Rabid Habs

The Coexistence of Analytics and Results

NHL stats

While enjoying televised golf from the beautiful coast of California, it donned on me the inroads that advanced statistics have made throughout sports. Driving accuracy, greens in regulation, percentage of putts made from various distances, par 3 scoring, scrambling, etc., help to provide up to date information regarding individual players. Chances are that if you hit it longest and straightest (overall driving) and have the least number of putts per 18 holes you are going to be winning golf tournaments. So, why do the leading indicators in hockey not always preclude or align to point production, wins and championships? Humbly, respectfully, it’s because you can’t qualify a team sport the way you can an individual one. While a golfer is solely responsible for his or her statistics, there are far too many variables to accurately assess a player’s worth through the application of puck possession alone.

Prior to proceeding further, I wish to offer a disclaimer. I believe that reviewing all available data helps to better evaluate both an individual and team’s successes and liabilities. Data helps to support on ice evaluations, which remain a critical component in assessing personnel. However, to solely rely on advanced statistics, while ignoring on ice results is equally detrimental. Lars Eller offers an excellent example. Advanced statistics suggests that Eller’s puck possession numbers are among the highest on the Canadiens roster. When I watch Eller play, I see a young man who establishes himself offensively through an effective cycle, but who lacks the ability to finish plays.  He carries the puck into the offensive zone, yet seldom does he drive the puck to the net creating scoring opportunities. This is reflected in his point production, which is on pace to be in the 23-28 point range, a far cry from his 30 points during the strike shortened season (2012-13). So, is Eller simply having bad luck? Is there less correlation between puck possession and points? Or, conversely, is puck possession alone a key indicator in determining a player’s success? Have we reached the point whereby a player will be paid, not based on points, but on advanced statistics alone? Personally, I think $3.75 million dollars annually is a lot of money for what might amount to 25 points. Then again, it might be a bargain compared to PA Parenteau’s 6 goals and $4 million annual price tag.

I’m an educated man, but I’m not so certain I see the value in singularly applying one statistic to justify an opinion. Data can be manipulated and small samples can be extrapolated to qualify a position. That said, to be objective, one must also open their mind to new ideas and consider all available information prior to rendering an opinion. The difficulty is, IMO, the fact that more statistics are required to evaluate a player in the face of decreased scoring. 25 years ago, hockey players’ successes were based on wins, losses, points and +/-. Now, as scoring tests new depths, advanced statistics are required to evaluate a player’s worth as we can no longer count on points solely in an era in which teams may not reach 200 goals over 82 games. I’ve often believed that such advanced statistics must be reserved for single events (golf), or match ups in which variables can be reduced (pitching vs batting in baseball), but not held static. Team based sports are so difficult to evaluate as the variables are so wildly inconsistent. But, alas, the fact remains that the evolution of sport, hockey specifically, might necessitate the further integration of analytics. The business of sports is just so big that I can’t imagine a player agent not focusing on positive puck possession or corsi to mask scoring deficiencies or lowered point production among clients.

In closing, can I help to close the divide that exists between advanced analytics and on ice results? No. Are they mutually exclusive, maybe. I require more information, a larger sample size and the continuation of an open mind prior to relying more heavily on individual statistics in a team based sport. Sure, it’s easy to see the correlation between advanced stats, points and results as it applies to PK Subban or other elite players. But, the process becomes significantly less clear as we, the fans, try to evaluate Christian Thomas (small sample size), or third liners who aren’t assigned offensive roles and responsibilities. Do we focus exclusively on points? On other measurements including hits or shots? In these instances I can see the value of advanced stats to determine where the player started his shift and where he helped to push the puck (into which zone). But, again, I struggle with isolating the other four players on the ice, opponent, etc. So, I continue to collect more data and further educate myself. In the interim I will continue to try and see all players through an unbiased lens. After all, it might be old school, but I still believe the best way – the truest way – to evaluate talent, is to watch the game live. In the absence of having a ticket, try to remove the emotional component and evaluate how an individual influences the outcome. In who’s end is the game most often played when they are on the ice? Lastly, points, wins and losses will continue to greatly qualify an individual or team’s success. Statistics can help to offer key indicators. However, they are not absolute. Sure, it’s unlikely that you will win more often than not if you are getting outshot every game.

But, it’s hockey. Expect the unexpected…..ole

I welcome your feedback. Please connect with me @LWOScjcasselman.