Update: Usage Adjusted Possession Numbers

Some interesting nuances introduced yesterday with respect to the way we calculate zone starts by @arikparknass. The same points were made in this piece by @hockeyanalysis which was posted a few weeks ago.

The point made was that we should only be counting faceoffs for individual skaters when they were off the ice prior to the draw. Why? Because any faceoff that occurs afterwards is influenced by the possession game of the player of interest. Poor possession players are more likely to be in their own end and have their goaltender cover the puck. They then remain on the ice for the ensuing faceoff and their OZS% decreases. Zone starts, the way they are calculated at the moment, are NOT independent from possession.

In light of all of this, I will be offering usage adjusted possession numbers that do not adjust for zone starts in addition to the usage adjusted metrics already available here at OwnThePuck. The R^2 between U.A. Corsi% (-ZS) and U.A. Corsi% (+ZS) is 0.9239 (400+ min).

Click here to view the new tableau host containing both Usage Adjusted Corsi% (+ZS) and Usage Adjusted Corsi% (-ZS) data dating back to the 2007-08 regular season.

Advertisements

2 thoughts on “Update: Usage Adjusted Possession Numbers

  1. Having thought about this more we really need to be subtracting zone finishes from zone starts before conducting any ZS adjustments. We should also be including shifts that start/end on the fly in the equation. Regardless, I think the conclusion will remain that zone starts are not a significant factor in most players stats.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s