While thinking about the Rangers-Capitals Game 7 the other day, I started to think about the Stanley Cup Play-off’s many upsets. One of the Stanley Cup’s main selling points is the fact that there are so many of these. In particular, I began to think about the late runs by the Kings (Yay!) and the Capitals. Then, two possibilities popped into my head:
either a) these teams are better than their records showed or b) a best-of-7 is a woefully inadequate way to determine the best hockey teams
In all probability, I think this is a combination of the two factors. (The upset-by-upset data can be found here at Play-offUpsetsMatrix.) I decided to investigate further for all completed play-offs since the NHL Lockout. I devised a formula to calculate the “magnitude” of all upsets:
Magnitude=(high seed wins-low seed wins)*(4-loser wins)*(seed difference)
I classified any lower-seed win as an upset, including finals appearances where the low seed actually had a better record. Even though hockey and basketball have different ways of determinig play-off teams (with hockey awarding two points for a win and one for an overtime loss instead of just one for a win), I have a chart (shown here at Points Differences) that shows that the point differentiation in the two sports is very similar. When I point that that the very best teams in hockey often drastically more points than all the other teams, the gap shrinks further. Therefore, I did not see the need to adjust the formula for hockey.
(I also noticed that both leagues have one conference where the separation between teams is smaller than the other, although the NHL’s Eastern Conference in 2011 had a very pronounced spread.)
Then, I simplified the data into two charts (shown here at Upset Meter), one for each sport. First of all, I found that the NHL has had one-and-a-half times as many lower-seed victories between the 2006 and 2011 Play-offs as the NBA had. Furthermore, they had twice the average magnitude and nearly three times the overall magnitude. The gap would have been even greater had there not been the Warriors-Mavericks and Grizzlies-Spurs 8 seed upsets in the bookend years of this study. This, in combination with the similarity in regular season record spreads between play-off teams, has led me to believe that the NBA does a better job in determining the best teams, both during the regular season and the play-offs, than the NHL does.
Okay, fine, what am I going to do about that? Well, I believe that the NBA should not change its season or play-off length in any way; it does a great job of determining its champions. In 2008, there was only one lower-seed victory, and it had a magnitude of 0 because the teams had identical regualar season records. Take away the Warriors-Mavericks upset and the years 2006-2008 seasons had a combined upset magnitude of -13! While upsets have become more frequent in the past three postseasons, the NHL had higher magnitudes in 2006, 2009, and 2010 separately than in the last three seasons as a whole.
That brings me back to the NHL. The issue here is that hockey is such a grueling sport that more games could not be easily added without putting additional, otherwise unnecessary wear-and-tear on its players. However, I believe there is a two-part solution to fixing this.
For one, the NHL could increase its season length back to eighty-four games. This is not a big jump at all; just about 2.5%. Second of all, the NHL played eighty-four games in every season in the 1992-93 and 1993-94 seasons. I’m not sure why they changed it, although it may have to do with the 1994-95 lockout. They’ve played that many games before, and they can do it again. However, every increase to the schedule heightens the possibility that the best teams do better (and of injury, but that’s another story).
The other thing they can do is make the play-offs (at least after the first round) best-of-nine. Again, it’s not much, but it stills increases the probability of the best teams winning. In all, it might not change the probability that the better team will eventually win by that much, but every game helps.
On second thought, the NHL will probably not want to do that. If upsets became more rare, it may lose some viewership who rooted for the underdogs. My Kings may not have gotten past the Canucks in a best-of-nine.
On third thought, it might actually help. With this proposal, when Cinderella does come to the ball, maybe she’ll receive an even warmer welcome.