Villanova ignores analytics for the biggest play of the year | Mike Jensen
Fourth-and-1 on their own 29, the Wildcats didn't even look at the analytics book. They went for it, made it, and changed their season.
Mark Ferrante, Villanova’s head football coach, knew what his players wanted to do. His offensive guys yelled over pretty much what you’d expect them to yell.
“I was telling Coach, ‘Go for it,’” said tailback Justin Covington. “ ‘Let’s do this.’”
Going for it was not a no-brainer. You could make the case either way and have it sound all sorts of sensible. Villanova had fourth-and-1 at its own 29-yard line, 116 seconds left, up 28-27 at James Madison, the third-ranked team in FCS. End the game with a first down. Miss the conversion, however, and it would put JMU practically in game-winning field-goal territory.
» READ MORE: Trey Patterson will help Villanova hoops
Sure, the JMU kicker had missed a couple of field goals. That made Ferrante more certain Ethan Ratke wouldn’t miss another one, since he’s the all-time FCS leader in career field goals, and hadn’t missed since 2019 before this game.
“He’s an All-American,” Ferrante said.
Also, Ferrante said he had great faith in his own punter, not pointing out he was a fill-in, a backup QB taking care of things with Villanova’s star punter out hurt. Wasn’t that playing with fire a little bit?
“He’s done a great job for us,” Ferrante said of the fill-in, Connor Watkins, who had punted twice in the game, for 50 and 31 yards.
Most college programs have analytics offering suggestions for such situations. What did Villanova’s paper say?
“It’s not a paper, it’s a book, up in the booth,” Ferrante said the next day, explaining that he made a point of not asking what that book said, nor was such a suggestion offered.
“It’s a tool,” Ferrante said.
His own head placed a bet on his offense to get the yard. Nothing fancy. Get the ball to Covington, one of the great Villanova backs of its FCS era. The kind of guy who routinely turns a half-yard into a yard, a yard into 2.
Even at the snap, the drama was baked in.
“They looked like they had a pretty good defense called,” Ferrante said.
Covington got the handoff from QB Daniel Smith and the first JMU would-be tackler, a safety, blitzed through untouched and had a shot at Covington in the backfield. A cleaner angle, or getting there just a step sooner, and the play would have been a loss. Instead, Covington barely noticed the touch at his ankles. The next touch came at the line of scrimmage, also low, but cleaner. Covington, a sixth-year player, had his momentum going. He got the yard. Maybe.
“I kind of spun,” Covington said the next day over the phone. “Because I was moving my feet, I reached the ball out.”
Then, he said, he knew he had it. His coach said Covington had it anyway, pre-reach. That was an insurance yard, officially a 2-yard gain. Game over. A huge win, but not a crazy upset. The week before, Ferrante had run through all the league statistics, how these two programs were so close, leading all the relevant statistics.
As for the analytics, Ferrante made the relevant point — a tool, not a crutch. Fourth-down go-for-it analytics was a discussion point at the weekly college football luncheon earlier this season after Penn had faced an interesting scenario in its season opener at Bucknell. The home team had fourth-and-1 at its own 27, first series of the game. Bucknell went for it, was stopped. The Quakers got a quick field goal in what turned out to be a 30-6 victory.
Bucknell, as an underdog, probably knew it had to take chances, even if that one didn’t work out. That’s the thing about going for it. The book obviously is only as good as everything put into it.
“The analytic group — I sidelined them this year,’’ said Penn coach Ray Priore. “I put them on the sideline. What they don’t know is you haven’t played in two years.”
The Ivy League didn’t play any 2020-21 football.
“They don’t know your team,” Priore said. “A guy who runs it through a program someplace, in Whereverville, I get it, I get the analytical part of it. In 2019, I did use it. They told me to go for it. We didn’t make it a couple of times. I probably wouldn’t have done it except the piece of paper said go for it. Somebody whose job wasn’t on the line was listening to some computer that spit out some information.”
Priore is a veteran coach, but no Neanderthal. He’s won Ivy titles as a head coach and as a defensive coordinator. He’ll get back to looking at the numbers.
“I think it’s a good tool, I really do,’’ Priore said, and he knows the analytics offer percentages, not assurances.
“We all know, you have to go where your strengths are,” Priore said, outlining how when he had a veteran star QB throwing to a future NFL receiver, taking a chance throwing wasn’t really much of a chance at all. “But a first-time QB behind a young offensive line, different.”
Good analytics factor all that in. This season, Priore was saying, is so different the percentages could be impacted more than normal.
“It’s all predicated on how you feel your team is playing at that point,” Priore said. “And that’s the part that analytics doesn’t always add up. How is your energy? How was your practice week? What are your injuries?”
“I always take chances early in the season,” said Rowan coach Jay Accorsi. “Then when it doesn’t happen, I take less chances. I go back to the players, ‘OK, you wanted to go for it, I went for it, you didn’t get it.’ "
Penn student Zach Drapkin has worked closely with analytics for both basketball and football. He was an analytics intern this summer with the Baltimore Ravens.
“Usually, these decisions are made with one of two types of models: Expected points or win probability,” Drapkin explained in a message. “Expected points [EP] is possession-level and mainly just deals with down, distance, yard line, and sometimes time on the clock. … Each play is assigned an expected point value.”
Win probability, Drapkin said, tends to consider other factors like team strengths “and whatever other information will improve the model’s accuracy.”
“Honestly, I’ve gone for things on fourth down forever as a head coach,” said West Chester coach Bill Zwaan, who became the school’s all-time winningest coach this season. “I’ve always been an aggressive, forward-leaning head coach. But I also understand that early in the game, to not put our defense in that situation, because it can flip the game quickly.”
Zwaan said he has “an idea of what the percentages are,” but always goes by feel.
“There are some quarterbacks I’d do everything with,” Zwaan said. “There are other quarterbacks I’d do …” He paused for a moment. “... nothing with.”
Saturday, Ferrante knew he had a sixth-year college QB handing off to a sixth-year college running back. He knew all the variables. When it all worked out, it all looked smart and gutsy and heroic, and Villanova, with just the one loss at Penn State, had risen to sixth nationally in FCS.
So did Ferrante ask later what the book said to do?
“I didn’t ask,” Ferrante said. “But they told me.”
And?
“The book said punt.”