This is our old blog. It hasn't been active since 2011. Please see the link above for our current blog or click the logo above to see all of the great data and content on this site.

AFL versus NFL: the Super Bowls

Posted by Jason Lisk on March 24, 2009

I'm going to start my discussion of the AFL and NFL by going where, for most people, the discussion begins and ends. The first four AFL versus NFL championship games (or Super Bowls as they came to be known) took on almost mythic significance. These 240 minutes of game action not only decided titles--they proved how good the two leagues were.

Once the American Football League established itself as having some staying power, the clamor for a championship game began. The December 16, 1963 issue of Sports Illustrated featured an article titled "The Two Pro Football Leagues Must Meet", which set forth an exchange between then AFL Commissioner Joe Foss and NFL Commissioner Pete Rozell. That issue also featured a point/counterpoint, with
Dan Jenkins setting forth the case for the AFL being competitive with the NFL,
and Tex Maule summarily dismissing the AFL with the retort, "Ridiculous! The NFL by 50 Points".

It would be three more seasons before the two leagues would meet in formal competition to test those arguments. By 1966, the AFL was around to stay and had recently signed a new lucrative television contract that would go through the 1969 season. The two leagues were also competing for the top draft picks in bidding wars, driving salaries and hitting the owners where they noticed. This was probably the leading impetus for the merger of the two leagues. So, in June of 1966, the AFL and NFL reached an agreement to merge after the AFL's television contract expired following the 1969 season. They also agreed to have a common player draft starting the next draft in 1967, and to begin playing exhibition games the following season as well. Oh, and the two league champions would meet at a neutral site following the 1966 season.

In Super Bowl I, the NFL Champion Packers defeated the AFL Champion Chiefs 35-10, managing to half Tex Maule's 50 point claim and "proving" that the NFL was far superior to the AFL. In Super Bowl II, the Packers again won, beating the Oakland Raiders 33-14 to affirm that the AFL was still light-years away from competing with the best of the NFL. But then, in Super Bowl III, faster than the speed of light, the New York Jets shocked the establishment by beating one of the most dominant regular season teams of the past three decades in the NFL, stunning the Baltimore Colts 16-7. Then, the AFL evened the score as the Chiefs avenged their earlier Super Bowl loss, and won Super Bowl IV by a score of 23-7 over the Minnesota Vikings.

I’ll go ahead and make some bold claims here. The Chiefs did not improve by 41 points relative to the NFL between 1966 and 1969. Similarly, the AFL did not make a 28 point leap in one season. It’s possible that the AFL did improve some, but there is simply no way that the improvement was as large as looking at the outcome of those four games would suggest. To try to answer how much we can take away from these games, I’m going to look at other “winner take all” games—the NFL championship games from 1933-1969, and the Super Bowl results since 1970.

I used the “simple rating system” regular season ratings for 1970-2008, to figure out an expected score, based on the power ratings, which I could then use to compare to the actual Super Bowl results. For the seasons pre-merger, I do not have the simple rating system rankings calculated, so I used each championship game participant’s regular season point differential (points scored – points against / games played). Then, for each, I calculated the absolute value difference between the actual result and the expected result. For example, according to SRS, the 2007 New England Patriots were +16.8 points better than the New York Giants in the regular season (which is roughly similar to the point spread for that game), and the Patriots lost by 3, for a difference between actual and expected of 19.8 points. I did that for all championship games and Super Bowls, and then found the standard deviation for all of the expected and actual results.

If you want to cut through the math, basically what I’m trying to do here is find out how much variation we should expect when two teams get together in a winner take all setting, using history as a guide. For example, how likely is it that the Chiefs and Packers were roughly equal teams in 1966, knowing that the Packers won the only game played by 25 points?

In The Making of the Super Bowl, Don Weiss wrote about the importance of the Jets' victory in Super Bowl III:

That NFL Championship games were far more competitive than the Super Bowl game was equally undeniable, while the NFL's overall superiority over the AFL was difficult to dispute. Had the Colts defeated the Jets with anything approaching the ease with which the Packers had routed the Chiefs and Raiders in the first two games, the Super Bowl's viability would have come under a dark cloud and prompted serious discussions about how much longer this apparent mismatch could continue . . .

With all due respect to Mr. Weiss (for I do think he was capturing the spirit and attitude at the time), I'm not sure to which NFL Championship games he was referring that were "far more competitive." In fact, the Colts had just destroyed the Cleveland Browns 34-0 in a NFL Championship game far more lopsided than either of the Super Bowls. The standard deviation of the actual versus expected results for all of the pre-merger NFL championship games (1933-1969) was a whopping 22.8 points. The standard deviation for the post-merger Super Bowl games has been a little lower, at 17.6 points. The Bears destroying the Redskins 73-0 back in 1940 probably comes to mind, but even setting that game aside, NFL championship games were full of blowouts and massive upsets, at least judging by the regular season numbers. Consider just these results in the twelve years prior to Super Bowl I, in evaluating how much a single result proved regarding the relative merits of the teams or leagues/divisions that those teams came from:

  • In 1964, Cleveland dismantled the Western Division Champion Baltimore Colts 27-0. Cleveland was +122 in point differential and Baltimore was +203, and Cleveland played in the far weaker division in the regular season, with the East going 4-10 against the West.

  • In 1961, Green Bay destroyed the New York Giants, 37-0. Based on comparative point differentials from the regular season, the Packers were closer to 2-3 points better than the Giants over the course of the 1961 season.
  • In 1959, Baltimore followed up the 1958 victory over the New York Giants with a comfortable 31-16 win in the title rematch.
  • In 1957, the Detroit Lions hammered the Cleveland Browns 59-14. How dominant were the Lions that season? They went 8-4 in the regular season, with 251 points scored and 231 points allowed. Tobin Rote had replaced Bobby Layne at quarterback, leading a late surge and a playoff victory for the Western Division title over San Fransisco a week earlier. Cleveland led the league in point differential.
  • In 1956, the New York Giants beat the Chicago Bears 47-7. Chicago was roughly 4 points better than New York based on the regular season results.
  • In 1955, Cleveland, who was the class of the league in the regular season, comfortably handled the Los Angeles Rams on the road, 38-14.
  • In 1954, Cleveland and Detroit were the two most dominant teams in the league, posting +174 and +148 point differentials over 12 games. The NFL Western Division was the better of the two regular season divisions, going 8-3-1 against the East. Cleveland won the title game, 56-10.

I'm going to talk about the four Super Bowls and break down the teams and the key plays and players in a future post. Looking only at the scores of those four games, though, we don't really know a whole lot. I'm going to split the difference between the pre-merger Championships and the Super Bowls, and say that a reasonable standard deviation for these AFL versus NFL games is 20 points. We had actual score margins of 25, 19, 9, and 16. With the exception of the first game, all of these results would be within one standard deviation even if the opponents were exactly even. Could the Chiefs have been "only" 5 points worse than Green Bay in 1966? Sure. Could the Raiders and Packers have been roughly equal in 1967? Could have been. Might the Colts still have been the far superior team in 1968, after dominating the NFL while the Jets had the third best record in the AFL regular season, to the tune of as much as 11 points? They might. Could the Vikings have been equal to or slightly better than the Chiefs over the course of the 1969 season? Plausible.

I don't know where the actual numbers are going to lead us. From what I've looked at so far (and will be writing about over the coming months) I do have enough to go ahead and say that I think the game point spreads, which were approximately 14, 14, 18 and 12 in favor of the NFL champion in the first four games, were likely overvaluing the NFL champion relative to the AFL. But by how much?

These four games carried great importance for the pride and honor of the advocates and fans of the two leagues, and likely, for the immense popularity that the National Football League and its American Football League descendants enjoy today. What these four games cannot tell you, though, is how good these two leagues were relative to each other, with any kind of specificity. To do that, we are going to have to dig alot deeper. And that's where we are going to start next.