Corsi Is Dead. Long Live Corsi.

20150128-01

All eyes in the NHL this weekend were turned to Columbus, where you could see a once interesting and entertaining fixture in the hockey world now turned into a laughing stock, filled with professionals just going through the motions and putting on an embarrassing display.

No, I’m not talking about the All-Star Game, I’m talking about Ken Campbell and The Hockey News.

With news breaking that the NHL is about to revamp it’s website to add new statistics, as well as trialing new technologies to automatically track even more on-ice parameters, Campbell penned an obituary to Corsi.

Not only was he quick to dance on Corsi’s grave, but he threw in a few shots at the “pocket protector crowd” all the while showing a complete lack of understanding for what Corsi actually is and how it is used.

To guys like Campbell, Corsi has gone from the category of dumb idea to obsolete overnight. But I suppose we shouldn’t be surprised. I mean, Campbell has long associated Corsi with death for some reason:

Anyway, according to Campbell and his ilk, gleefully cheering about the death of Corsi, the NHL will soon be tracking things like puck possession times and territorial zone times, as if that is the only useful piece of information that Corsi provides. Sure, Corsi (the number of shot attempts, for and against, taken while a given player is on the ice, expressed either as a raw +/- or as a percentage) is a proxy for possession, but it is much more than that.

Corsi is also an indicator of productive puck possession.

For a crowd that for so long has criticized Corsi and Fenwick (same as Corsi but without blocked shot attempts) as not mattering because what really matters are goals, it’s amazing to me that they would now jump all over minutes and seconds as the thing that will kill Corsi.

Down with the calculator. Up with the stopwatch.

Do they not realize that possession and zone times are yet one more step removed from the almighty goal?

Anyway, I’m getting distracted from the point, which is this: Corsi is not just about puck possession. And additional tracking technologies and mechanisms will not do away with Corsi. They will either (a) provide new and additional data points to track, study, analyze and correlate with other pre-existing information, or (b) make the collection of stats that are already in use more accurate.

They will not kill Corsi. They will make it stronger.

In fact, there is news just yesterday that the 30 new stats categories being added to NHL.com include Corsi, Fenwick, PDO (sum of save% and shooting%, and which trend to 1.0 over long periods of time).

So no, Corsi isn’t dead. Corsi is alive and well and can now even be used in salary arbitrations.

That being said, there is something that is becoming more and more obsolete one mind-numbing article at a time:

20150128-02

But like the proverbial frog cooking in the pot as the water gets hotter, these guys will continue to go about their lives dismissing the changes around them.

Campbell is not alone. There are a host of hockey writers out there that either don’t get it, don’t want to get it, or just plain hate things that aren’t officially sanctioned in some way. If you’re wondering who I’m talking about, here’s a handy guide to the usual suspects:

20150128-03

Heck, much of the anti-stats movement among the traditional analysts and writers has been one long, drawn out appeal to authority. If it’s not from an NHL or team official, or from a former player, it can’t possibly be of value. Well, now that the NHL is going to start providing these stats directly, I’m sure you can expect a host of “well yeah, I’ve been using this for years” type of pieces.

Either way, one thing is sure: the rumours of Corsi’s demise have been greatly exaggerated.*

* This was supposed to be the title of this piece, but noted jerk, Andrew Berkshire over at Habs Eyes On The Prize had to go and steal it.

RECENT GRAPHIC COMMENTS

You can also check out the monthly collections of Graphic Comments over at The Sporting News, including:

  • Steampuck

    I’ve always said that Corsi cannot be looked at in a vacuum. Adding stats like zone times, and puck possession time is just a natural progression of hockey analytics. Hockey is a moving and dynamic game, the more stats that we can quantify/track the better we can paint the bigger picture.

  • Baumerman

    The problem with the old school “traditional stats” mentality is not that they are anti-progress and against rigorous study of the game, but rather they are reacting to the cognitive dissonance that it is creating.

    Many of these old school thinkers (coaches, managements, analysts, etc) have built their career on traditional stats and their eye test. As unfortunate as it sounds, they have a difficult time reconciling the fact that there way of life, the process by which they’ve built a career upon, is proven to be incorrect or at least seriously flawed. And to further infuriate them, who are these people that are telling them they’re wrong- young, educated people…who come up with ideas in their parent’s basement (I joke).

    For example, imagine you are a Leafs management executive. Who has made a 40 year career based on your opinion. All of a sudden a they higher a 28 year old as an assistant GM (Dubas). And he tells you that your opinions are wrong. What would be your reaction?

    Will this change? Well with old school thinkers on NHL teams it already has because of a plain and obvious incentive -winning. Results matters and analytics already has a proven track record. There will be hold outs, but this change has already started to happen.

    TV analysts are another story. Their incentive is to be appealing to the broad audience. If the viewing audience does not understand or appreciate analytics, there is less pressure for the TV analyst to learn and incorporate analytics. The more we call out bad analysis by TV analysts (like Garry Valk claiming the Canucks need to hit more to win games) the faster they will be forced to incorporate analytics.

  • Baumerman

    I will start off by saying that Corsi and the other statistical information used to assess a team and individual’s performance are excellent tools although, in my opinion, they are better suited for individuals. I spent a good portion of my life working with statistics and one of the first things I learned is that statistical information is meant to be used over an extended period of time, i.e. years and not the 3 periods of a hockey game. There are so many intangibles in a hockey game (illness, injury, hot goalie, bad ref’s etc.) that talking about Corsi on a game by game basis makes no sense. Statistics show wide fluctuations over the short term but over the long term everything evens out and gives a better indication of what’s going on. I remember a number of commentators remarked that Colorado was due for a fall and they used Corsi, as well as other statistics in coming to this decision. That was an excellent use of statistics but the commentators who analyse one game based on how good or bad the teams Corsi was are simply wasting our time because the information is of little value.