Hindering offence

A quick follow up to something I found last week when digging into Oscar Klefbom’s early season numbers.

To re-cap, with Klefbom on the ice with him this season at even-strength (5v5), McDavid’s on-ice Corsi For percentage (i.e., a proxy for possession) was 55.08% - a great proportion, right around where we would expect things to be when an elite forward is on the ice with a very good offensive-minded defenceman. And for good reason, we’ve seen the coaching staff play Klefbom as much as possible this year with the top offensive forwards to squeeze out every bit of offence possible.

Now I was expecting both players to post a a slightly lower Corsi For percentage without one another, but was surprised to see McDavid take a significant hit - dropping down to a 45.94% on-ice Corsi For percentage without Klebom this season. Below is what the results were after 13 games.

• McDavid with Klefbom (111 mins) – 55.08% CF%, 75.0% GF% (6 GF, 2 GA)
• McDavid without Klefbom (101 mins) – 45.94% CF%, 66.67% GF% (4 GF, 2 GA)
• Klefbom without McDavid (129 mins) – 55.17% CF%, 62.50% GF% (5 GF, 3 GA)
• Oilers without Klefbom or McDavid (271 mins) – 46.40% CF%, 35.0% (7 GF, 13 GA)

Knowing what we know about McDavid and his ability to drive offence, I think it’s fair to assume that something is dragging him down - either the play of an individual player or perhaps the team’s overall tactics. The fact that Klefbom’s Corsi For percentage has stayed above 55.0% without McDavid made me focus on McDavid’s numbers and what his numbers were like with the other defencemen.

To get a sense of how each defencemen has done this season after 14 games and what their on-ice possession numbers have been like with the four lines, I pulled their on-ice Corsi For percentages with each of the regular centermen, using these forwards as a proxy for the lines. The most common deployment has been McDavid, Draisaitl and Strome centering their own lines. And I picked Brodziak as a proxy for the fourth line as he really hasn’t spent much time with the other regular centers. I could have picked Khaira, but he’s played about 15 minutes with Strome this season.

Below are the results. I’ve highlighted the Corsi For percentages in green if it’s above 52.0% and highlighted them in red if it’s below 48.0%. Anything between 48.1% and 51.9% is in yellow. The Oilers team Corsi For percentage is currently 50.32% (14th in the league).

We’re obviously dealing with small sample sizes, so the idea here is to look at what has happened rather than predict what’s likely going to happen in the future. This isn’t a final word on what the Oilers should do. Rather it’s about highlighting these numbers and asking further questions around why they’re happening.

Focusing on McDavid’s numbers, we see that his on-ice Corsi For percentages are well above the team average of 50.32% except when he’s with Nurse or Russell. When McDavid has been on the ice with both of them, the team has scored only once and allowing none - an on-ice goals-for per hour rate (GF/60) of 0.94, well below the 3.0 mark that we would expect when McDavid is on the ice. What’s especially concerning is the drop in offence whenever McDavid is on the ice with Nurse. When McDavid is on the ice without Nurse, the team generates 66.6 shot attempts per hour (CF/60). That number plummets to 55.0 when Nurse steps on the ice with him. The team rate right now is 57.12 (14th in the league, right around league average).

The rate of unblocked shot attempts for (i.e., Fenwick, FF/60, a proxy for scoring chances) also takes a significant hit when Nurse is on the ice with McDavid. McDavid’s on ice rate goes from 58.4 unblocked shot attempts per hour without Nurse to 38.6 with Nurse. The team as a whole generates 44.54 unblocked shot attempts (12th in the league, right around league average). In my mind, at no point should McDavid have an on-ice rate of shots and scoring chances below the team’s average..

I don’t think Nurse has magically become a bad player. In 2017/18 he played most of his minutes with either Larsson or Russell. And with either partner, he didn’t drag down McDavid’s possession or offensive numbers this badly.

• McDavid with Nurse-Larsson: 300 mins (5v5), 53.04% CF%, 62.77 CF/60, 50.38 FF/60
• McDavid with Nurse-Russell: 147 mins (5v5), 49.69% CF%, 64.88 CF/60, 51.09 FF/60

I suspect that there’s something tactic-wise that the coaching staff is doing differently this season when Nurse is on the ice or they may have adjusted something about his game. Whatever the case may be - it’s impacting McDavid’s offensive output and should really be addressed. Especially if Nurse is going to continue leading the Oilers defencemen in ice time at even-strength.

Data: Natural Stat Trick