Hockey evaluation continues to evolve
Analytics has begun to change how hockey is being evaluated
Hockey is a sport that among fans, bloggers, reporters and even some teams within the league; vary on how they evaluate players. For the longest time, scoring production was always king when looking at statistics to determine how good a player was and how they should be valued in the NHL. The movement of analytics becoming more prevalent in the sport has caused a shift in player evaluation by some.
To be clear right off the top this isn’t a piece that is meant to cause a debate of analytics versus the “eye-test.” Everyone looks at the sport in different ways and that’s perfectly fine. It’s what makes debates and discussions on topics interesting. There’s room for both analytics and watching the game with your eyes to evaluate players. Looking at just numbers in the evaluation of the sport will trip you up at times and completely ignoring the numbers may give you a false understanding of players.
Over the last year, I’ve spent time studying the analytical data within the sport and utilizing them to improve my analysis of the game. Therefore, while the NHL is in a lull, I thought I’d spend some time explaining how analytics can show you which players are good or overvalued outside of scoring stats.
I’m not one that completely discounts goals and points of a particular player. At the end of the day, the team with the most goals wins the game. The flaw is that some people put too much into those numbers. The actual goal or assists are the focus, however, the play that created the goal by the other players on the ice is often left out of the conversation.
Ideally, you’re not looking to craft a roster of 21 skaters that score a bunch of points. Teams should be trying to acquire point producers, but also add players that drive offense, as well as dominate the roles that they’ll be assigned on a given team.
In theory, if you have the puck more than your opponent and create quality scoring chances when you have the puck, that team is more likely to win. Thus, general managers should look to get players that excel in those areas of the game.
This past season Nikita Kucherov, Connor McDavid, Patrick Kane, Leon Draisaitl, and Brad Marchand were the top five point producers in the league. Kucherov was rewarded for his 128-point season with the Hart Trophy as the league’s most valuable player. I’m not going to discredit Kucherov’s season because it was great, but was he the most valuable player in the league? Probably not.
Looking at analytical data such as Evolving Hockey’s RAPM expected goals plus-minus per 60 and Corsi plus-minus per 60. As well as goals above replacement (GAR) per 60 and wins above replacement (WAR) from their site, I crafted a listed of eight players that were arguably more valuable to their respective club. Those eight players are Mark Stone, Sean Couturier, Erik Karlsson, Patrice Bergeron, Sidney Crosby, Ryan O’Reilly, Brayden Point, and Aleksander Barkov. The chart below shows their impacts in the aforementioned stats from Evolving Hockey, as well as Kucherov’s metrics in relation to them.
This puts him on a similar level of impact as Point. Now, when you account for his 128-point season, he should be bumped into the top five on that list. Even so, there’s a good argument to be made that Mark Stone, Sidney Crosby, and Sean Couturier were more important to the success on their team than Kucherov on the Lightning. Crosby finished in second place in Hart voting, while Barkov and Point did receive votes as well. However, one of the best play-driving forwards in the game in Stone, didn’t receive one first or second-place vote. He received a third-place vote (1), fourth-place (3), and fifth-place (4).
The overall point is players like Stone and Couturier are not considered on the top tier with some of the best in the game like Crosby, McDavid, and Kucherov because they regularly score under 80 points a season. This is where the eye-test and data can intersect. Some people understand Stone is a good player by watching the game but undervalue his importance if they ignore the analytics that shows he should be considered on that top level.
That starts to bring us into the valuation of a player. The NHL is a salary cap league and having players that give you value on the cap is important in building a successful roster. At times you’ll see analytics people compare two players and say that they rather have player X on their team over player Y. In most cases, this is where those that don’t follow the numbers snicker, but these claims are valid.
One prime example is a comparison of Andrew Copp and Patrik Laine. The Buffalo Sabres have been tied to the Winnipeg Jets in trade rumors and those two players have come up in various conversations on the internet.
If I’m building a team I would rather have Copp on my team over Laine. It’s not because I think Copp is a better overall hockey player. It’s because for the value of their contracts and play I’m likely getting better results from Copp. Don’t get me wrong, Laine is a good player, but also a flawed one. He’s an elite shooting talent that is dominated when he’s on the ice at 5 on 5. Copp, on the other hand, is a bottom-six forward that has dominated his role. Laine will cost at least $7 million annually (maybe more) and Copp will come in around $3 million (maybe less).
The valuation here favors the player that will be cheaper and dominate their assigned role on the club. Where the other player may produce at a higher rate but will very likely be crushed in their assigned role at 5 on 5, resulting in being a net negative for his team.
Again, this isn’t saying that one player is better than the other. It all ties back into using analytics to have a better understanding of the players on your roster to know which of them bring value to the team. The advanced stats are not meant to come in a replace what you can see when you watch a game or replace entire scouting departments. They’re meant to provide that next layer of analysis and get a deeper understanding of the sport that in some cases are not noticeable when you watch a game.
The common misconception in this debate is that you have to be on one side or the other. In reality using a mix of watching games and understanding the importance of the underlying numbers, will give you the best opportunity to evaluate various players properly.