When Seattle coach Mike Holmgren joined the San Francisco 49ers as an assistant in 1986, one of his first tasks was to go through the Sunday game films and chart tendencies of the upcoming opponent.
Numbers NFL coaches ignore:
» Passer rating system:
"I think it's the most useless thing in sports, because it leaves too much out, and it doesn't take into account the system that you're running," said Baltimore coach Brian Billick. "I don't know of anybody who uses that as a viable tool as a way of measuring quarterbacks, either by scouts or coaches."
Seattle coach Mike Holmgren agreed, saying, "If you have a really low passer rating, then you're not playing very well, but you don't have to be in the 100s to be playing well at quarterback."
"A quarterback is about winning games," said San Francisco coach Mike Nolan. He uses as examples Brett Favre, John Elway and Jim Kelly. "They could throw the interception, but they were always one (play) better than you," Nolan said.
» Team rankings in rushing and passing: "If people can run the ball on you, you're going to have better pass defense statistics (because they won't bother to throw)," Miami defensive coordinator Dom Capers said.
"If you throw for 500 yards of offense but you still (allow) 10 sacks, I don't know how you can sit there and rave about being, say, third in the league in offense," said Green Bay coach Mike McCarthy. "That can be skewed."
This one is a bit of a surprise. San Francisco's Mike Nolan, a defensive coach by trade, thinks the statistic of sacks is "very overrated." Nolan cites the 2000 Ravens, whose 35 sacks ranked them only 22nd in the league.
"If you're just beating the hell out of quarterbacks, it's a great stat to have," Nolan said. "But I know they have not transferred into wins for as much attention as they get."
And winning, after all, is the most important statistic of all.
-- Ira Miller
Yes, film. That was before the NFL switched to video tape.
That also was before it all became computerized. He did it with pencil and paper.
"I would spend all day Monday tabulating that stuff," Holmgren, now in his 16th year as a head coach, said recently.
It was an early form of compiling all the oddball statistics you often see plastered on your TV screen during telecasts of NFL games.
"Now, with computers, that's zapped out in book form by noon Monday so you can study it -- and use your time in other ways," Holmgren said.
As a result, coaches today have extra time on their hands to study statistics. But which ones do they look at?
Well, you can bet they're not all that concerned about some of the dizzying array of information that has been thrown at the fans already this season ... Tom Brady is the first to throw at least three touchdown passes in each of the first six games of a season ... The Detroit Lions set a record for points scored in the fourth quarter ... Vinny Testaverde came out of retirement and completed his first eight passes ... Devin Hester returned his 10th kick for a touchdown in 25 career games ... Brett Favre set career records for both touchdown passes and interceptions thrown ... Under coach Andy Reid, Philadelphia never has lost a game following a bye week.
Coaches have their own numbers to put under the microscope. They look at turnovers and takeaways, the obvious figures of importance -- but beyond that, they all have individual methods of deciphering statistics and what is important to them.
We surveyed 10 current and former NFL head coaches, and the numbers that concerned them varied widely:
» Yards per pass attempt. Pro football is a passing game, no question, and the proposition is that if you win the passing game, you win the game.
» Field position after kickoffs. Not the yardage on the returns, but the actual starting position of each drive. Call this "hidden yardage."
» Red zone scoring efficiency. We all know what that means -- completing a drive by scoring, preferably a touchdown. Not just scoring in the red zone, but how many points.
» Big, or "explosive," plays. Not all coaches define them the same, but in general, it's a run of at least 12 yards and a pass play of at least 20. The bigger the chunks of yardage a team can gain, the fewer chances there are to mess up on a long drive.
» And, very significantly, tendencies. As in what a team is likely to do on second-and-long or third-and-short ... or, more refined, what's the specific play or pass route that creates success or causes trouble?
Conversely, there are statistics that get discussed every day that many coaches find unreliable and meaningless. Chief among them is the passer rating system, which addresses some passing statistics, but doesn't really measure the overall effectiveness of a quarterback.
Or perhaps you didn't notice that Brett Favre and Brian Griese have comparable career passer ratings.
Here's a closer look:
Yards per pass attempt
This stat is a favorite of at least two coaches who won recent Super Bowls -- Tony Dungy of Indianapolis and the retired Dick Vermeil.
"It doesn't matter how often you throw, if you're throwing and having great success ... that's a determining factor," Dungy said. "There are times in every game when you have to throw the ball, and if you're throwing it efficiently, you're going to win most of the time."
"Yards per pass attempt has one of the most direct correlations to the won-loss record," Vermeil said. "It covers a wide range (pass attempts, completions, sacks, net yardage) and it reflects on a lot of different things."
Dungy, of course, has one of the greatest throwers of all time, Peyton Manning, on his team, and Vermeil coached the Rams when their offense reached record heights in 1999. Dungy said his reliance on the passing game numbers helped keep him and his team going a year ago when critics were harping on the Colts' problems stopping the run on defense.
Indianapolis gave up more than 185 yards rushing in four of its first seven games in 2006. But the Colts won all four. One reason? Their average gain per pass was more than a yard better than the opponents in those four games, nearly two yards better than their opponents over the course of the season.
"We weren't that concerned about (the rushing stats)," Dungy said. "They're trying to keep our offense off the field by running the ball, and they're kicking field goals, and we're scoring touchdowns."
He will take that trade-off every time.
For Dungy, the defining figure is seven yards per pass, and last season, the Colts were the only team in the AFC to average more than seven yards per pass play and the only team in the NFL to average more than 7.5. They averaged 7.53 yards per pass play on offense and 5.79 on defense.
"I know most people look at rushing (yardage) and the number of rushes," Dungy said. "We didn't want to be where we were, but there are times that can be misleading ... In the long run, you're going to have to throw the ball efficiently to win in the NFL. That's just how the game is now."
Field position is one of the canards that Dom Capers used to sell his players when he coached expansion teams at Carolina and Houston, and still preaches as defensive coordinator for the Miami Dolphins. So did Bill Cowher, coach of the 2005 champion Pittsburgh Steelers.
In 1996, Capers' second season with the expansion Panthers, Carolina had the league's best average starting position after receiving kickoffs, and its defense had the second best position in the league after kicking off. Those numbers translated to an 8.6-yard differential for every pair of kickoffs; more than a quarter of the time, Panthers' opponents started inside their own 20 following a kickoff, twice the league average.
Multiply that through an entire season, and you get a lot of so-called hidden yardage.
That wasn't the only reason, but clearly it was one of them, that, in their second season in the league, Carolina compiled a 12-4 record and reached the NFC Championship Game. That same year, the New York Jets had the worst field position differential after kickoffs -- minus-5.7 yards. The Jets finished 1-15.
Cowher, a former special teams coach, understands the concept.
"I didn't look at (kickoff return) average," Cowher said. "I just look at starting position. It's an indication of how long a field are you creating when you kick off, and how short a field are you creating when you return it?"
Red zone scoring efficiency
Last Sunday, the Minnesota Vikings scored 34 points in beating Chicago without once penetrating the Bears' 30-yard line; the Vikings scored four touchdowns on plays of at least 35 yards and got two long field goals. But that kind of cross-country scoring, if not rare, is nonetheless atypical. More than two-thirds of the touchdowns scored in the first six weeks of this season started in the red zone.
Stats that matter
Which makes the winning formula rather clear: To win, a team should score touchdowns when it gets into the red zone and not permit the opponent to score touchdowns there.
"A team can drive the ball and take it down there and (you make them) have to settle for a field goal, that's very deflating for an offense," Cowher said.
For Baltimore coach Brian Billick, red zone scoring efficiency always has been near the top of his list of key stats. He says it's the statistic that is "most statistically relevant to winning and losing."
Billick points to the 2000 season when his Ravens won the Super Bowl with one of the all-time great defensive years in NFL history -- but ranked only No. 2 in the league behind Tennessee, because the league measures rankings by yards and nothing else.
"I'm biased, but I think that was the greatest single season of defense in the history of the game, and we were ranked second in defense," Billick said. "But does anybody argue that wasn't the best defense, certainly that year or the last 10 years? Quite frankly, the NFL rankings as a whole are useless."
In 2000, opponents penetrated the Ravens' red zone 27 times and had only 73 points to show for it, an average of 2.7, or less than a field goal for each drive inside the 20. The next best was Carolina, which allowed an average of 3.2 points on red zone penetrations, but the Panthers were so bad they allowed 61 drives inside their 20, which just might have had something to do with their 7-9 record.
"People are going to get down there, but if you can keep them out of the end zone and make them kick field goals, that four-point differential will be staring you in the face in the fourth quarter of games," said Cowher.
What a team or a player does in certain situations -- what Holmgren worked on as a young assistant -- always has been fascinating, and teams react to it in different ways.
In Denver, Mike Shanahan took the information a step further, establishing a routine that has been copied in Houston by Gary Kubiak, the Broncos' former offensive coordinator: The training camp schedule is dictated by the previous season's situations.
Kubiak explains it this way: "We go back and look at our season, and we look statistically at how much time each phase of the game took, what percentage. Let's say we were on the goal line three percent of the time. Then, three percent of training camp is going to be devoted to the goal line. We reset all our practice schedules to coincide with how much time we were spending on that phase of the game."
Holmgren says he's not much of a stats guru, but earlier in his career, he was. He focuses first on interceptions, something he got from his NFL coaching mentor, the late Bill Walsh, with the San Francisco 49ers in the 1980s. Few things ever bugged Walsh more than a quarterback throwing an interception.
It's no coincidence that Joe Montana and Steve Young, the 49ers' Hall of Fame quarterbacks, were two of the hardest-to-intercept passers in NFL history. But it is the weirdest of coincidence that it's necessary to go to four decimal places to separate their interception percentage. Montana's interception percentage was 2.5783, Young's 2.5789.
"Interceptions come in all degrees and sizes," Holmgren said. "Just the number itself can be misleading at times. That was a big one with Bill. I look at what happened."
In 1989, when Holmgren became offensive coordinator of the 49ers, he decided to study every interception Montana, who already had won three Super Bowls, threw in his entire career. When Holmgren put the information together, he realized that about one-fourth of Montana's interceptions came on a single play.
Montana liked the play, a vertical stretch route down the middle of the field, because it appealed to his gambling instinct, Holmgren recalled. Walsh had liked it, too, but Walsh retired after the 1988 season. So Holmgren went to George Seifert, Walsh's successor, and they agreed to remove the play from the playbook.
As it turned out, Montana had his best season in 1989. He set a league passer rating record, since broken. He threw only eight interceptions, the lowest total of his 49ers' career, and had his highest completion percentage, 70.2. The 49ers rolled to their fourth Super Bowl victory.
San Francisco probably would have won it all that year, even if Holmgren's statistical study did not remove that one play from the offense. But ...
"If you can change the behavior a little bit with arguably the greatest quarterback ever, then numbers can make a difference," Holmgren said.
Statistics, of course, don't always tell the whole story.
In 2001, when New England won its first Super Bowl, the Patriots ranked 19th in the NFL on offense and 24th on defense. They ranked near the bottom of the league in yards per pass attempt and in getting their quarterback sacked. They ranked near the bottom of the league in field position after kickoffs, giving their opponents a 3.1-yard edge. They did rank among the league leaders in red zone defense, but the rest of their numbers were mediocre.
Which only proves that a Tom Brady, like Montana and the other greats, is much more comforting to a coach than a glitzy stat sheet.