Match-fixing in professional sumo

Match-fixing in professional sumo is an allegation that has plagued professional sumo for decades. Due to the amount of money changing hands depending on rank, and prize money, there had been reports of yaochō (八百長^?) (corruption, bout-fixing) in professional sumo for years before it was finally definitively proven to exist in 2011. The hierarchical structure of the sport, in which a minority of top-ranked wrestlers have great advantages in salary, privileges and status over the lower-ranked wrestlers that make up the majority of sumo participants, may have contributed to the use of match-fixing in order to prolong careers for top-ranked wrestlers and assist in the distribution of promotions.

In The Joy of Sumo: A Fan's Notes (Charles E. Tuttle, 1991), David Benjamin determined that over a span of ten basho in the years 1989 and 1990, wrestlers who entered the final day with records of 7–7 emerged with an unlikely winning ratio of .813 (39–9). He offered further evidence of cheating in sumo, but his analysis of Day 15, "The Last-Day Blues", represents the first effort to demonstrate statistically that rikishi almost certainly trade favors late in sumo tournaments to facilitate the all-important goal of having a majority of wins in a tournament (8-7 or better, called kachi-koshi), which guarantees a wrestler will not be demoted the next tournament.

In 2002, Steven Levitt and Mark Duggan replicated and expanded upon Benjamin's research, although not crediting The Joy of Sumo. They published a paper using econometrics in order to suggest that corruption in sumo exists. Popularized in Levitt's book Freakonomics, the study found that 70% of wrestlers with 7–7 records on the final day of the tournament (i.e., seven wins and seven losses, and one fight to go) won. The percentage was found to rise the more times the two wrestlers had met, and decrease when the wrestler was due to retire. The study found that the 7–7 wrestlers won around 80% of the time when statistics suggest they had a probability of winning only 48.7% of the time against their opponents. Like Benjamin, the authors concluded that those who already have 8 wins collude with those who are 7–7 and let them win, since the 8-win wrestlers had already secured their ranking.

...
Wikipedia