Ex-coaches chide reliance on analytics
The risk-taking in the NFL these days is off the charts. It’s almost impossible to watch a game without second-guessing the seemingly boneheaded decision of a coach to, say, go for it on fourth down from his own territory. Maybe they’ll convert. Maybe not. Just hold your breath.
No, this is not your daddy’s NFL. The analytics might suggest the play has a 60% chance of success – or 80% chance of prompting you to scratch your head.
“This analytics stuff, I don’t care for it,” Bill Parcells, the Hall of Fame coach, told USA TODAY Sports.
You’d figure this would be the Tuna’s position. Never mind that he spends his retirement tending to a stable of nine racehorses. When it comes to gamemanagement in the NFL, he’s proud to consider himself old-school.
“This is a ‘get-off-my-lawn’ guy talking,” Parcells added. “When you start talking about analytics, everybody’s an expert. I saw this happen in boxing when they came up with CompuBox – this guy had so many punches and so many power-punches, so therefore he should win the fight. It made everybody an expert. In reality, there’s more criteria than just punching. Like ring gamesmanship.”
The NFL numbers have an undeniable “CompuBox” flavor when it comes to situational decisions.
Through 13 weeks, teams have attempted to convert on fourth downs and two-point conversions at a record pace, based on NFL data. The 528 fourthdown plays (275 conversions, a 52.1% success rate) are on pace to top the 658 fourth-down plays in 2020 (55% success rate), while the 111 two-point conversion calls – including Baltimore’s failed attempt in the closing seconds of a loss at Pittsburgh on Sunday – are in line to top 131 tries from last season -- the most since the league began officially tracking that category in the early 1980s.
In real life, the numbers reflect cases like Lions coach Dan Campbell going for it on fourth and 1 from the Detroit 28yard line with just over 4 minutes to play against the Vikings last weekend. The Lions were leading 23-21 when Jared Goff was stuffed for no gain on a sneak. The Vikings quickly turned the field position into a touchdown. Campbell wound up getting his first NFL victory when Detroit drove to win on the game’s final play. But it was hardly a victory for analytics.
They almost blew it because of the decision by Campbell that has become almost typical.
“I think a lot of the decisions are made with analytics and percentages, and they forget about their team and their opponent,” Jimmy Johnson, recently minted as a Hall of Fame coach, told USA TODAY Sports.
Coaching in the NFL has always been risky business. It’s just that in today’s game, too many of the decisions – and particularly some choices early in the game – are difficult to justify. Except when adding analytics to the mix.
“Some coaches fall back on analytics too much,” Johnson said. “I feel like the risk-taking has gone to another level.”
Interestingly, during their heyday, both Parcells and Johnson, each of whom won two Super Bowls, were considered on the liberal end of risk-taking. Johnson chuckled when reflecting on fake punts and surprise onside kicks during his Dallas years. And Parcells knows his strategy wasn’t always by the book.
“When I was coaching, I went for it on fourth down quite a bit,” said Parcells, who won two Super Bowls with the Giants and led the Patriots to a Super Bowl. “I wasn’t doing it because of the percentages, I was doing it because I had Jumbo Elliott and Mark Bavaro as blockers. They were pretty dominant. You could count on them.
"But for someone who says you should go for it on fourth-and-one because that’s what the book says, what they don’t know about are the mismatches. You’re a little more reluctant to go for it when Reggie White is over there.”
Both Parcells and Johnson enthusiastically lauded Harbaugh’s decision on Sunday at Pittsburgh for its merits as a calculated risk. After Lamar Jackson’s 6yard TD pass to Sammy Watkins cut the Steelers' lead to 20-19 with 12 seconds to play in the fourth quarter, Harbaugh opted not to tie the game with an extra point from the NFL’s best kicker, Justin Tucker, that likely would have set up overtime.
Harbaugh explained that with star cornerback Marlon Humphrey suffering a season-ending torn pectoral injury during the game, he had little confidence his depleted defense could hold up in overtime. So he played the percentages and went with a two-point try that resulted in Jackson’s incompletion to tight end Mark Andrews in the flat – the flow hurried and pass altered by T.J. Watt’s rush on the quarterback.
“Harbaugh had a good reason,” Johnson said. “His cornerbacks were hurt. He didn’t think he could stop them in overtime. I like his decision. A lot of times, it’s simple: Do you think you can make it?”
Parcells: “The truth of the matter is that he had the play. They couldn’t get it executed. Just by a hair. Game of inches.”
Johnson also loved what he saw from Bill Belichick during New England’s 14-10 win at Buffalo on Monday night. The Patriots coach went for two after scoring the game’s first touchdown, converting with a Brandon Bolden run.
“That changes the strategy for the whole game,” Johnson said, mindful of the wicked weather elements. “He had a one-point lead, then a four-point lead. So it set the stage right off the bat and gave him an edge.”
In the fourth quarter, Belichick went for it on fourth and 1 from Buffalo’s 33, converting with a Mac Jones quarterback sneak that was reminiscent of Tom Brady so often plowing ahead on sneaks. On Monday night, the conversion kept alive the 14-play drive that led to a safer (and successful) field goal from Nick Folk that provided the final margin.
So surely, the second-guessing goes both ways. And when the pivotal calls work, the coach is a genius – or in Belichick’s case, confirmed as more genius.
Then again, there are armchair analysts who would contend that despite the record pace of pivotal calls, coaches still don’t go for broke enough.
That speaks to the heat on Giants coach Joe Judge after the loss at Miami on Sunday when New York had three fourth-down cases near midfield – needing two, three and four yards to convert – and punted each time. Judge explained that with his defense playing well, he opted to win at field position.
It backfired. The Giants lost. And they handed an L to old-school philosophy, too, at least in hindsight.
“Analytics are nothing but statistics,” Parcells said. “I don’t want to sound like I don’t believe in statistics. I used statistics. But I don’t believe the percentages apply in every situation.”
In other words, risky calls are hardly created equally.