Machine Learning Madden NFL: The best player position switches for Madden 17

A couple weeks ago, I wrote about my initial efforts toward using machine learning to model the “master equations” that govern the Madden NFL player ratings system. This week, I’d like to put those models to use to compute the best player position switches for Madden 17.

Sometimes NFL players are assigned to play a particular position for their team even though they’re better suited to play at another position. Perhaps their team had a glaring weakness at that position, or there was already a better player filling their preferred position.

In Madden 17, these position mismatches cause some players to have an embarrassingly low overall player rating (OVR) even though they’re star players waiting for their time to shine. I’m going to highlight a handful of those players below, but I’ve also uploaded a spreadsheet of Madden 17 player ratings at all positions so you can sort through them yourself.

The >95 OVR club

Perhaps the worst position mismatch in the Madden 17 preseason teams is Khalil Mack:



Image courtesy of Wikipedia

Technically, Khalil Mack should’ve been a part of Madden 17’s coveted 99 OVR club, but since he started the season at left outside linebacker (LOLB), he was assigned a 94 OVR. A wise Madden coach will quickly move Mack over to the left end (LE) position, where Madden will rate him at 99 OVR and Mack will rain sacks on QBs all season. Perhaps it makes sense to put Mack at LOLB in a 3-4 defense, but don’t put his pass rushing talents to waste as a linebacker in a 3-4 defense.

Here’s all of the Madden 17 players that could’ve made the >95 OVR club:

Player Team Position OVR Recommended Position New OVR
Khalil Mack Raiders LOLB 94 LE 99 (+5)
Derrick Johnson Chiefs MLB 90 SS 98 (+8)
Pernell McPhee Bears LOLB 92 DT 98 (+6)
Brandon Marshall Broncos MLB 90 SS 96 (+6)
Sean Lee Cowboys ROLB 89 SS 96 (+7)
Thomas Davis Sr. Panthers LOLB 88 SS 96 (+8)

Biggest gainers

Below is a list of players who improve in OVR the most simply by changing positions. Notably, Lee Smith can jump from a mediocre Madden tight end (TE) to an above average fullback (FB) thanks to his exceptional blocking skills.

Player Team Position OVR Recommended Position New OVR
Zak DeOssie Giants TE 55 ROLB 75 (+20)
Lee Smith Raiders TE 75 FB 87 (+12)
Christian Kirksey Browns MLB 74 SS 85 (+11)
Jonathan Anderson Bears MLB 66 SS 77 (+11)
Cordarrelle Patterson Vikings WR 70 HB 81 (+11)
Deone Bucannon Cardinals MLB 78 SS 88 (+10)

Best out-of-position kickers

And just for fun—and because the NFL had a bit of a kicker SNAFU this year—below is a list of the non-kickers who could become the best kickers if necessity called. Funny enough, according to Madden, Ryan Tannehill (the Dolphins’ starting QB) would make almost as good a kicker as he would a QB.

Player Team Position OVR Recommended Position New OVR
Ryan Tannehill Dolphins QB 79 K 76
Tim Hightower Saints HB 73 K 75
Odell Beckham Jr. Giants WR 93 K 74
Mohamed Sanu Sr. Falcons WR 77 K 67
Dontrelle Inman Chargers WR 72 K 67
Ben Roethlisberger Steelers QB 93 K 66

Feel free to look through the full spreadsheet and report your findings in the comments. There’s far more useful position switches than what I reported here.

Dr. Randy Olson is a Senior Data Scientist at the University of Pennsylvania, where he develops state-of-the-art machine learning algorithms with a focus on biomedical applications.

Posted in analysis, machine learning Tagged with: , , ,