Welcome to my first official Round 1 simulation of the 2021 NFL Draft! My analytics-based mock is based solely on a contextual, data-driven model that aims to do one thing: maximize each team’s potential to win as many games as possible in 2021. So, before you read any further, take note:
I am NOT attempting to predict or divine what teams will ACTUALLY DO on draft day.
For this particular file, the model considered current rosters, the overall market of potential free agents and 2021 draft prospects. How, exactly? Well, here’s how my mock works …
I use my draft prospect model, explained at the top of this article, to create a numerical value for each player. These ratings can be compared across years. Then I use my NFL model, which considers the market of potential free agents at each position, to create projected win-contribution metrics by player, position group and side of the ball. These get added up to predict win totals for the season. (Here’s an example of these metrics for WRs.) The results quantify strengths and weaknesses of current NFL rosters. My model also factors in as many known elements of coaching philosophies (of the current staffs) as possible, and each team’s 2021 opponents. Then, my model “selects” the draft prospect that would yield the highest win total for each team in the coming season.
Here’s the part that’s extra: I have projections and results for all teams and the draft prospects they selected over the past 15 seasons. I examine each season’s on-field results, objectively analyzing what happened while identifying the trends and strategies that led to success or failure. I also ask coaches, front office executives and even players to help me understand why results occurred. These subjective inputs help shape the results, meaning the model gets “smarter” each season.
There are a lot of real-life efficiencies that could be realized via draft-pick trades. I can’t help but to note them in certain cases. Still, for the sake of this particular mock, I did not allow for trades. If I worked for an individual team, an analysis like this could aid in creating a strategy for identifying potential trade partners, as well as vulnerable spots where other teams are most likely to scoop up particular players — especially given free agency.
Finally, another change to this year’s mock is a real refinement of how the on-field computer vision weighs on the predictions. Normally, it’s a huge factor, as a prospect’s most recent season-long on-field performance is the most valuable ingredient. This college season was odd (to say the least); the number of games played, when they were played and even situations/contexts faced were quite different from in years past. So I had to look for factors over longer periods of time (and their trajectories) and normalize all past measurements (like we’d get from the combine) for all 15 seasons using computer vision to make sure every comparison was as apples to apples as possible.