NHL Attendance Predictions for the 2013-14 Season

After all the great feedback I received on the Road Attendance analysis last February, I got inspired. I dug into years and years of game by game data to understand trends and impacts to attendance. Some of that research you can see via the analysis I did on the Phoenix Coyotes mid-season. I really wanted

to decipher what moves the day to day attendance. At the core, my findings show that there are 3 main influences: 1) Opponent 2) Time of Year and 3) Day of the week. Although people love to say “winning” as an indicator, my research doesn’t call that out specifically. It is more of a long term effect on attendance and not as variable year to year. i.e. Dallas missing the playoffs for years has an impact on attendance but it is a steady decline vs a cliff from playoff year to non-playoff year. Also when you look at certain franchises like Edmonton, Toronto and Calgary. Losing has not impact on their regular season attendance.

With all that research in my back pocket and a handful of time on my hand since game 6 of the Stanley Cup Finals otherwise known as baseball season, I decided to systematically predict the attendance of every NHL game this season. For comparison purposes, I’ve excluded Winter Classics, Heritage Classic, Premier games (Europe) and the new Stadium Series.

Using my systematic approach, I am projecting a league dip in attendance and not just from 2013 but 2011-12 Season as well.


I have only used math to attack this challenge. I have not taken into account any Olympic impact, ownership resolution (Phoenix, New Jersey, Florida & Dallas) or playoff bump (New York Islanders  & Minnesota). I will do a deeper dive on predicting the impact to some of those teams.

Deep dive into Phoenix 2013 Attendance

I received a lot of feedback and questions on my last post on NHL attendance year over year. Several of the questions were around the Phoenix Coyotes attendance and comparing those to like months since there is conventional wisdom that Phoenix does well later in the year due to snowbirds coming down for the Winter. I went ahead and looked at the average attendance for the Coyotes by month.

As you might be able to see, there is a snowbird effect as Phoenix sees a 30% increase in Attendance for February games over October games. I also looked at how this compares with basic league increase for teams that have Attendance Volatility (teams that don’t sell out every game). For these teams they only see a 6 to 7% increase from October to February. So that starts to explain why the Coyotes attendance is so much higher this year through the end of February games when looking at the first 13 games (in October for 2011-12 season vs. the 2013 season).

Double clicking one more time I looked at the 13 games after January 19th over the last three seasons. I didn’t like how the graphs looked so thought I would just display the table:

This starts to tell a little more concerning story as three games under 10,000 while there was only one in each of the last two seasons. Attendance is essentially flat year over year and off of the 2010-11 season. I will continue to monitor these numbers and unfortunately they are not as rosy as originally reported.

NHL Attendance Numbers for 2013

Inspired by Day one of the MIT Sloan Sports Analytics Conference, I have updated my Year over Year NHL attendance figures. I am going to make this short and sweet

  • League attendance is up 3.4%
  • 4 teams saw a decline in attendance but two of them (Leafs & Rangers) are due to configuration changes in their arenas over the off season
  • 4 teams saw growth over 20% (Hurricanes, Devils, Coyotes and Dallas)
  • Dallas is seeing a 70% increase
  • Nashville Predators have sold out every game this season.
  • That makes 18 of the 30 teams that have sold out every game

For the heck of it I thought I would break out the Coyotes attendance since there appears to be a lot of sensativity about them right now:

As always, let me know your thoughts and if you want to see any specifics.

Road Attendance Volatility in the NHL

With the modified schedule and in conference only games of this shortened season I began to wonder what were the impacts to limiting the set of opponents for a given team. So I took a deep dive into the Attendance figures over the last 2 years to understand a little more of the business side of schedule dynamics.

I assumed that any game that was 99%+ of arena capacity was essentially a sellout,  16 of the 31* teams over the last two seasons have virtually sold out every regular season game played in their barn. Those 16 teams are Boston, Buffalo, Calgary, Chicago, Detroit, Edmonton, Montreal, New York Rangers, Philadelphia, Pittsburgh, San Jose, St. Louis, Toronto, Vancouver, Washington and Winnipeg. Here is how the capacity matrix breaks out:
In addition to teams that have a lot of green, there are several home teams with a lot of red or yellow in their columns. With the 16 teams that have sold out every That leaves 15 teams with what I call Attendance Volatility. It is the teams with Attendance Volatility that I want to dig into. Let’s go ahead and isolate them to help us understand what we are dealing with:

The far right column of the grid above is the volatility index. Another way to look at it is the expected capacity when the visiting team plays a team that does not sell out every game. Looking at some of the detail you can see the best teams on the road are Boston, Detroit, Pittsburg, Winnipeg & Chicago.
The bottom 5(or 6) are Carolina, Dallas, Phoenix, Atlanta*, Minnesota and Nashville.

The most surprising team in the Top 5 is Winnipeg as all other 4 have won the Stanley Cup within the previous 5 years. With determining a common thread (Recent Stanley Cup Champions) of 4 of the top 5 I went back to 2007-08 & 2008-09 seasons to see if the Anaheim Ducks winning the Stanley Cup in 2007 had a material impact on road attendance the following two seasons. The results were rather inconclusive. The Ducks only played 15 teams on the road that they played on the road in the 2006-2007 season. Of those 15 teams, they averaged more than they did in 2006-07 season in only 5 of the 15 sites another 5 saw a decrease in attendance and the remaining 5 were neutral. 

Another interesting take on the data is evaluating the worst match ups for attendance. There were 5 matchups in those two years that averaged less than 50% capacity.
                Anaheim @ New York Islanders

                Carolina @ Phoenix
                Columbus @ Atlanta
                Nashville @ Phoenix
                Ottawa @ Phoenix

As information trickles in on the realignment plans of the NHL, this becomes an interesting reference to determine who are the real winner$ and loser$ on hosting teams that might not draw well on the road.     

Some odds and ends:
  • I broke out Atlanta from Winnipeg to help with comparability
  • According to Wikipedia, the hockey Capacity for the Tampa Bay Times Forum is 19,204 but on February 17th 2011, the reported capacity was 20,849 when the Lightening welcomed the Red Wings to town. After further investigation, the arena had some renovations in 2012 that eliminated 500+ seats.
  • For obvious reasons, I’ve removed counting Winter Classics, Heritage Games and Premier games in the data