skip navigation

Monyepuck: A Reader's Guide to Advanced Statistics (Part 1)

By Andy Blaylock, 10/22/14, 2:45PM CDT

Share

Part 1 of a 2 Part Series on How NHL teams use Statistics

Here in Minnesota, “advanced stats” at the NHL level have told an interesting story in 2014.  This comes from the mostly unknown fact that the Wild were actually favored over the Colorado Avalanche in their playoff series this past April by various internet-based NHL commentators who champion “advanced stats”.

Advanced stats allow you to see things about the game that the final score may not tell you.  Why would this be?  The final score depends a lot on luck.  Players often refer to “puck luck”.  You can make a lot of good plays, put high quality shots toward the net and come up empty handed or just get robbed by a goalie whose performance is mostly out of your control.  The result can be a player making a quality contribution and simply not showing up on the score sheet.  And this luck can also lead to a loss for a team that should have won.

There are several examples of these advanced stats that are used to analyze player and team performance by NHL front offices to assist in deciding which players to acquire to create the most successful team.  The example we are going to focus on today is a stat called “Corsi”.

Advanced stats seek to smooth out the luck factor in order to speak to who consistently is making the better plays in all areas of the ice and thus “deserves to win” more games at a team level or contributes most to their team winning at an individual level.

Bringing this back to the Wild’s playoff series, advanced stats experts had said that the Wild won roughly as many games as they should have this past NHL season based on how they played.

On the other hand, the Colorado Avalanche supposedly won a lot more games in the 2013-2014 season than they should have based on how well they played.  The advanced stats experts said the difference between their winning percentage and how well they played is explained by an abnormally high save percentage (see Semyon Varlamov’s status as finalist for the NHL’s Goaltender of the Year Award, the Vezina Trophy, in 2014) and an abnormally high shooting percentage.

Analyzing historical NHL save percentages and shooting percentages, advanced stats experts say that you can’t rely on keeping an abnormally high shooting percentage or save percentage going even if you keep our team mostly intact and keep your goalie.  These things just simply regress, or “come back”, to the average over time.  Because they come back to average values over time they say that save percentage and shooting percentage are mostly out of a team’s control, and because of this, they said that the Avalanche were “lucky” last year to pile up a good enough record to win the best division in Hockey.  As such, they said the Wild were the favorite in the series, even though the Wild were the seven seed and the Avalanche were the two seed in the conference.

That luck factor seemed to hold true throughout the classic series as it seemed the Wild deserved to win six of the seven games (at least from a Minnesota perspective).  But, in the end, what a great series that made!

Another notable fact with respect to NHL use of advanced stats that may pique your intertest is that two of the strongest users are the LA Kings and the Chicago Blackhawks.  Two facts make that interesting.

1.  The prominence of the use of these stats has really hit the mainstream in the past 7 years with the most recent years being the strongest examples.

2.  Those two teams own the most recent three stanley cups and four of the past five.

As you can imagine the rest of the NHL has taken notice.  This off season, most all organizations are looking to hire “math people” who can give them the tools to take numbers and turn them into things that “hockey guys” can use.  News of new “advanced stats” hirings in NHL front offices has been hitting almost weekly this past summer.  If you are interested in reading more on this, this article (pretty long read) sums the NHL’s push in the advanced stats direction.

Understanding all of that, I would like to conclude part 1 by explaining what the most prominent advanced stat in Hockey (which is named “Corsi”) is.  Most hockey parents understand the concept of “plus-minus”.  Plus-minus is a stat that considers who was on the ice when a goal is scored.  If your team scores when you are on the ice, you add one to your plus-minus stat.  If your team is scored upon when you are on the ice, you subtract one from your plus-minus stat.

Corsi is a similar concept except it is designed to factor in all shots.  This means, not only shots on goal, but every shot attempted.  This is done in order to increase the “sample size” of the data (larger sample sizes make for more robust conclusions from any data set).  The basic version of Corsi is simply a plus-minus stat that considers your team’s performance while you are on the ice and gets at this by adding one to your Corsi for each shot attempt by your team and subtracting one for each shot attempt by the other team.

So, why is this stat (and similar statistics concepts) all the rage in NHL front offices these days?  What value could understanding this development have for Youth Hockey parents?

In part 2, we will consider what parents can learn and apply to their child’s Hockey by understanding what makes Corsi valuable to NHL decision makers.

About the Author

One of Minnesota's premier hockey trainers, Andy Blaylock joins the YHH Staff to write about the dynamics of training, both on ice and off.  Andy is the General Manager of Competitive Edge Hockey in St. Louis Park. His content will emphasize the importance of high quality in-season and off-season training. In addition to running his own private clinics and camps, Andy has trained several organizations including Andover, Anoka, Edina, Hutchinson, STMA and Wayzata.

To view all of Andy's articles on YHH click here. Andy can be reached via email at Andy@compedgehky.com 

 

Recent MN YHH News