OK, so here is what I did. It is somewhat complicated and I might not have the right approach so I appreciate feedback, but bear with me while I try to explain.
This is just using +/- as a first approach, the same technique could be applied to shot attempts or whatever you prefer. Let's not focus on that for now, more on the approach.
For each player, I took their +/- with other players and compared it to the second player's average +/- (weighted by TOI) when not playing with the first player.
For example. Ben Hutton when not playing with Gudbranson has a TOI-weighted average +/- of -1 with every other player.
Then I took that metric and divided by the players ice-time, and multiplied by their ice-time with the "evaluated player" to get an expected +/-. So Ben Hutton, since he has spent about 70% of his TOI with Gudbranson, has an expected "+/-" of -0.70 with Gudbranson. Basically. Then, Gudbranson's "impact" on Hutton is (-19) - (-0.70) = -18.3. Finally, I then applied this to every player to calculate the average impact that Gudbranson has on each player.
Then, finally, I calculated the "average impact" that every player has on every other player on the team, again weighted by TOI.
Does that make sense? I know, it's confusing and took me some time to get my head around it.
Here is the Canucks:
Player | AverageImpact |
ERIK GUDBRANSON | -6.39 |
BRANDON SUTTER | -4.09 |
BEN HUTTON | -3.59 |
TYLER MOTTE | -2.09 |
MARKUS GRANLUND | -1.81 |
BRENDAN LEIPSIC | -1.36 |
TIM SCHALLER | -1.34 |
CHRISTOPHER TANEV | -1.07 |
LOUI ERIKSSON | -0.87 |
JAY BEAGLE | -0.82 |
JAKE VIRTANEN | -0.53 |
ALEXANDER EDLER | -0.37 |
ADAM GAUDETTE | -0.35 |
SVEN BAERTSCHI | -0.35 |
BO HORVAT | -0.11 |
ALEX BIEGA | 0.11 |
MICHAEL DEL ZOTTO | 0.44 |
NIKOLAY GOLDOBIN | 1.01 |
ANTOINE ROUSSEL | 1.22 |
DERRICK POULIOT | 1.51 |
TROY STECHER | 3.17 |
JOSH LEIVO | 3.33 |
BROCK BOESER | 3.43 |
ELIAS PETTERSSON | 5.17 |
[TBODY]
[/TBODY]
Pouliot is pretty surprising here, as is Tanev, but it does line up with what we saw from the individual WOWY's, where Gudbranson and Sutter have basically terrible scores with everyone, and Pettersson great scores with everyone, so we accomplished our goal.
But the question was, is Erik Gudbranson the worst in the NHL? Are there other players who consistently have worse "with" than "without" on their teammates?
Here is your answer:
Player | AverageImpact |
ERIK GUDBRANSON | -6.39 |
MATT NISKANEN | -6.01 |
DMITRY ORLOV | -5.78 |
IVAN PROVOROV | -4.99 |
MARC-EDOUARD VLASIC | -4.27 |
MAGNUS PAAJARVI | -4.26 |
CODY CECI | -4.25 |
RYAN CARPENTER | -4.18 |
BRANDON DUBINSKY | -4.14 |
BRANDON SUTTER | -4.09 |
[TBODY]
[/TBODY]
So there you have it. At least in terms of goal +/-, nobody consistently has worst +/- with everyone on their team, than their teammates do without.
Having said that though, this list indicates that this is likely not a perfect approach. Most of these players are not considered terrible players. In fact, Vlasic is often considered elite. Is this just because we are using goal +/- instead of something like shot attempts?
Here is what happens if we apply the same methodology to shot attempts.
Player | AverageImpact |
MAXIME LAJOIE | -69.43 |
CODY CECI | -67.03 |
ERIK GUDBRANSON | -56.5 |
TOM PYATT | -53.95 |
CHRIS TIERNEY | -53.39 |
NEAL PIONK | -50.94 |
MATT NISKANEN | -47.96 |
BOBBY RYAN | -45.36 |
BRETT HOWDEN | -45.33 |
RYAN DZINGEL | -44.78 |
[TBODY]
[/TBODY]
People might like this list better, but on the other hand there is a whole lot of Ottawa on that list! Which indicates to me that I haven't appropriately accounted for the bad team effect.
Anyway, so this is a work in progress but feel free to let me know what you think.