I’m waiting for the post-season to finish before I get started with my model building process. So, in the meantime I want to do a bit of writing. The plan is to write short articles about ideas and concepts that fascinate me. These may or may not be hockey statistics related.
I will call these short articles “Perspectives”.
In the first perspective I want to talk about Team sports and how they’re different from Individual sports. This is something that I have briefly mentioned in a Twitter thread before, but I think it’s worth re-examining.
In team sports you say that:
The productivity is simply the team performance – the results.
The Potential is the Quality of the individual players. The name potential is somewhat misleading here since we’re talking about current quality/ability and not future potential ability.
Processes is how well the players work together.
So, basically all the equation says is that the productivity of a team depends on two things.
- The Quality of the players.
- How well the players work together.
This shouldn’t be surprising to anyone. Sometimes a team performs better than the sum of its parts, and sometimes a team performs worse than the sum of its parts.
A team consisting of all the best players in the world may not be the best (most productive) team in the world.
In an individual sport the best player is simply the player with the highest productivity (the best results). Otherwise, he wouldn’t be the best player. This doesn’t mean that the most talented and most gifted player always ends up as the best player. You’re still dependent on circumstances like injuries, training facilities, coaches etc.
…But you can never be better or worse than the sum of the parts, because there is just one part – You. I get that this is probably an oversimplification. You could argue that a coach deciding on the strategies is also part of the equation.
Let’s get back to the interesting part – Team sports. So, what is a process? This can be difficult to define, and honestly the concept can be somewhat “fluffy”.
Some processes are easy to understand. What tactics are the team using? Do the tactics fit the player types? How well do the players communicate on and off the ice. Are the players eating and sleeping well?
Other processes are harder to explain. Do the players have on-ice chemistry? Do they believe in the concept? A coach can have the greatest concept in the world, but it won’t matter if the players don’t buy-in. Do the players trust each other? A defender that doesn’t trust his goaltender will do everything in his power to block shots, but then he might give up passing lanes. And a goaltender that doesn’t trust his defense will cheat towards the pass rather than focus on the shot.
And then there are the very “fluffy” processes. Are the players happy (work-life balance)? Do the players like each other?
The role of the GM vs. The role of the Coach
In a Pro league like the NHL, the ultimate goal is to win the Stanley Cup by optimizing productivity. If we simplify things, you could say that the job of the GM is to optimize the potential (put together the best possible players) and the job of the coach is to optimize the processes (get the players to work together as well as possible).
In real life the roles are less clear, though. Most GMs will look to add players that fit the team’s needs and style, so they aren’t just looking to improve the potential but also look at how an addition will impact the processes.
Likewise, the coach can impact the potential of the team. Player development is one way to improve the potential of a team… But you could even argue that player usage is a way to change the potential of the team. If a coach uses his best players the most without making them too tired, then he can increase the potential of the team.
In youth hockey the end goals should be different. Here the main focus should be on Player development, and not necessarily so much on the processes that lead to good results/productivity. Especially not, if process optimizing lead to worse player development.
Taking one step backwards in order to eventually take two steps forward
Sometimes a team need to take one step backwards, so that they can later on take two steps forward. This concept is obvious when a team goes through a rebuild.
However, the coach can even do this… but it takes courage to do so. If a coach wants to try something completely new, he must expect a lower productivity during the implementation period. This is difficult for a coach to do because he is judged solely on the team’s productivity. Unfortunately, I find most coaches very conservative for this exact reason.
How is this relevant to Hockey Statistics?
So, how is all of this relevant to hockey statistics? It’s important to understand that almost all statistics measure productivity and NOT potential. So for instance, when you see a GAR value for a player, then it estimates the player’s impact on the productivity given the team’s processes. It doesn’t estimate the player’s true potential… So, on a different team with different processes the player might perform differently.
Again, I don’t think there’s anything groundbreaking about this theory. It’s just important to know that statistics measure and project productivity and not true potential/quality.
That will be my first “Perspectives” article. I hope you enjoyed it, and there will be more in the coming days and weeks.