Will the performance of new NFL athletes be worth the price?

By Laura Stassi

Jason Merrick, Ph.D., and his colleagues initially developed a statistical modeling method as a college admissions tool, but at the request of a certain NFL team, it’s been repurposed. The team — Merrick’s not allowed to name names — is utilizing the model to more accurately predict whether money it spends in the draft on the most desirable amateur football players will pay off after they go pro.

“You see all these teams that are trading multiple draft picks away to get a top draft pick,” said Merrick, professor of supply chain management and analytics in the Virginia Commonwealth University School of Business. “But our model shows the risk of chasing those players.”

This year’s draft is April 23–25. During seven rounds of picks, the 32 NFL teams choose players to add to their respective rosters. Teams can give away or trade their turns in the draft but typically in the first round, teams with the worst records the previous season choose first, in ascending order. That means this year in the first round, the 3-14 Las Vegas Raiders get first choice, and the Super Bowl-winning Seattle Seahawks go last.

Scouts for the teams already have evaluated the prospects on traits including prowess, perceived fortitude and college game-time performance. With those ratings, “team management determines an overall rating to decide which players they’ll go after and which of their precious draft picks to use on them,” Merrick said. The earlier a player is chosen, the higher his pay.

To develop the model, “the NFL team gave us 14 years of rating data on every player they scouted, not just those who were drafted or who played for this specific team,” Merrick said. “So we’ve got multiple predictions that we aggregate into a rating that we then chart against percent of cap for a second contract.”

“A player’s first contract is based on draft position, so it’s a prediction: how good a team thinks you’ll be,” Merrick said. “But your second contract is based on your market value to the team, so it’s an outcome.”

Percent of cap is the portion of the team’s salary budget devoted to each player. Merrick noted it is difficult to determine how good a player has been because “you’d use different metrics” to measure the performance of a quarterback, for example, than for a defensive lineman.

“So our way isn’t necessarily a perfect way,” Merrick said, “but it’s the least bad way of doing it.”

Merrick said his model shows that “maybe you don’t go for the highest-rated player in the draft because overspending on them doesn’t necessarily lead to a winning outcome. We see that the high-rated picks performed a bit better on average. But you could end up with nothing from that expensive choice.”

Merrick and his colleagues’ model is detailed in the Wharton School research paper “Decision-making with Ordinal Ratings.” It also has been presented at academic and technology conferences, Merrick said, adding that he’s working on revisions for upcoming publication in a top scientific journal.

Meanwhile, Merrick has his own prediction. If the unnamed NFL team sticks with its draft strategy, “it will trade down,” he said, “recognizing that the players rated in the middle of the pack have a very decent chance of being as good as those rated at the very top. So if I can trade down and get three picks in a later round instead of one pick in the top round, I have a better chance of a better outcome with one of the three than with that one player.”

This story was originally published on the School of Business website.