clock menu more-arrow no yes mobile

Filed under:

Wild About Numbers: An Introduction to dCorsi With Wild Defencemen Rankings

In a change of scheduling here on Hockey Wilderness, from now on, instead of 5 Noon Numbers a week, there will be just 2, both written by me. For now, they won't have a set day but they will both be posted between Monday and Friday. The plan is that both articles will, hopefully, be more informative and longer than the old Noon Numbers. Also, instead of "Noon Number" these will be called "Wild About Numbers".

Just Spurgeon, doing his tank-thing.
Just Spurgeon, doing his tank-thing.
Justin K. Aller

Stats-orientated hockey writers are always looking for ways to improve on the advanced stats metrics they currently have at their disposable. While Corsi, Fenwick, PDO etc. have been remarkably successful in predicting future performance, there are still people looking for ways to make them better and develop new metrics.

Steve Burtch, who writes for our good buddies over at Pension Plan Puppets has, with the help of PPP commenter 'Frag', been slowly developing a new metric called dCorsi (Delta Corsi).

As you know by now, a player's Corsi number is calculated by weighing up the shot attempts for and against while he is on the ice at even strength. It is expressed as a % or in +/- form. A variation on this is Corsi Rel, which shows how his Corsi number or percentage relates to his teammates. For example, Johan Larsson has a Corsi % of 49.4%, which is negative. While not awful, that doesn't exactly look impressive. But when you take into account that he plays on the worst Corsi team in the league and look at his Corsi Rel, you see it is a team-leading +7.2%. So, while his 49.4% isn't that great by league standards, when you see his Corsi Rel it becomes clear that he is being held back by his teammates and is greatly out-performing them.

Another way of applying context to individual Corsi numbers is by looking at Quality Of Competition, Zone-Starts and Quality Of Teammates. Generally we look at these numbers, maybe on a Player Usage Chart, and make vague assessments like "that player has a very bad Corsi, but he plays really tough minutes so it's understandable" or "that player has a great Corsi but is very sheltered so take that with a pinch of salt".

While these statements aren't false, it would be better if we could take all these different contextual factors along with a player's Corsi and turn it into one number so we could really see which players are under-performing or out-performing the expectations set by the degree of difficulty of their role. This is what dCorsi effectively does. Or at least, that's my understanding of it. I recommend you read Steve's original post about it for a better explanation.

From that post, here's a basic description of how dCorsi is calculated:

So over the off-season I (with the help of Frag) slowly morphed what had been SDI into a Multi-Variate regression that allowed us to compare Actual Corsi Performance for a player to Expected Corsi Performance now known as dCorsi. The expected results were then output from a regression that factors in a players' avg TOI, Zone Starts, and Weighted Average Team-mate and Opposition Corsi Results (both For and Against).

Since that original post, Steve has worked-out the updated numbers and last night he posted a link to the Excel spreadsheet on his Twitter account, which I downloaded. So far he has only done the defencemen, setting the cut-off point at 200+ minutes played this season, which left 179 candidates.

-Here are the Wild defencemen who made the cut sorted by dCorsi For/20, dCorsi Against/20 and then cumulative dCorsi:


dCorsi For/20


Jared Spurgeon



Marco Scandella



Clayton Stoner



Ryan Suter



Jonas Brodin



  • What this tells us is that Spurgeon is out-performing his expected Corsi For by a wider margin than any other defencemen in the league. He beat-out Andrei Markov and P.K. Subban by a pretty wide margin at the top.
  • Brodin is the complete opposite, finishing dead last, with his Swedish buddy Oliver Ekman-Larsson, who is having a tough season, finishing 2nd last.


dCorsi Against/20


Clayton Stoner



Ryan Suter



Jonas Brodin



Marco Scandella



Jared Spurgeon



  • This tells us that Clayton Stoner is out-performing his defensive expectations better than all the other Wild defencemen. This means he is limiting Corsi events against very well relative to the minutes he plays.
  • This chart shows that all 5 defencemen are performing very close to their defensive expectations. Spurgeon, Scandella and Brodin are just below zero, while Stoner and Suter are only a bit above it.




Jared Spurgeon



Clayton Stoner



Marco Scandella



Ryan Suter



Jonas Brodin



  • Weighing up both the dCF20 and dCA20 against each other leaves the 5 Wild defencemen with the above rankings.
  • Spurgeon, Stoner and Scandella are all impressive, slotting into the top-60 in the league. Spurgeon finishing just inside the top-20 is great.
  • Suter and Brodin are both performing a fair bit below their expectations. Brodin in particular put-up rough numbers, finishing second to last overall, with only Slava Voynov finishing below.

This metric isn't perfect, but it should grow and develop over time and for now it's something interesting to look at. Hopefully it is available on sites like in the near future so we can track it during the season, rather than waiting for Steve to compile it every few months. Also, major thanks to Steve for doing all the work he put into this.

Follow me on Twitter for more talk about stats and opinions on hockey.