Steve Keim is an old-school scout at heart, most comfortable using tape and in-person observation to guide his decisions as the Cardinals’ general manager.
He also knows football is evolving, with data playing a more essential role every season. Keim and team president Michael Bidwill took a notable step in tapping into that market this offseason, creating a three-person analytics team to assist in a wide variety of decision-making.
“Michael and I talked it through, and for two middle-aged guys, I feel we are both forward-thinkers that aren’t so proud to think you can’t learn something new,” Keim said.
While scouting remains Keim’s bread and butter, he wants to know if the analytical research coincides or disagrees with his views.
“When you start out in the scouting business, you don’t want to hear the word analytics because you know (scouting) is something that is an instinctive thing, something you were born to do as a scout, evaluate talent,” Keim said. “Analytics generally scares the old-school mentality because you think it’s put in position to do your job for you, which is not the case. For me, it’s a great tool to either back up the theories you have or to maybe put into question some of your theories and create a checks and balances system.”
Purposeful or not, the Cardinals’ decisions this offseason seem to be in lockstep with analytical philosophies. They hired coach Kliff Kingsbury, who is expected to orchestrate the type of pass-heavy offense favored by the number-crunchers.
Their draft was praised by Pro Football Focus, as the data-based site approved of selections ranging from quarterback Kyler Murray with the No. 1 overall pick to tight end Caleb Wilson as Mr. Irrelevant.
Kingsbury, 40, is one of the youngest coaches in the NFL and known for his innovative offensive ideas. Unsurprisingly, he stays in close contact with analytics and research coordinator Charlie Adkins, assistant Meredith Manley and assistant Kevin Jordan.
“They have a plethora of information every day – really, whatever you want,” Kingsbury said. “However deep into that wormhole you want to dive, you can get there. I’m big on collecting information and trying to make better decisions from it, whether it’s fourth-and-1 calls or what plays to run in the red zone. They are going to have some information that can help you curtail your game plan.”
The NFL is still far behind Major League Baseball, which has undergone a complete analytical revolution. Not only do its coaches and general managers rely heavily on data, but so do the players.
Hitters have changed their swing paths and pitchers have altered their mechanics and pitch usage in the name of sabermetrics.
After an informal poll of Cardinals players, it is clear the NFL is not there yet. Reserve quarterback Charles Kanoff is an avid analytics consumer – regularly reading FiveThirtyEight.com and the musings of Warren Sharp – but is battling to find its practical application.
“There are just so many conflating (variables),” Kanoff said. “It’s not the same every time. A pitcher goes up to the mound and throws one pitch to the same catcher and nothing changes. A football play -- this time (when) the fade is thrown, he was under pressure – there’s a million things that change every single play. Eleven guys are factoring into things, and you also don’t know the assignment.”
While Kanoff is reticent to implement anything now, as sample sizes grow and consensus is reached, he believes data could affect a player’s mindset.
The Princeton product brought up a histogram he saw recently that tracked the performance of throws outside the numbers compared to ones down the seams.
“It was (inefficient to throw to) both sides, and down the middle it was 80 percent complete,” Kanoff said. “It was just interesting, and I was thinking about that. Like, one of the guys was saying, ‘Don’t throw fades,’ and I was thinking about all the fades I’ve thrown in my life, and it was like, ‘Yeah, not a ton of them were complete.’ So, yeah, maybe in the future there will be routes you don’t throw because you didn’t realize how little they were being completed.”
At this stage, Kanoff believes analytics are most beneficial to playcallers. There is evidence teams should throw more on first down, and, while it sounds unconventional, that they should go for two after cutting a 14-point deficit to eight late in games.
“It makes sense,” Kanoff said. “You just literally increase your chance to win.”
Kingsbury is the one in charge of decisions like those -- and will be the one who takes the heat if he bucks tradition and fails. He pledges to minimize that risk by having a clear understanding of the probabilities before each major call.
“You have the final say, but I think having pertinent information, relevant information, is huge,” Kingsbury said. “It’s not just winging it. If they can give you the stats and give you a percentage, then you feel better about your decision.”
Analytics in the NFL are still a bit abstract, but Keim said he has already noticed practical gains. The Cardinals’ number-crunchers have helped him compare player contracts during free agent negotiations and assess the risk-reward of preseason playing time.
It remains to be seen if critical football inefficiencies can be identified through data -- which was the impetus for the famous book-turned-movie ‘Moneyball’ -- but the Cardinals are one of the NFL teams seeking those answers.
“The ability to constantly evolve in every way is the key for us to have success going forward,” Keim said. “It’s not just staying up with the Joneses, but implementing things you truly believe in.”
Images from the Dignity Health Arizona Cardinals Training Center heading into the third preseason game