Like any other business in the world, the NHL is rapidly finding new ways to analyze market inefficiencies. In today’s league, those teams that have bought into analytics are ahead of the curve.
In a salary cap sport, finding cheap talent is the difference between winning and losing. Finding all of the right parts to go around your long term controlled players is the recipe for today’s league. Capitalizing on the entry level contract structure is paramount to fielding a serious contender. In addition to the entry level contract, there are players that are undervalued by the league’s decision makers.
Simply put, in any complicated system, there will be assets improperly valued. In the case of the NHL, the league’s dinosaur mindset has held several organizations back. Instead of being open to new ideas, the old guard has clung onto their old ways. Learning how to adapt to a changing game is one of the league’s biggest challenges.
Currently, we are witnessing the transition to a speed oriented league. The league currently over values one dimensional defensive defenseman. These players provide no upside in the offensive zone, yet get long term contracts. Learning how analytics work can begin to bridge the gap between the results on the ice and the concepts everyone is familiar with.
Related Story: Five potential trade targets this summer
In addition to explaining how basic possession metrics work, this will explain their shortcomings.
Possession and the new school
In a nutshell, the new school of hockey thinking is straightforward. When a team possess the puck, they are generating shots. Therefore, if a team is generating shots, they are not playing defense. In theory, this means that a team that possess the puck more should generate more shots which should lead to more goals. Of course, like any rule, there are exceptions. Occasionally a team that is thoroughly outshot will still win games from time to time.
Now, to measure the control of the puck there are the stats Corsi and Fenwick. Corsi For Percentage or CF% measures shot differential for two teams while at even strength. For example, the Rangers CF% as a team last season was 45.92%. This means that the Rangers as a team had only 45.91% of the shots during its games this season. A good possession team will as a collective be over 50%, meaning that they generate more than half of the shots within a game.
There is significant evidence that proves that over the course of an entire season, the CF% metric is a solid indicator of team success. Obviously there are more factors at play than just the raw total of shots compared to the other team. Another stat that comes into play is the high danger scoring chance. Using this metric which defines a high danger shot as a shot that comes from the inner half of the circles in the offensive zone that then cuts in at about five feet from the net mouth.
In this realm, the Rangers were middle of the pack, finishing 19th in the league in high danger CF%. Meaning that New York had a greater share of the high danger scoring chances in a game than only 12 other teams.
The over/under valued
In this school of thinking, the most overvalued players are players who do not drive possession. Using this past season’s Rangers, the player with the worst CF% who played more than 20 games was forward Cody McLeod. This is one area of overlap between the new school analytics and those who favor the eye test. Enforcers do not provide anything in the means of playing hockey other than emotional support. McLeod’s unsightly 37.3 CF% means that the Rangers would have been better off on the penalty kill than with the forward on the ice.
That means that when McLeod was on the ice, the other team had nearly two thirds of the even strength shot attempts.
Now, on the opposite end of the spectrum, Mika Zibanejad had the team’s best CF%. The forward drove possession with a 51.8 CF%. Zibanejad was only one of just three players on the team that finished with a positive CF%. For context, the Pittsburgh Penguins, Washington Capitals, Nashville Predators and Winnipeg Jets had 16, 7, 13, and 17 such players respectively.
It is this school of thinking that is trying to root out inefficient players that do not drive possession. The theory being that the more shots a team generates, the more goals they will score, which means they are more likely to win. That is why giving defenseman like Dan Girardi and Marc Staal, who are both negative possession players did not make sense. The league is trending towards two way defenseman and neither could do much in the offensive zone
Final thoughts
Although possession metrics are a relatively new tool, they are considered commonplace amongst most progressive thinking hockey organizations. The mindset being that in conjunction with a scout’s eye test, analytics help round out more of the picture. That is all that analytics of any kind are, a tool for measuring. Not a one size fits all bill.
For example, the Washington Capitals are a negative possession team for the 2017-2018 season and they are just one win away from winning the Stanley Cup. In fact, both teams in the Stanley Cup final are riding a high wave of PDO. PDO, a stat which is used to quantify luck, adds together a team’s shooting percentage and goalie save percentage to give a single number. The league average for shooting percentage is nine percent and the bare minimum for an NHL quality goaltender save percentage is 90.
These averages are why the baseline for PDO is 100, therefore any team that has a PDO over 100 is considered playing lucky or above their talent. For the season, the Capitals have a PDO of 101.5%. This number means, that even if the caps are just one win from the Cup, they are getting lucky.
Next: Could the Rangers return to playoff contention soon?
Analytics will never be a perfect measure of a hockey player’s talent. In the postseason they are outright unreliable due to a small sample size. However, over the course of an entire regular season, they typically hold true and are an indicator of success.