Would Connor McDavid vs. the Rangers Guarantee Big Interest in the Stanley Cup?
"Biggest Market versus Biggest Star" sounds good on paper, but how much has that mattered in the past?
Ahead of the the conference finals between the Stars and Oilers in the West and the Rangers and Panthers in the East, Front Office Sports speculated about whether a Stanley Cup Finals matchup featuring the NHL’s biggest market (New York City) and its best player (Connor McDavid) would represent the dream scenario for league honchos:
[A]fter last year’s steep viewership drop in the Stanley Cup Final, the league office will undoubtedly be hoping for a juicy matchup to attract the biggest TV audience possible.
New York is a slight series underdog to Florida at most sportsbooks, but having the nation’s largest media market back in the Stanley Cup Final for the first time in a decade would most certainly be the preferred scenario for NHL stakeholders…
Playoff series featuring Canadian teams often don’t draw as well for U.S. TV ratings, since viewership from their home market isn’t measured by Nielsen. But the Oilers have the consensus top current player in the NHL, Connor McDavid, who would be playing in his first Stanley Cup Final since being drafted with the No. 1 pick in 2015. That would be a great test case to find out whether the league’s top talent can draw casual fans on the biggest stage.
This got me wondering: what combination of market size and star power — or just general team quality — has predicted Finals interest in the past?
To begin to answer that question, I looked at all Finals since 2010, gathering data on Stanley Cup Final “interest” through two different measures: Television viewership (using the average rating for each game in the series, per Wikipedia) and internet search traffic (as determined using Google Trends for the topic of the Stanley Cup Final).1 I used the ever-handy geometric mean to average the two together — since that particular mean allows you to properly average things on different scales — giving us a measure of how much interest the general public had in the past 14 years of Finals matchups.2
Here are the most (and least) interesting Finals by this measure:
The next task is to gather data on features like star power, market size and team quality.
The latter category is easy enough — I’ll be using the pre-series Elo ratings of each team to represent their strength going into the Finals. For market size, I’ll use the estimated 2023 metro-area population for each city, with a twist: For metros with multiple teams, I’m splitting the population in proportion to the Google Trends search interest for each team within that metro. And finally, we’ll measure star power using the adjusted Goals Above Replacement (GAR) for the top player on each Finals squad.
Armed with all of that, let’s check out the correlation between each category and our measure of Finals interest, to see what actually tracks with how popular a matchup is:
Perhaps surprisingly, there was little correlation between a series’ interest level and the quality of its higher-rated team, nor was there much of a link around the quality of the best player in the Finals. (In that sense, maybe we ought to be skeptical that McDavid playing for the Stanley Cup would make a dent in fan interest overall.)
Market size was more correlated with Finals interest, with a combination of the two markets involved scoring best, but the factors most strongly linked to more popular Cup Finals were the quality of the lesser star in a series, and the Elo of the lower-rated team. This implies that fans are drawn in by matchups where both teams are great — and where two great players are going head-to-head.
Going back to the original question about what combination of factors are the best predictors of Stanley Cup Finals interest, we can also set up a regression model to predict matchup interest based on the most highly correlated item from each subset of categories in our table above. Granting that there are other factors we’re not including or cannot include (market size, star power and team quality only explain around half of the variation in Finals interest), here’s the relative importance of each factor in our regression:
Team quality: 45%
Star power: 32%
Market size: 23%
Because of this, the regression would predict that New York versus Dallas would be the most compelling matchup. It would mix two great teams — the Rangers and Stars are Nos. 3-4 in the Elo ratings — and the highest possible population combo — the NHL’s largest and third-largest markets3 — even though the all-around star power would be lacking. (New York’s Artemi Panarin can put his résumé up against anybody, but Dallas was led by a good but not great season from Jason Robertson.) By contrast, the Oilers have the star power with McDavid, but not the market size — Edmonton ranks 30th out of 32 NHL markets — and they are easily the lowest-rated remaining team by Elo.4
If the Oilers and Rangers do clash in the Final, it will be a great marketing opportunity for the league to pit a huge fan base against a star that even casual sports fans might know (a rarity in hockey). But if history is any guide, just having great teams skating for the Cup is an even more important ingredient for generating interest in the Final.
Filed under: NHL
Those who click through to the Trends dashboard will see that I compared the Finals to several vegetable-related search terms. How silly! What’s that about? I learned this fun little trick from
: A benign search topic like “celery” or “spinach” works well as a common comparison point for other topics in Trends, since it’s a fairly stable and non-seasonal food item that people are searching for in the course of their normal lives.I was tempted to include more years, but Google Trends data can be a little less reliable (or just wonky) the further back into the 2000s you go.
Remember, the Los Angeles–Long Beach–Anaheim, CA MSA is being split between Kings and Ducks fans.
They’re still good, though! (6th-best in the league, to be exact.) That speaks to just how many great teams were able to advance deep into the playoffs this year.
From looking at the data, it looks like "years that teams have been in league" or "years that teams have been in their current city (for Colorado, New Jersey) would have a strong correlation. The Original 6 and teams from the initial expansions in the years after seem to do much higher ratings and have higher search interest. Given the relatively longer tenure of the Blackhawks vs the Golden Knights, you might have to do something like the Square Root of the years, but I bet that would be an even higher correlation than anything else. You also might have to throw out 2020 and 2021 b/c those were years when the Cup was moved from it's traditional early June time frame.