COLUMN

The three rounds that disappeared

13 November 1982, Las Vegas. The world lightweight fight between Ray Mancini and Duk-koo Kim ended in the fourteenth round. Kim died five days later.

The following month the WBC announced that world title fights would be cut from fifteen rounds to twelve, and put it into effect on 1 January 1983. The WBA followed in October 1987, the IBF in June 1988.

Three rounds disappeared. What changed?

First, the wrong answer

Split the record at the change and compare, and this is what you get.

PeriodBoutsInside the distance
Through 198836,67747.2%
From 198958,63657.9%

Every bout, split at 1988

The fights got shorter and the finishes went up. It looks like an interesting result. It is wrong.

Of the 36,677 bouts on the left, 16,504 are from before 1960. That era finishes at 35.5%, and its records were kept differently from today's. It is simply a lot of old data in one half, and it has nothing to do with the rules. Restrict both sides to 1960 onward and the gap falls from 10.7 points to 1.1 points.

The right comparison

Only world title fights went to twelve rounds. Everything else was already scheduled for ten or fewer. So if you take how world title fights changed and subtract how every other fight changed over the same years, the era cancels out.

Then match the distance as well. A ten-round bout in the 1970s and a ten-round bout in the 1990s are the same length of fight.

GroupPeriodInsideDecisionMean rd
Title, 15 rounds1973-198258.5% (593)40.8%8.21
Title, 12 rounds1989-199849.5% (1,460)49.7%6.17
Non-title, 10 rounds1973-198251.7% (3,362)47.7%4.59
Non-title, 10 rounds1989-199854.6% (3,967)44.7%4.23

Title fights scheduled for 15 rounds, and non-title fights scheduled for 10 (bout counts in brackets)

World title fights finished 9.0 points less often, while over the same years non-title fights finished 2.8 points more often. The difference is 11.8 points, and that is the estimate of the part the rule can be held responsible for. Four different ways of drawing the comparison all land between -11.5 and -16.0, and none of them changes the sign.

So the answer was the other way round. The fights got shorter and the finishes went down.

When rounds 13 and later went away

YearStopped in 13-15Per 1,000 bouts
19781215.2
1980910.2
198198.7
198298.6
198300.0
198477.4
198687.7
198754.8
198821.9
1989-199900.0

Bouts stopped in rounds 13, 14 or 15 (decisions excluded)

The order the bodies moved in is visible in the numbers. The WBC went to twelve in 1983; WBA and IBF fifteen-rounders carried on from 1984 to 1987; by 1988 they had nearly gone; and from 1989 there are eleven straight years of zero. The yearly counts are single digits and should not be read one year at a time, but smoothed over five-year blocks they fall 10.98 → 3.12 → 0.00.

Round twelve became a different round

It was not only rounds 13 to 15 that went. Round twelve itself changed.

Round15-round era (1973-1982)12-round era (1989-1998)
1-11 total42.5%47.5%
125.2%2.1%
13-15 total10.8%0.0%

Share of bouts ending in that round (title fights, eras matched)

Inside a fifteen-round fight, round twelve is one more round a little past the middle, and 5.2% of bouts ended there. Inside a twelve-round fight it is the last round, and the figure is 2.1% — less than half. Non-title ten-rounders show nothing of the kind.

The same numbered round is a different thing depending on whether three rounds are still to come.

Where the lost finishes went

Counted per hundred bouts, 13.9 finishes from round twelve onward disappeared. They went here:

11.9 more bouts went to a decision (after subtracting the change in non-title fights).

2.2 more bouts were finished in round eleven or earlier (same subtraction).

Almost all of it is decisions. Fights that would have been stopped went to the scorecards standing.

What did not change

How likely a fight was to end in rounds two to eleven did not move. The differences scatter around a point either way and do not keep a sign. The idea that twelve rounds made fighters push harder early is not visible in this data.

Whether the better fighter got harder to beat is also unresolved. In world title fights the higher-rated fighter won 3.4 points more often after the change; in non-title fights the figure is 5.0 points. Subtract and the sign is slightly the other way, and the rating-gap bands do not agree with each other.

This question did not get an answer.

The fights that would have gone to the cards

The world title fights that ended in round thirteen or later, in the fifteen-round era, have names.

1 October 1975, Muhammad Ali vs Joe Frazier. Frazier's corner retired him at the end of the fourteenth. Under twelve rounds, a decision.
16 September 1981, Sugar Ray Leonard vs Thomas Hearns. TKO in round fourteen.
30 November 1979, Sugar Ray Leonard vs Wilfred Benítez. TKO in round fifteen, with six seconds left.
12 November 1982, Aaron Pryor vs Alexis Argüello. TKO in round fourteen — the day before the Kim fight.
11 June 1982, Larry Holmes vs Gerry Cooney. TKO in round thirteen.

All five would have ended on the scorecards under twelve rounds.

What these numbers cannot say

It cannot be said that twelve rounds caused the drop. What can be said is that across the change, world title fights fell 12 points relative to non-title fights. Other things that happened to world title fights alone over the same years — television, where a referee draws the line, the splitting of the weight classes — are not separated out here. And Kim's death, which is the direct reason for the twelve-round rule, must also have changed other safety practice at the same time. Those cannot be told apart in principle.

Nothing at all can be said about the years before 1960. There is not one world title marker in that era, so the framework of this article does not reach it.

A fight recorded as not being a world title fight may still have been one. The marker is read out of the notes column, and it catches 78% of them in the 1970s and 50% in the 1960s.

The source is the record tables on Wikipedia. They are edited thickly for world title fights and thinly for everything else. That the records themselves are made differently for the two groups is a bias no check here removed.

The figures in this article are based on data as of 21 September 2026. The data is updated as it comes in, so the last digit may differ from the current tally.