If one team is leading Sunday's Super Bowl heavily by the end of the third quarter, as often occurs, should you turn off the television?
This is a big concern. After all, Super Bowl parties are at stake. Someone who stops watching the game may even be forced to do chores around the house. How persuasively can the fan say to a wary spouse that the television must remain on in case the losing team makes a thrilling comeback, as the Buffalo Bills did in their playoff game against the Houston Oilers.
We have some good news for the fans and bad news for the spouses. With our colleague Kristina DeNeve, we recently analyzed hundreds of sporting events for Chance, a magazine on statistics. We found that comeback victories are more likely in football than in most other major team sports.
Late comebacks occur about 20 percent of the time in football, basketball and hockey. We found a 23 percent chance that a pro football team losing at the beginning of the fourth quarter will go on to win, and about a 21 percent chance that a basketball team will surge ahead in the final period. In hockey, this occurs 19 percent of the time. In only 6.5 percent of baseball games does the team losing after seven innings win the game.
We analyzed a carefully selected sample of 200 basketball games, 100 baseball games, 100 hockey games and 100 football games. We used the 1990 regular seasons for baseball and football and the 1990-1991 regular seasons for basketball and hockey.
The data showed that football is a good sport for comebacks not only during the final period but throughout the game. In baseball, the team losing after three innings wins only 19 percent of the time. In basketball and hockey, teams overcome first-period deficits about 30 percent of the time. In football, however, the team losing after the first quarter has a 45 percent chance of winning.
Why such a big difference? One explanation is that football is most responsive to mid-game alterations in strategies. Coaches may use the first quarter primarily to test out plays. Another possibility is that first-period scoring opportunities are most uneven in football. One team may have three possessions while the other has only two, a substantial difference. In baseball, by contrast, both teams have made nine outs after three innings. Over the course of the entire football game, this random unevenness becomes less of a factor in the score.
The Super Bowl is played at a pre-determined location, so there is no home-field advantage except by coincidence. During the regular season, football teams playing at home win 58 percent of the time. That's about the same as in hockey (54 percent) or baseball (53 percent), but less than in basketball, where the home team wins 64 percent of the time. Not surprisingly, home teams are much more likely than visitors to make a comeback, especially in basketball. The home team in basketball is at least three times more likely than the visiting team to overcome a deficit at the end of the third period. In the other three sports combined, the home team is only 1.8 times more likely to pull out a last-period victory.
One reason home-team comebacks are most common in basketball may be that sport's rapid scoring pace. In the other three sports, there can be long gaps between scores. The fast scoring in basketball promotes a more constant level of fan reaction, which becomes intense as the game nears an end. Basketball fans deserve their reputations as the home-team "sixth player" during fourth quarters.
Our findings were based on regular season games rather than playoffs. But if you are a sports fan, they do suggest that you should not miss the beginning of a basketball or hockey game, and especially a baseball game. You can tune in a bit late for a football game. And if you are at a basketball game and the local heroes are behind after three quarters, stay in your seat.
Harris Cooper is a professor of psychology at the University of Missouri. Frederick Mosteller, a member of the National Academy of Sciences, is professor emeritus of statistics at Harvard University.