Eighty-seven finals: how the championship game changed

The championship game is the one thing this archive has never missed. All 87 of them, 1939 to 2026, both scores, every year, each one cross-checked against the champion of that tournament. Almost everything else about the old tournament is patchy — and that turns out to be the interesting part.

Eighty-seven finals

One game a year, the same game every year, played under whatever the rules happened to be at the time. That makes it the cleanest measure of how the sport changed that this archive can offer: no sampling, no missing rounds, nothing selected.

Championship games by decade. Combined score is both teams added together; margin is the median winning margin.
Decade	Combined	Median margin	Finals
1940s	89	10	10
1950s	141	11	10
1960s	144	14	10
1970s	151	10	10
1980s	136	2	10
1990s	150	8	10
2000s	149	10	10
2010s	133	6	10
2020s	138	10	6

The finals of the 1940s average 88 combined points. The finals of the 1970s average 151. A title game finishing 46–33, as the 1939 one did, is not a slower version of a modern final. It is a different sport wearing the same name.

Then there is the 1980s, where the median winning margin in a championship game is 2 points. Ten finals, and the middle one decided by a single possession. No other decade in this archive is close.

What changed in the mid-eighties

Two rules arrived within about a year of each other: a shot clock, and then the three-point line for the 1986–87 season. Either side of them, across every game the archive holds in a stretch where its coverage is essentially complete:

1980–1986: 134 combined points a game, across 370 games
1987–1993: 152 combined points a game, across 444 games

A shift of about +18 points a game. This archive cannot tell you how much of that belongs to the clock and how much to the arc — they land too close together to separate, and none of this is a controlled experiment. It can tell you the shift is real, and large.

Twenty-five champions, and one program

25 different schools won a title in the 46 tournaments before 1985. UCLA won 10 of them. Each of those is cross-checked the same way: the champion recorded for a tournament matches the winner of that year's title game, for all 87 seasons.

UCLA’s full record is here, and every other program’s is here.

What this archive cannot tell you

This is the honest part, and it is the reason this page is built on the final rather than on the whole tournament.

The early rounds are incomplete. The archive holds three games for 1941 and thirteen for 1956. Those are gaps in transcription, not small tournaments. So nothing on this page counts teams, counts games, or says how big a field was in a given year. Those numbers would look precise and be wrong.

Overtime is not recorded before 1987. The transcription matched a literal “OT”, and the older bracket pages mark overtime with a bare asterisk under a footnote instead. Around seventy overtime games carry no flag. No overtime figure is computed here for the early era, and a blank in that column means “not recorded”, never “finished in regulation”.

Seeds. This archive carries no numeric seed before 1979, and one 1978 entry holds a non-numeric value. Everything on what a seed line is actually worth therefore begins in 1985, where the modern bracket does.

What is solid: champions, runners-up, title-game scores, and the score and winner of every game that is present. That is what this page is made of.

A number that looks precise and rests on a gap is worse than no number at all.

Playing them

82 seasons are playable in the game, the earliest 1944. The smaller early fields enter at their first complete round, because a partial bracket cannot be drawn as a clean draw. What you get is the real opponents and the real results as the record holds them, with a roster you can change.