How to predict a Super Bowl winner

[PonyExpress]

How to predict a SB winner

Offseason musings...

Part I.
Treating 1978 as Year 1, when the 16 game schedule, the Wild Card Playoff System, and the easing of the passing rules came into full effect.

These rules work only when the two teams playing in the SB have already been set, i.e. after the AFC and NFC Championship games are over.

Rule 1: the team that allowed the fewest points overall during the regular season will win, except in the case of Rules 2 & 3. Success rate: 16/16. 100%

Rule 2: if a team trying to repeat as champions scored more points overall during the regular season than its SB opponent, it will win, fewer points it will lose. The exception to Rule 2 is Rule 3. (5/5) Success rate: 100%

Rule 3: If the teams already met in the regular season, and the game was decided by a TD or less, the team that lost will win the SB rematch. Exception: Dan Reeves’ Broncos in 1986 (arguably the worst SB coach of all time). (6/7) Success rate: 83%

Overall success rate: 27/28 or 96%.

There have been 8 outright victories by the underdog, as established by Las Vegas, since 1978: Oakland in 1980, Washington in 1982, Oakland in 1983, Washington in 1987, NY in 1990, Denver in 1997, NE in 2001, and TB in 2002. Those three rules would have predicted all 8.

Part II.
How a SB winner could have been predicted between the merger in 1970 and the advent of the 16 game schedule in 1978, the Wild Card, and the easing of the passing rules (assuming the two opponents have already been set).

Rule 1: the team that scored the most points during the regular season will win. Exception: Pittsburgh 1974 (arguably the greatest defense of all time). Success rate: (7/8) 88%

Between 1970 and 1978, there were two outright victories by underdogs established by Las Vegas: Miami in 1972 and Baltimore in 1970. Rule 1 would have predicted both.

 

[patchick]

Pony, I usually groove on your data crunching but I have to take exception with this one. Seems to me this isn't "how to predict a Super Bowl winner" but "how to describe Super Bowl winners post hoc." Some stat is bound to correlate the highest...how many different measures did you look at?

 

[PonyExpress]

Pony, I usually groove on your data crunching but I have to take exception with this one. Seems to me this isn't "how to predict a Super Bowl winner" but "how to describe Super Bowl winners post hoc." Some stat is bound to correlate the highest...how many different measures did you look at?

Good point. I may have overstated my case. However, using the three rules the winner could have been predicted prior to the game itself simply by observing very simple regular season data... which is why I found the observations interesting.

 

[Coast2CoastPatsFan]

I take it you're referring to straight up SB winners as opposed to Against the Spread (ATS)?

 

[PonyExpress]

The three rules represent commons sense truths; for instance (1) a good team with a superior defense will beat a good team with a superior offense on a neutral field with a title on the line; (2) A good team with a superior defense can win one championship, but it takes a great team to win back to back titles, and that great team needs a superior offense to carry it through its second title run; if the O fails, so will the team. (3) A team that loses a close game has an emotional edge in a rematch on a neutral field with a championship on the line. Those rules I mentioned in the opening post are just an expression of common sense, which happens to have proved correct 27 out of 28 times on Super Bowl Sunday...

 

[PonyExpress]

I take it you're referring to straight up SB winners as opposed to Against the Spread (ATS)?

Exactly. But the betting angle comes in when you consider the predictions of upsets. I found it interesting that the eight upsets since '78 seem to flow from those three rules. But I'm definitely making too much of it. It's the offseason after all.

 

[8man]

In the rule below:

Rule 3: If the teams already met in the regular season, and the game was decided by a TD or less, the team that lost will win the SB rematch. Exception: Dan Reeves’ Broncos in 1986 (arguably the worst SB coach of all time). (6/7) Success rate: 83%

When you say "a TD or less" is that 6 points of 7 points?

 

[PonyExpress]

In the rule below:

Rule 3: If the teams already met in the regular season, and the game was decided by a TD or less, the team that lost will win the SB rematch. Exception: Dan Reeves’ Broncos in 1986 (arguably the worst SB coach of all time). (6/7) Success rate: 83%

When you say "a TD or less" is that 6 points of 7 points?

7 points or less.