The Feigned Distaste for Analytics and a Parallel to Phaneuf

Over the last couple of years, one trend I have noticed during the rise of hockey analytics is the shared negative reaction to the presumptive complexity in aggregated numbers.  Or maybe numbers are not complex, but the ignorance of the complexity in hockey.  The truth is that hockey analytics is still in the embryo stages.  The accumulation of data is still on-going and so are the many discussions of what part of the game is in need of being scrutinized.  But it doesn’t hurt to be critical, does it?

It’s an interesting dynamic watching analytics supporters and traditional viewers fight over rooted beliefs of what exactly can or cannot be quantified.  Having played a very minor role in the discussion – which is just another way of saying that no one really listens to me – I feel that most of the problem lies at the feet of those who support analytics.  They just aren’t very good at conveying the nuances of the data they have.  Of course, it would be even better if someone could collect the cumulative knowledge out there and put it into an easy-to-find and accessible place on the internet where interested readers can catch up.

However, I’m not here to enlighten everyone on what’s right or wrong.  I am the skeptic after all.  I think the root of the issues goes back to what I said in my very first post: “We watch hockey and sports in general because we believe that in the many events that repeat itself, there’s that one event out of a thousand that can change the complexion of a game completely.”

With people invested in analytics, it’s a matter of wanting to have something to quantify events into weighed values of importance so that they know, to some degree, that a certain way of playing tends to win more often than it loses.  That same analytical mindset is why we assign a personal preference to players we enjoy watching – it’s entirely theoretical, but not necessarily practical.  As a fan of the game in its unpredictable state, I think it’s okay to be skeptical, but at the same time it is important to acknowledge the shortcomings of watching with your eyes as well as trying to allocate numbers of hockey’s many nuances.  It is better to be aware of the statistical merits of what we can try to quantify in hockey rather than completely dismissive – and that goes for both sides.

On that note, I have two links for people interested in following up on analytics.  The first link is Aaron Chan’s excellent introduction of Corsi and Fenwick.  Aaron, by the way, is probably a bigger skeptic of hockey analytics than I am.  The second link is more cognitive.  If you’re more interested in the psychology of theory, intuition and mental observation, Steve Burtch gave me a heads up a while back to read “Thinking Fast and Slow” by Daniel Kahneman.

On Dion Phaneuf

This is a subject I have been kind of putting off from writing because Mike Stephen wrote up an amazing summary of how good Phaneuf has been in the past couple of years – last season in particular.  So until recently, I didn’t have it high on my list of priorities to write.  But sometimes, I come across tweets and posts overflowing with condemnation and vitriol towards the Leafs captain, and I can’t help wonder if people understand the absurd role Phaneuf has to play.  I’m not going to name names or post specific examples, but I really want to extol Phaneuf’s value to the Toronto Maple Leafs.

Let’s start with some of the more traditional aspects of his game.  Phaneuf brings a unique mixture of size, skill, skating ability, compete, physicality, and robust defensive play to his position.  I guess there might be some questions about his sense for the game and his hockey IQ, but I find a lot of his mistakes are reflective of the mistakes his own teammates make.  Last night’s game against Minnesota was a good example — he dropped a pass behind him in the neutral zone and Franson wasn’t there to pick it up like he was supposed to.  But that’s just one example out of many.

Anyway, that he merges his impressive mixture of size, skating, and skill on the first pairing of the second youngest team in the NHL cannot be understated – he eats up the ugliest minutes that few other defenseman in the NHL plays.  It is in Phaneuf’s role that we can see just how useful Corsi can be to understand how he mitigates the difficulties the other pairings have to face.

And for the record, I want to make it clear that Corsi isn’t an exact science, but it does help us understand a player’s role by contextualizing the statistical merits we have available to us.

Right now, Phaneuf ranks 13th overall for Corsi Relative to Quality of Competition (Corsi Rel QoC) among defensemen.  Corsi Rel QoC is a measure used to determine a player’s possession numbers vs. his opponent’s possession numbers weighed by the amount of ice-time they are on the ice against each other.  In other words,  Phaneuf is on the ice against players who positively drive possession – and if he’s 13th overall in this category, he is facing the league’s best players.  Last season, Phaneuf was 2nd overall, just behind Olivier Ekman-Larsson (who happens to be a pretty damn good defenseman already).

Another way to understand the statistical uphill battle Phaneuf has to face is to look at how his team performs when he’s off the ice.  This category is known as Corsi Relative (Corsi Rel).  This statistical merit is a pretty interesting feature because usually, if you have a minus number in this category, you’re probably a pretty terrible player.  Except that we already know that Phaneuf eats the dirtiest minutes so that the pairings behind him don’t have to.  Last season, Phaneuf’s Corsi Rel was a -7.2, placing him 21st overall among defensemen with 40 or more games played – this year, Phaneuf’s Corsi Rel is -9.7, placing him 41st overall in the NHL among defensemen.

Obviously, the numbers for this season are too early to place any emphasis on, but at least we know that that Corsi is showing a similar trend in which he is, once again, facing some of the stiffest minutes in the NHL.  However, consider this: If Phaneuf’s Corsi Rel is a negative, while the rest of the defense have had better possession numbers against weaker possession drivers, why are we not celebrating his role so that we can watch Jake Gardiner, Cody Franson, and Morgan Rielly develop at their own pace?

When you couple in the fact that Phaneuf was 10th overall in scoring by a defenseman in 2012-13 and is currently 14th overall this season, the picture painted shows a defenseman whose difficult role and elite production beggars belief that he’s not treated as one of the NHL’s best defensemen. Without Phaneuf, the Leafs don’t have anyone to protect Gardiner, Rielly, or Franson.  Be happy he’s on our side.

Some news

The recent blog traffic has caused me to reevaluate if I want to take the purpose of this blog into my own domain.  I have decided that I would like to see how people react to some of the more statistically inclined writers, so I brought in two friends of mine — once they post something, I will link their Twitter accounts for anyone interested in following them.  They will be posting sometime in the near future.  Until then, I will likely be writing a post on Morgan Rielly by the end of the week.  If you got any questions, give me a shout in the comment section or add me on Twitter.