Does the best team win?
In the end, maybe so.
The team that wins is qualitatively the best team.
The question is: Does the quantitatively best team win the championship? Is it the team with the best record going into the championship game? Basketball, the younger brother in the sibling rivalry that is American sports, does not have that definitive game, but rather a series during which two teams duke it out for supremacy (or at least national supremacy) of their sport. It does come down to one game sometimes, but not like it does in football.
The NFL has a winner-take-all mentality that often offsets records going into that final game: the Super Bowl. If you have made it that far, then whatever your record (regular season) was coming in doesn’t particularly matter. It is about that miracle push into the playoffs and winning games when it matters.
Since the merger in 1970, when the AFL and NFL became once league, only nine teams claimed (or tied) the best regular-season record and won the Super Bowl (1997 Packers, 1993 Cowboys, 1989 and 1990 49ers, 1986 Bears, 1983 Redskins, 1979 Steelers, 1975 Steelers, and 1973 Dolphins). That means that about 21 percent of teams with the best regular-season record end up taking it all and hoisting the Lombardi trophy. Compare this to fourteen teams since the ABA-NBA merger in 1976—a total of 37 percent. The Boston Celtics and the Michael-Jordan-led Chicago Bulls are among those who have done it multiple times. I did not run advanced statistics on those two numbers, but I think we all might look at those and cock our heads ever so slightly.
Why is there a discrepancy?
My guess is that it has to do with the illusion of home-field advantage (which has a long, sordid history among sports statisticians), and the lack of said advantage in a Super Bowl, which is played at a neutral site. The NBA Finals, on the other hand, have everything to do with home-court advantage—especially given recent changes to the sprawl of the games by league administration. Additionally, the top record in the NFL is afforded a first-week bye. Does this give credence to the idea of teams going on a streak, something that fundamentally defies classic statistical modeling? Do statistics matter in the world of sports?
Yes and no. Statistics, at best, are used to describe phenomena and give reasonable predictions within a given margin of error. However, I do not think we should prescribe to an absolute predictive validity. That is a jargon-laden way of saying what beer commercials have been telling you: superstitious rituals are only weird if they don’t work.
I think it means the best team emerges, whatever that elusive term means in the end.