For years, the National Hockey League held a place of its own as the league of unpredictability.
Low-score tight games have been a staple of the sport, providing the exact recipe you’re looking for if you want to see recurring upsets. Combined with the standings structure – where teams can only capture one of two points by losing in the extra time frame or on penalties – the league’s ecosystem is one of perpetual convergence.
It’s a huge difference to other North American sports leagues and comes with advantages and disadvantages. Tight, low-score, and often random single-game results can keep most teams in a theoretical tie-break for long stretches of the regular season.
Compare that to the National Basketball Association – a league where you may only need a few weeks and a dozen games or so to see talent beat a random chance in the rankings. The NBA regular season, especially the tail end, is often painful to watch. But the main draw for this league is the prospect of the most talented and capable teams in general winning, which is not always the case in the NHL.
Michael Lopez, the NFL’s senior director of data and analytics, did some great work on this in 2018. He and a team of researchers discovered that the best (by regular season ranking) team in the NBA is advancing from a playoff streak. 80 percent of the time. For the NHL to achieve a similar rate of progression to the superior team, it would require a best streak of 51 – marked teams.
Why am I bringing this up as we approach September? We’ve talked at length here over the past two years about NHL offensive renaissance – There are more goals across the league, the average total number of goals per match, therefore, the average winning margin is higher. This is a great sign if you are looking for greater predictability within the sport; Talents within the league are given more opportunity to showcase their skills, and the best ever hockey game has become a compelling watch.
The current environment – where teams score 6.3 goals per game, compared to just 5.4 six years ago – will further differentiate the haves and have-nots. Strong teams will cement their place early in the regular season; Weak teams on the sidelines of playoffs will be less frequent in March.
Consider the past two seasons versus the first in the modern era (2007-08), as well as the league’s scoring low point (2015-16). Boxplot shows the average winning margin for all teams:
The smaller the box, the more compact the data set. In other words, the 2007-08 season did not see a significant spread of talent among the majority of the league’s teams. Compare that to the 2021-22 season, in which the fund’s size has increased dramatically – an indication of a significant disconnect between the two clubs.
Instead of focusing on outliers like the best and worst teams in the league by season, consider the quarters where the majority of teams sit. If we look at what the 25th and 75th percentiles mean for a team in 2007-08 versus 2015-16 and 2021-22, the difference is astounding.
The gaps between these two groups — the inherent teams and teams bad enough to draft the lottery but not bad enough to win it — grew as the scoring increased. While the gap between Buffalo and Columbus in 2007-08 was just 37 goals (0.45 per game), the gap between San Jose and Minnesota last season was nearly tripled at 106 goals (1.29 per game).
If this trend maintains itself, it changes the behavioral economics of the university. On a relative basis, years ago it was much easier to take a limited talent team into a playoff race – these teams were confusingly progressing as deadline trade buyers because they were just a few points from the playoff race.
This is not the case now. More scoring means fewer penalties and less extra time, fewer penalties and less extra time, which means fewer extra points thrown around the standings, and fewer extra points thrown around the standings, which means more visibility into the team’s talent real.
The NHL doesn’t want a league as predictable as the NBA, but I think it seeks to reward teams that perform well, and reducing perceived in-game randomness can do just that.
This is a league built on parity, and that won’t change – not until the league reviews the win, loss, and overtime/penalties points structure. But the NHL’s welcoming of an offensive decision not only showcased more skill. It also added a touch of credibility to the sport, as more talented teams are likely to win matches.
This is a good long-term development for the league.
Data via Natural Stat Trick, NHL.com, Evolution Hockey, and Hockey Reference