Anaheim DucksArizona CoyotesBoston BruinsBuffalo SabresCalgary FlamesCarolina HurricanesChicago BlackhawksColorado AvalancheColumbus Blue JacketsDallas StarsDetroit Red WingsEdmonton OilersFlorida PanthersLos Angeles KingsMinnesota WildMontréal CanadiensNashville PredatorsNew Jersey DevilsNew York IslandersNew York RangersOttawa SenatorsPhiladelphia FlyersPittsburgh PenguinsSt Louis BluesSan Jose SharksSeattle KrakenTampa Bay LightningToronto Maple LeafsVancouver CanucksVegas Golden KnightsWashington CapitalsWinnipeg Jets


Note*** A full NHL schedule breakdown by week is available here on a Google doc.

While doing some amateur scouting in the QMJHL for McKeen’s during the lockout, I found myself asking questions I couldn’t answer:

1-      How accurate is my assessment of players on a limited amount of views?

2-      To what degree are the skills I am witnessing (and reporting) relative to the level of play in this particular game?

3-      How will those skills project at the pro or NHL level? (Because that’s what we’re all trying find out in the end, isn’t it?)

4-      How can we find those late blooming ‘diamonds in the rough’ BEFORE they bloom and everybody finally notices them? (Because that’s every scout’s fantasy, isn’t it?)

5-      What are these elusive ‘intangibles’ everybody always talks about, how can we make them more tangible and how much will they eventually affect the player’s development?

We will probably never have definitive answers to these questions, but as a community of people trying to reach a better understanding of this great game, I believe it is our duty to investigate them as thoroughly as we can.

Baseball fanatics have SABR (Society for American Baseball Research), which has done (since the mid 80’s) a tremendous job of helping to better understand the mechanics of the sport.  Hockey now has its own, yet unnamed research brotherhood.  More and more bloggers have joined the movement toward using available data to uncover the game’s secrets, to put ‘conventional wisdom’ to the test by observing it through a statistical microscope never before employed.

Many findings have already been made public all over the internet and in a few publications, but after sifting through many of them, I have come to realize that hockey suffers from a bad case of data-deficiency.

Baseball has always been a statistician’s preferred sport.  But even baseball’s conventional numbers were proven by sabermetricians (members of SABR) to be misrepresenting of players’ actual skills and value.  Consequently, researchers came up with new metrics, but even more importantly new ways of charting the field of play.  I would argue that hockey is way overdue for a statistical revolution that would bring about better empirical support to traditional scouting methods and therefore paint a more accurate picture and projection of teams and players.

The challenge facing this emerging faction I will propose to name C.A.D.H. (Community for Advanced Data in Hockey) is that contrary to Bill James (founding father of SABR), we have so little relevant data to draw from considering hockey’s poor record of scorekeeping.  Much of the new statistical research centers on shot attempts at the opposition net and shooting percentage (see research on Corsi, Fenwick via NHL Numbers and PDO).  While this has proven to be helpful by multiplying game events, thus creating a much better tool than shots on goal for assessing team and individual performance, more detailed scorekeeping is needed to further CADH’s cause.

Where are these shot attempts taken from exactly?  How has the shot opportunity come about (created by the shooter, a pass, a rebound, a one-timer, a deflection or a turnover)?  Was the shot’s path clear or was the goalie screened?  If the shot resulted from a pass, was the pass deflected?  If it was a one-timer, was it clean or did the shooter have to adjust to a pass in front or behind him?  Was the puck lying flat or bouncing/fluttering as the shot was taken?

I know, I know, after reading the last paragraph, you’re all thinking: “This guy is out of his mind.  Why would we need that much information?”  Well, simply put, because it is the only way to truly understand the dynamics of the game: quantify and qualify as many of the events that occur during a hockey game.

Corsi and Fenwick metrics rely on shot attempts only because no better data is available.  The reasoning behind the use of shot attempts is that it ‘probably’ is the best indicator of possession/zone time that we have.  Then why don’t we start recording ZT?  PDO relies on shooting and save percentages, but we all know the weakness of those statistics: not all shots are created equal in terms of their odds of ending up in the net.

This brings us to the heart of the matter: scoring chances.  We know NHL teams track them; coaches even like to refer to them in their post-game interviews when commenting on their team’s performance.  Some TV broadcasts post scoring chances within their stats package during intermissions.  I certainly believe that defining what is a scoring chance (SC), qualifying SC’s by type and charting them during NHL games would be the start of a statistical revolution that would forever change the way we analyze the game of hockey.  It would open up many new avenues for team and player evaluations by creating a new set of stats that is a better reflection of the reality (and complexity) of the sport.

The first step is developing the charting tools to record scoring chances by type during a game.  (Editors Note: This is well under way, as well as zone entry data that is likely the beginning of the acceptance of this data on a game-by-game basis). This is what I propose here and it will be my next contribution to the CADH movement.  Stay tuned, hopefully this is only the beginning of a beautiful thing.

Follow the McKeen's team on Twitter: