clock menu more-arrow no yes

Filed under:

Evaluating Defensemen using their effect on both team and opponent Corsi%

New, comments

There is more than one way up the mountain when it comes to player evaluations.

Jana Chytilova/Freestyle Photo

Tyler Dellow has recently pushed for defensemen evaluations via their effect on their team's Corsi%, also known as relCorsi. Over at Hockey-Graphs, I published today that there is another angle to look at as well: how a defenseman affects their opponents Corsi%.

The article is worth a look and some discussion. The punchline though is in this graph here:

Bqc02rvceaa0vu3_medium

For fun, I thought we'd like to look at how different Jets defenders fared in these different metrics.

Player

delTmC%

-delOpC%

distance from (0,0)

Combined (dTm+(-dOp))

Tobias Enstrom

2.70%

2.40%

3.60%

5.10%

Dustin Byfuglien

1.80%

1.50%

2.30%

3.30%

Paul Postma

0.00%

-0.30%

0.30%

-0.30%

Grant Clitsome

0.00%

-0.60%

0.60%

-0.60%

Jacob Trouba

-0.70%

-0.40%

0.80%

-1.10%

Zach Bogosian

-1.00%

-0.40%

1.10%

-1.40%

Mark Stuart

-2.40%

-2.20%

3.30%

-4.60%

Jets defensemen with 500+ 5v5 minutes accrued on team and signed for next season.

A positive delTmC% means that Jets players are in general better with the player on the ice than off.
A positive -delOpC% means that the Jets opponents do worse with the player on the ice than off.

Some thoughts:

* There is an obvious link but  some differences exist.

* Tobias Enstrom is really good at hockey. I don't care what anyone says about his size. They are misguided.

* The Jets will likely miss Byfuglien on defense, although if Trouba can be freed of his anchor, maybe we can start to see an awesome Enstrom-Trouba pair.

* Both Trouba and Bogosian seem to hurt their team's Corsi% more so than than they help their opponent. This discrepency is likely due to their defensive heavy deployment.

* Clitsome has the most extreme opposite effect, meaning his relCorsi may be overstating his effectiveness.

* Mark Stuart received an extension and...  well...