Big D By Three

Twenty-one weeks from now, on the 21st day of January, the defending champion Dallas Cowboys will beat the hometown Miami Dolphins by three points in Super Bowl XIII. So says computer statistician Bud Goode. Or rather, so says the Univac 1106 computer that Goode (pronounced Goody) has fed with relevant data. The computer has "played" the entire 1978 National Football League schedule, thus providing SI a sneak preview of the season (see charts beginning on page 48).

The major revelation of this forecast is that NFL '78, in sharp contrast to such sequels as Airport '77 and Jaws 2, may actually turn out to be better than its predecessor. The reason is not merely a schedule expanded from 14 to 16 regular-season games per team but a new format that matches strong teams against strong teams and weak against weak. Last year 82 NFL games, less than 42% of the total, were decided by a touchdown or less. This year Goode's computer predicts that a similar margin will settle 144 games, or more than 64%. "You cannot overemphasize the importance of the new scheduling philosophy," says Goode. "An average team could easily qualify for the playoffs, particularly since there will be an extra wild-card team from each conference."

Close races should be the rule in the AFC in each of its three divisions. Miami, New England and Baltimore in the East; Pittsburgh, Houston and Cincinnati in the Central; and Oakland, Denver and San Diego in the West are all playoff contenders. In the NFC, Dallas appears to have only one real challenger: Los Angeles.

This season 14 NFL coaches will have Goode on the payroll for his weekly analysis; his track record is simply too good to ignore. In 1972 he correctly forecast that the Dolphins would become NFL champions with a perfect 17-0 record. Last season he predicted that underdog Denver would make it to the Super Bowl, only to lose to Dallas.

For the 54-year-old Goode, a native of Los Angeles, football statistics are an avocation turned vocation (SI, Jan. 14, 1974). He has applied a sophisticated computer process known as Multiple Regression Analysis to the myriad data available. Essentially, Goode's computer breaks the game down into its component parts. Take two parts hydrogen and one part oxygen, put them together with a bang and you get water. Take 44 rushes, only half as many passes, don't get intercepted or sacked, and you get victory. It isn't quite that simple, of course, but Goode's numbers provide a blueprint for winning football. This is what he sells coaches, not the predictions the computer spins off from its analyses.

And just what sort of attack does the computer recommend? Well, unfortunately for the three television networks and everyone else who thinks of pro football as show biz—and what else can you consider a business that promotes the bump and grind while outlawing the bump and run?—the computer and Woody Hayes think very much alike. According to the Univac 1106, the way to make the most of your chances of being a winner is to run the ball as often as possible.

Skeptics might argue that the winningest team of the '70s, the Oakland Raiders, has the game's best passing attack in Quarterback Ken Stabler and such capable receivers as Fred Biletnikoff, Cliff Branch and Dave Casper. For runners the Raiders have always had guys from Colgate. Yet the fact is that last year Oakland led the league in rushes. The Raiders ran the ball almost 49 times a game, about 10% more than their closest rival, Los Angeles. Meanwhile, 18 teams threw the ball more than Oakland in 1977.

What's more, the computer says you don't have to run the ball well, you just have to run it. Thirteen teams gained more per rush than the Raiders last year. For their part, the Atlanta Falcons conducted a sort of controlled experiment in this theory. Under new Coach Leeman Bennett, a Goode client, the lowly Falcons appeared to have no offense. After six games Atlanta ranked last in the league in yards per rush. However, because Quarterback Steve Bartkowski was injured, the Falcons stayed on the ground. And what did this do for their record? Atlanta was the surprise of the NFL, leading its division at 4-2.

One reason Goode's computer rates the rush so highly is that running the ball helps a team's defense almost as much as its offense. "The run in football is the closest thing to the stall in basketball," Goode explains. Just the thing for prime-time television. Once again the Falcons illustrate the point. Last year the Falcons set an NFL record for fewest points allowed in a modern 14-game season—129. On the basis of that record they can be considered the top defensive team in the history of the game. Yet can anyone recall the names of three members of the Falcon defense? If you could ask Goode's computer to name Atlanta's defensive star, it would probably answer, "An injured Steve Bartkowski."

If a team insists on putting the ball in the air, Goode's computer offers three other criteria that can separate winners from losers. They are: yards per pass attempt, sacks and interceptions. In running, quantity is more important than quality, but the reverse holds true for passing. Teams that had 40 or more rushes won 85% of the time in 1977, while no team won a game in which it threw 40 times or more. As for total passing yardage, the team that threw for the most yards last year, Buffalo, won three games. The team that threw for the fewest, Tampa Bay, won two. The key passing statistic is yards per attempt; the six teams that led the league last year—Dallas, Baltimore, Oakland, New England, Pittsburgh and St. Louis—averaged 10 wins.