(11/19/98)
Was Denver's victory in Super Bowl XXXII really an upset?

I tried to stay away from any Super Bowl preview analysis during the weeks leading up to the Super Bowl last year, because I was afraid my lack of objectivity regarding the Broncos would get in the way. Now that we are more than halfway into the following season, I will look back at the Broncos Super Bowl victory. Some people have said that this was an upset only topped by the N.Y. Jets defeat of the Baltimore Colts led by Joe Namath. It halted what was an apparetnly endless streak of NFC victories in the Super Bowl. Denver was also the wild card, a position from which hadn't been won since Oakland beat Philadelphia in 1980. Denver had to win road games at Kansas City and Pittsburgh (and both of them barely) just to get into that position. It all added up to a Pakcers rout and Las Vegas agreed giving listing the Pakcers as two touchdown favorites. However, statistical analysts (as opposed to the general football analysts) actually had other ideas. ESPNet's Rob Neyer ran a column (which I don't have the link for any more unfortunately) showing that by using adjusted yards per pass attempt (the basis of the QB ratings on this page) that Elway had a great year and that few teams with a QB with that good of a rating were as bad as everybody was saying. There were some advantages for both teams, but he picked a close game and couldn't see the blowout. Bud Goode, another statistical analyst showed that Denver should be a 2-3 point favorite according to his computers. After adding some subjective modifications, he came up with a Green Bay team favored by 9. But why did he put in the subjective modifications? It was because of John Elway's historically poor performances in the Super Bowl. However, using a simple system such as the pythagorean winning percentage used on my pages, Denver should have been signifcant underdogs all three times Elway had led them to the Super Bowl (yes, that includes the Redskin game). So what did such a simple stat say about last year's Super Bowl? Denver's pythagorean record was actually 12-3-1 (vs. an actual record of 12-4) while Green Bay's pythagorean record was 11-4-1 (vs. an actual record of 13-3). Advantage Denver. Of course, it's not quite that simple. Green Bay had played a harder schedule (at leat in terms of # of playoff teams) and Denver had recently lost 34-17 to San Francisco, a team Green Bay handled in the NFC title game. These were the reasons Neyer gave for favoring Green Bay, despite the minor statistical advantage that Denver apparently had. In terms of Modified Drive Value, Denver's offense was rated at 32.4 vs. Green Bay's 27.7. Denver's defense was rated at 21.5, virtually equivalent to Green Bay's 21.4. Again, advantage Denver. In straight up yards/drive Denver actually had the advantage over Green Bay in both offense and defense. Using these statistical methods, it actually pointed to the Broncos as the favorites. Due to the schedule advantage, one can understand the Packers being favored. But the only reason they were favored by so much was the streak. Maybe the streak didn't mean as much as some simple statistics.

Back to the football analysis page