Does the team that scores first in the Superbowl tend to win?Just prior to the 2005 Superbowl, the Wall Street Journal reported that "one key marker" to watch was which team scored first, since the team that scored first had won 26 of 38 Superbowls. That was, just barely, statistically significant at the two-sigma level. But I did not think it was a very powerful predictor (although, scoring being conducive to victory, it would be surprising if it had no predictive power at all). Indeed, although the team scoring first had won 26 out of 38 Superbowls, the team who scores second had won 28 out of 38 Superbowls. So watching for the team who scores second was an even better "key marker" than watching who scores first. In the 2005 and 2006 Superbowls, the team that scored first lost the game, causing the WSJ "key marker" to drop below the magic line of statistical significance at the two-sigma level. In the 2007 Superbowl, the team that scored first lost again! Take that, you Wall Street Journal dunderheads. 2008: In 2008, the team that scored first did win the game. At 27 out of 42 correct predictions, the "who scores first" rule is still just shy of statistical significance. |