Author: Harrison Rubin (harrisonrubin2023@u.northwestern.edu), Weinberg ’23

Each year in the hockey community, there is a debate about who are the most underpaid and overpaid players in the league. Twitter users would go back and forth using WAR and other standardized metrics as proof of why a certain team gave a certain player a certain contract. As a data analytics nerd and hardcore hockey fan, I decided to end this debate once and for all, using a statistical regression analysis. Using R, I built a regression model to figure out (1) the main factors that influence a player’s salary, (2) how accurate the model is (using root mean squared error as the key metric), and (3) who the most underpaid, overpaid players are, as well as whose salary was closest to their projected salaries.

I decided to only examine players who played more than 13 games last season– out of a possible 56 games in the shortened season — as well as rookies that are on their entry level contracts (maximum of $925,000). This is because players who played only a small handful of games, as well as rookies who performed as well as more experienced players, adversely affected the model’s salary projections.

The model was fairly accurate, but not totally accurate, as the root mean squared error was about $1.5 million. With that in mind, listed below are the most important variables separated by forwards and defensemen.

 

The most important variables for forwards are not incredibly surprising, as forwards are typically paid more money to score more points, therefore increasing their total of point shares. However, for defensemen, the results are a bit more shocking. In hockey there are two types of defensemen: offensive-leaning and defensive-leaning, or shutdown defensemen. Offensive-leaning defensemen tend to shoot more, while shutdown defensemen are on the ice longer and usually do not score as many points. The model results show that offensive-leaning defensemen are paid higher salaries, as the importances of offensive-leaning metrics are greater than those of defensive-leaning metrics.

After the model was created and analyzed, I decided to figure out which players are underpaid, overpaid, and which players’ actual salaries closely match up with their projected salaries.

Underpaid Players:

Adrian Kempe (LW, Kings)

• Predicted Salary: $4,847,815
• Actual Salary: $2,000,000

Last year, Kempe scored 14 goals, 15 assists, and 2.5 point shares, while attaining a -18 plus/minus rating. Over the course of 56 games played last year, Kempe averaged 16.9 minutes of on-ice time. Kempe’s scoring total ranked second on the Kings, while his overall point total ranked fifth on the team.

MacKenzie Weegar (D, Panthers)

• Predicted Salary: $5,590,361
• Actual Salary: $3,250,000

Last season, Weegar took a giant step forward in his development, totaling 6 goals and 30 assists while achieving 7.4 point shares and amassing a +29 plus/minus rating. Across 54 games played, Weegar averaged 22.1 minutes of on-ice time. Among his fellow Panthers teammates, Weegar had the highest plus/minus while having the most points among the team’s defensemen.

 

 

Got it (Almost) Just Right:

David Pastrnak (RW, Bruins)

• Predicted Salary: $6,628,294
• Actual Salary: $6,666,666

Pastrnak continues to show why he is one of the league’s best forwards, putting up 20 goals and 28 assists in 48 games played last season, which is tied for second on the Bruins. Pastrnak also amassed 6 point shares, a +10 plus/minus rating, and was on the ice for 18.6 minutes per game (on average).

Travis Hamonic (D, Canucks)

• Predicted Salary: $3,004,388
• Actual Salary: $3,000,000

One of the Pacific Division’s most hard-hitting shutdown defensemen, Hamonic put up 3 goals and 7 assists in 38 games played last season. Hamonic also had 1.9 point shares and a -3 plus/minus rating while being on the ice for 19.4 minutes per game (on average).

 

Overpaid:

Jamie Benn (LW, Stars)

• Predicted Salary: $5,462,822
• Actual Salary: $9,500,000

Compared to years past, Benn had a down season in 2021. The Stars’ power forward and captain scored 11 goals and 24 assists last season, while accumulating 3.8 point shares and earning a +13 plus/minus rating. In 52 games last season, Benn spent, on average, 17.8 minutes of on-ice time.

Drew Doughty (D, Kings)

• Predicted Salary: $6,050,733
• Actual Salary: $11,000,000

Similar to Jamie Benn, Doughty had a down year after putting up several Norris Trophy-worthy seasons. Across a full slate of 56 games last season, Doughty scored 8 goals and 26 assists, while earning 5.5 point shares and a -14 plus/minus rating. Similar to previous seasons, Doughty led the Kings in time on ice, spending 26.4 minutes per game on the ice in 2021.