Why enhanced statistics are a game changer for NHL contract negotiations
Published 29 September 2015 By: Ryan Lake
The 20th century saw multiple attempts to quantify and measure every aspect of an athlete’s performance. The desire to measure and compare the performance of one player to another to develop criteria in which to establish differences and find similarities between and among players has driven individuals ranging from team executives, players, agents, and fans to push the limits of traditional statistics.
This article examines how players and player agents have been and could further utilize enhanced statistics to bring a new and more sophisticated approach to the contract negotiating table, which has the potential of revolutionizing the sport of hockey. The article will first discuss the genesis of advanced statistics, second highlight the enhanced statistics recognized by the NHL, and third explore how players and their agents can utilize the information provided by the analytics to change the way player contracts are negotiated.
The genesis of statistics
Baseball pioneered the field, which has become known by many names including: Advanced Statistics, Sabermetrics, Analytics, and now Enhanced Statistics. Earnshaw Cook was one of the earliest innovators in the field.1 Cook’s book, Percentage Baseball, which was first published in 1964, was widely criticized at the time.2 However, Cook’s research laid the groundwork for Bill James and others that developed the modern day advanced analytics, a field that has become widely accepted in baseball and has now started to cross over into other sports.
One of the most notable examples of the use of advanced analytics was demonstrated by the Oakland Athletics in the 2002 season. Billy Beane, the manager of the Oakland A’s was presented with a unique situation where his team did not have the budget to compete with the other teams in the league to acquire and retain the top talent in Major League Baseball.3 In an effort to find talented players on a budget, Beane turned to the use of sabermetrics.4 The team assembled by Beane and Paul DePodesta also known as Peter Brand, the assistant general manager of the A’s, set a league record in consecutive wins and reached the postseason.5 The success of this season was detailed in the book, Moneyball: The Art of Winning an Unfair Game, by Michael Lewis.6 Later the true story of the A’s unique and unprecedented approach to building a team was immortalized on the silver screen in the movie, Moneyball, starring Brad Pitt as Bean.7
Many, have dismissed the applicability of advanced statistics in sports other than baseball. The cynics have long said that advanced analytics work in baseball due to baseball’s unique characteristic of being a team sport defined by a series of one v. one battles.8 However, this has not stopped dedicated believers from developing algorithms that are capable of accounting for the team nature of the other sports.
In the last ten years, there has been a revolution, developing under the surface in the hockey blogosphere. The movement, largely led by Matt Fenwick and Tim Barnes (aka Vic Ferrari), the creators of the Fenwick stat and Corsi stat respectively, burst onto the hockey scene in 2015.9 Throughout the summer between the 2013-14 season, several teams in the National Hockey League (“NHL”) hired experts in advanced statistics to help scout and acquire talent.10 In February of 2015, the NHL formally recognized advanced statistics, or as the league refers to them enhanced statistics.11
Traditionally advanced statistics or enhanced stats, have been a tool utilized by teams to build a successful team on a budget. However, analytics are starting to have a much broader application.
In February of 2015, the NHL formally announced a partnership with software company SAP to manage and host enhanced statistics on the league’s website.12 The enhanced statistics that were chosen by the NHL to be hosted on their web page include Corsi, Fenwick, PDO, Zone Starts, a variety of modifying statistics and rate statistics.13
Corsi: Tim Barnes, the creator of the statistic, developed a method utilizing shot attempts to measure puck possession. Under the Corsi measurement, “a shot attempt is counted any time a player tries to shoot the puck. They are counted as a shot on goal, blocked shot or missed shot. By adding those three types of a shot together, you get the number of shot attempts.”14 Over time, the number of total shot attempts has been proven to correlate with total puck possession. This statistic can be utilized to examine both a team’s puck possession as well as the performance of an individual player.
Fenwick: Matt Fenwick, developed another statistic in order to measure puck possession, however, this does not account for shots that are blocked by the opposing team. Fenwick considered shot blocking to be a skill and should not be included in the shot attempts metric. One can gain great insight when examining the Fenwick metric against Corsi, as this will provide insights into how effective a team or individual player is in blocking shot attempts.15
PDO: Brian King, developed a statistic that added on-ice shooting percentage to on-ice save percentage to measure what is commonly referred to as “puck luck.”16 When combined the average percentage of the two values should equal 100%. When a the resulting value is above 100% it can be said that the team or player is on a lucky stretch, and conversely a resulting value below the average can indicate that the player or team is suffering from bad “puck luck.”17
Zone Starts: The NHL also tracks the zone of the ice in which players start their shift. This metric only takes into account for those times when the player is on the ice for a face-off. This statistic helps provide context to the other enhanced statistics. For example, a player may be deemed more valuable if they have a higher percentage of zone starts in the defensive zone and are still able to put up respectable Corsi and Fenwick numbers.18
Modifying Statistics: Modifying statistics, while based on the foundational statistics of Corsi, Fenwick and the other long established traditional statistics, modify these metrics to account for even strength vs. shorthanded vs. power play situations as well as score situations. These stats can provide a more complete picture of a teams or players' performance based on game situations.19
Rates: The amount of time players are actually on the ice during any one game vary from player to player and game to game. Rate statistics were developed in order to account for the difference in game time. Rate statistics, in effect, create a level playing field in which all skaters can be compared and measured.20
While this is not an exhaustive list of enhanced metrics currently in use in hockey today, these statistics will provide several examples that we can utilize in examining how a player or their agent can level the playing field when it comes to negotiations with teams.
Player contract negotiation and salary arbitration
It is not uncommon for sport organizations to place restrictions on which evidence can be utilized in the negotiation of a player’s salary. The NHL is no different. In January 2013, the NHL and the National Hockey League Players Association (NHLPA) came to an understanding on a new Collective Bargaining Agreement (CBA). A CBA is an agreement that employers and unionized employees have negotiated with regard to wages, hours, and working conditions.
One key aspect of the CBA is set out in Section 12. This section details a process known as salary arbitration, which is a formalized arbitration process in which a player, who meets certain qualifications, and his agent are unable to come to an agreement regarding the salary terms of a new contract, with the team who owns the players' rights. During this process, a neutral arbitrator listens to arguments from both the team and the player and his representatives as to what an appropriate salary would be.21
The salary arbitration process can become quite contested and cause an uncomfortable amount of tension between the team and their player. An important element of the argument for both sides deals with measurable statistics that point to how impactful the player is to the team. The NHL in the 2013 CBA included language regarding which statistics are admissible in the salary arbitration process.
Section 12.9 (h) “Statistics” governs the type of statistics that can be utilized in the formal process of salary arbitration. This section states that the “League shall obtain and provide to the NHLPA any statistics relative to any aspect of Player performance: (i) kept or maintained by the League; or (ii) retained by any club.”22
As of February 2015 the NHL has been maintaining enhanced statistics, which means that they are now admissible in the salary arbitration process. Additionally, since these statistics are admissible evidence in the arbitration process, teams and agents are starting to make them a greater focus in contract negotiations.
In a recent article in the Hockey News, Allan Walsh, a prominent hockey agent with Octagon hockey, shed some light on how the utilization of enhanced statistics can change the playing field during contract negotiations. Walsh explains the impact enhanced stats had on one of his contract negotiations earlier this year. In the article, Walsh described a negotiation with executives of an NHL team, in which he was involved in a discussion of traditional metrics such as goals, assists, and points. During this discussion, Walsh presented the executives with an enhanced stat book, which highlighted metrics such as Corsi and Fenwick. After providing the book to the executives, Walsh states “I saw them open the first page, and I saw the GM and the assistant GM lock eyes with each other, and the look on their faces was, ‘Oh s—, he knows.’”23
While the utilization of enhanced statistics will not turn an average player into Sidney Crosby, it can be highly useful in a situation where you represent a player who does not have the best traditional statistics but does have strong numbers in other less traditional metrics.
When appropriate, enhanced statistics can provide an extra tool in the agent’s tool box when negotiating with clubs. The utilization of how an agent could have improved their negotiating position can be seen when examining the recent contract signed by Pierre-Alexandre Parenteau.
On the first day of the NHL free agency period, Parenteau signed a one-year, $1,500,000.00 (USD) deal with the Toronto Maple Leafs. Parenteau will serve as a good example to demonstrate how enhanced statistics can be utilized in contract negotiations for several reasons: 1) the Toronto Maple Leafs front office is at the forefront of the enhanced statistics revolution24; and (2) Parenteau had lower than anticipated traditional statistics during the 2014-15 season with the Montreal Canadians.
The Toronto Maple Leafs, became one of the first teams to establish a front office position for a “stat’s guru” with the hiring of 28 year old Kyle Dubas as an assistant general manager in 2014. The Leafs have continued this progression in 2015 with the hiring of coach Mike Babcock and former New Jersey Devils general manager Lou Lamoriello, both of which have recognized the power of enhanced statistics.
One of the first tasks faced by the reshaped front office for the Maple Leafs was free agency. Free agency is a period, commencing on July 1st of each year, in which free agent players are allowed to negotiate with other teams in the NHL in order to enter into a new contract. According to the NHL CBA, a free agent is a term used, in accordance with certain specified criteria, to describe a player whose contract has expired.25
One of the first free agent players the Maple Leafs targeted was Pierre-Alexandre Parenteau. Parenteau was the ninth round selection of the Mighty Ducks of Anaheim in the 2001 NHL Entry Draft and began playing in the NHL in 2003. Parenteau has traditionally been a scoring right winger posting 227 points in 347 NHL games and earning a selection to the 2012-13 NHL All-star game.26 Despite solid traditional numbers, Parenteau has traveled around a league more than one would expect, having played for the Chicago Blackhawks, New York Rangers, New York Islanders, Colorado Avalanche, Montreal Canadians and is currently a member of the Toronto Maple Leafs.27
The 2014-15 season was a down one for Parenteau, in terms of traditional statistics. In 56 regular season games with Montreal, Pierre-Alexandre posted 8 goals and 14 assists for a total of 22 points. This was by far the lowest production by Parenteau since the 2009-10 season. The low numbers were partially instigated by injuries which caused Parenteau to be in and out of the lineup. After failing to find a consistent role for Parenteau and with a plethora of young, talented forwards, the Montreal Canadians decided to buy out Parenteau’s contract, which resulted in him becoming a free agent.28
When presented with a player who has had a down year, many agents will advise the player to sign a short-term agreement. A short term agreement will allow the player to reestablish themselves as a productive player and allow them to explore free agency, which will hopefully result in a more lucrative long term contract. One can speculate that this was the strategy implemented by Parenteau and his agent.
On July 1st, hours after the start of the free agency period, Parenteau signed a one-year $1,500,000.00 (USD) deal with Toronto. This was a substantial pay cut from Parenteau’s previous contract that had an annual average of $4,000,000.00 (USD).29
As with any contract, the factors that go into determining a player’s value is multifaceted and complicated. However, it can be argued that the Maple Leafs were able to sign an impact player at a significantly lower price point than the market would demand.
When evaluating a player's market value, it is important to find common factors in which you can compare to other players. These comparables lay the groundwork for the agent’s argument as to why a player is worth a certain amount. In Parenteau’s case, and for the sake of brevity, we will explore a very high level and limited application of how an agent can utilize both traditional and enhanced metrics to evidence the player’s true market value.
Since the summer of 2015 was the first off-season, after the NHL’s formal recognition of enhanced statistics this analysis will limit the comparable players to only those that were free agents at the same time as Parenteau. It is also important to determine a traditional metric in which to start the comparison. One well-established metric in which shows the player’s average point production per game is known as points per games played (“PTS/GP”). This statistic can be broken down further still to show an average point production based on an increment of 60 minutes of ice time. This metric is known as points per 60 minutes (“PTS/60min”). Both PTS/GP and PTS/60min help create a standardized measurement of production that takes into account the fact that not all players play the same number of games in a season and further not all players play the same amount of time in each game.
One hundred and thirty-five players were unrestricted free agents during the 2015 free agency period. In order to determine players that could be used as comparables for Parenteau, certain parameters would have to be determined. First, to find useful comparables only skater’s whose primary position is classified as a forward should be considered. Eliminating defensemen and goalies brings the number of available free agent players to sixty-nine.
Sixty-nine forwards is still too large of a pool to be useful in developing players of similar value. Therefore, other factors need to be considered when deliberating what may make one player more valuable than another. One such factor that tends to play a large role in both the term of a contract and the value is the age of the player. At the time of receiving the offer from Toronto, Parenteau was thirty-two years old. Once a player reaches his late twenties, and early thirties teams tend to be more conservative in their offers in terms of both the length of the contract as well as the value. In order to take this factor into consideration, it would be prudent to only examine players near the age of Parenteau.
Utilizing the established metrics of PTS/GP and PTS/60min along with the factors of age and position the comparable pool of players decreases from sixty-nine to six including Parenteau.
Figure 1 below takes a high-level look at free agent forwards, based on their PTS/GP and PTS/60min production from the 2014-15 season. This chart further indicates the average annual salary for each player based on their post free agency contract.
The chart above identifies Eric Fehr, Chris Stewart, Antoine Vermette, Daniel Winnik and Joel Ward as comparables. An examination of Figure1 reveals that Parenteau had comparable PTS/GP and PTS/60mins to the other players, however Parenteau agreed to a contract substantially less valuable than the others.
Other factors such as injury history, team needs, and the role each player plays on their respective teams may have had an influence on the contract that was offered to Parenteau. Additionally, Parenteau and his agent would have had to overcome the stigma that follows a player after being bought out of a contract. However, the utilization of advanced statistics such as Corsi For Percentage and Fenwick For Percentage may have provided Parenteau and his agent leverage at the negotiating table, especially when negotiating with a team as sophisticated as Toronto.
Having established a pool of comparable players, one can then utilize enhanced statistics to further show strengths, or in some incidences weaknesses to a player’s game. The following chart explores how Parenteau compares to the other free agent forwards identified in Figure 1, in terms of Corsi For Percentage. Corsi For Percentage, or CF%, refers to the total shots taken by the player’s team while he is on the ice divided by the total number of shots taken by either, his team or the opposing team while the player was on the ice. Looking at Parenteau’s CF% for example, indicates that for all of the Corsi events that occurred while Parenteau was on the ice last year, 50.7% of them were taken by Parenteau’s team. This CF% is the highest such percentage among the pool of comparables.
As a means of providing a wider perspective, Sidney Crosby, arguably the best player in the world, posted a CF% last year of 56.3%. Parenteau’s CF% is a strong number that indicates that his team possess the puck the majority of the time when he is on the ice. Additionally, Parenteau has been very consistent throughout his career, in terms of his CF%, posting a 50.4% for his career. Parenteau’s strong possession numbers are further evidenced when one looks at his Fenwick For Percentage. Fenwick For Percentage or FF% is similar to CF%; however, Fenwick events do not include shot attempts that were blocked by the opposing team. Again Parenteau posted the highest percentage among the comparable players, with an FF% of 51.5%. This means that Parenteau’s team controlled the puck more often than not when he was on the ice last season. Parenteau’s FF% is a very strong number, by a means of comparison, Sidney Crosby posted an FF% of 56.3% last season.30
While, the statistics highlighted above are only three of many different analytics currently in use in the hockey world, they provide another tool in the negotiating tool box. It is interesting to consider whether Parenteau is undervalued under his new contract with the Maple Leafs or if there is an underlying reason that Parenteau accepted what appears to be a contract offer that was well below market value. It is also important to note that the Maple Leafs are at the forefront of the enhanced statistics movement and have clearly recognized Parenteau’s value.
The use of enhanced analytics is revolutionizing the business of hockey and is following in the path of baseball, where such statistics have become a staple in contract negotiation and salary arbitration. The full application of enhanced statistics is just beginning to be explored and has the ability to drastically change not only player contract negotiations, but also the way sponsorship and endorsement contracts are negotiated. Additionally, other sports have started to take notice; in the summer of 2015 at least one team in the National Football League began tracking advanced analytics. The movement started by Earnshaw Cook in 1964 has the ability to completely revolutionize the way players are evaluated and enhance the way fans experience sport throughout the world.
This work was written for and first published on LawInSport.com (unless otherwise stated) and the copyright is owned by LawInSport Ltd. Permission to make digital or hard copies of this work (or part, or abstracts, of it) for personal use provided copies are not made or distributed for profit or commercial advantage, and provided that all copies bear this notice and full citation on the first page (which should include the URL, company name (LawInSport), article title, author name, date of the publication and date of use) of any copies made. Copyright for components of this work owned by parties other than LawInSport must be honoured.
- Tags: American Football | Baseball | Contract Law | Employment Law | Ice Hockey | National Football League (NFL) | National Hockey League (NHL) | National Hockey League Players Association (NHLPA) | NHL Collective Bargaining Agreement
- Terminating a contract for “just cause”: how bad do things have to be? A review of CAS’s decision in Salkic
- A full review of the National Labor Relations Board’s decision in NCAA v. Northwestern
- MLS’s CBA negotiations: federal mediation, salary cap and steps toward free agency
- How social media and smart devices are influencing the negotiation of Sports Media Rights Agreements