Time to draw the hypotenuse. I’ve already looked at how Max Pacioretty’s and Erik Cole’s goals vary with and without their usual center, David Desharnais. In the previous posts, we’d found out that both wingers had scored more without Desharnais than with him; while that didn’t mean that Desharnais was pulling them down, it did cast some doubt on the narrative surrounding the unusual chemistry of that line. This post examines their goal scoring with or without each other, completing the triangle and finishing the story about the perception.
|5-on-5 with Cole||16||13.58||1.18|
|5-on-5 with Gionta||5||2.06||2.42|
|5-on-5 with others||5||3.08||1.63|
|5-on-4 with Cole||4||3.09||1.29|
|6-on-5 with Cole||2||N/A||N/A|
|5-on-5 w/o Cole||10||5.14||1.95|
|5-on-5 with Pacioretty||15||13.58||1.10|
|5-on-5 with Cammalleri||3||2.49||1.20|
|5-on-5 with others||5||3.41||1.47|
|5-on-4 with Pacioretty||10||3.09||3.24|
|5-on-4 with others||1||N/A||N/A|
|5-on-5 w/o Pacioretty||8||5.90||1.36|
All the previous caveats apply: the samples are small, goal-scoring is very “luck-driven”, even the lowest G/60 numbers on this table are very good, and the highest numbers would’ve been unlikely to be sustainable. Nonetheless, the results are in the same vein. And the small sample has to be kept in mind, yet we find that Pacioretty scored more goals per minute played 5-on-5 without Cole than with him, and likewise that Cole scored more 5-on-5 goals per minute without Pacioretty than with him. That’s a rather counter-intuitive result and a direct contradiction of the perception that this line had some sort of special chemistry.
Because of the small sample sizes, I hesitate to draw too much from the other entries. But that Gionta-Pacioretty combo is intriguing. The Pacioretty-Plekanec-Gionta line has been suggested here before, and not solely by me. The data here, while very fragmentary, suggests that beyond its two-way ability it has value as an offensive unit, should Therrien elect to use a power-on-power line matching scheme. We would not expect Pacioretty to continue scoring well above Stamkos’s pace in the long run as part of that unit, but at the very least, it has the history of a small explosion of goals in the short span it was used in 2011-2012 to recommend experimenting with it.
Still, this data is interesting because it gets us to revise our perception of the effectiveness of the famous 67-51-72 line that was treated as an unbreakable unit for most of last year and seems to be in the plans to start next year again. In reality, both wingers are highly effective in their own right, and both scored more goals, relatively speaking, when the line was apart. It’s possible that this result means something different than reflecting the overall effectiveness of the line. For example, it may reflect a propensity to score during line changes. Some of the difference may also be influenced by the difficulty of opposition and/or quality of teammates (read “was Subban also on the ice”?) or a factor of both wingers hogging the puck more when apart than when together (and thus reducing the shooting of their linemates). This is a question that can be answered only with the time-honored, favorite tactic of quantitative analysts everywhere: applying more data.
The next step is to examine the story in terms of scoring chances. With larger samples and generally less impacted by shooting luck, this should give us a better picture of the overall effectiveness of the various combinations, even though I lack the data on who personally shot each scoring chance.