atlanta’s asking price would be for the fourth pick.
Additionally, the Buffalo Bills decided to go all in and trade up to select Heisman winner Devonta Smith to replace John Brown, who was lost in free agency. In 78% of simulations run, a team attempted to trade up to 14 to select Smith and the Bills were the most likely team, trading up to this spot 31% of the time. Minnesota would likely get a nice array of picks in return, and still pick up a quality player in Texas EDGE Joseph Ossai at pick No. 30. So far in his tenure as the Vikings’ GM, Rick Speilman has traded down 28 times, while only trading up nine. Meanwhile, Bills’ GM Brandon Beane has so far traded up six times and only traded down once.
This simulation attempts to mimic the way teams make draft decisions while baking in the inherent randomness in the process. The model’s predictions will be updated in real-time during the draft as teams deviate from the expected draft. Be sure to follow The 33rd Team on Twitter — @The33rdTeamFB — for live updates.