Next up in our series of Advanced Stats for Dummies (last was Corsi For Percentage) we are taking on a stat that helps to understand what a player does best. It’s called “Zone Starts” and it measures the number of faceoffs a player is on the ice for in the offensive or defensive zone.

Zone Starts are calculated using a very simple procedure. Any time a player is on the ice for a faceoff in the offensive or defensive zone he is scored with a Zone Start. (Neutral zone faceoffs are ignored). If the player is on the ice for a faceoff in the offensive zone, it goes for an offensive Zone Start, if his team is in the defensive zone, he gets a defensive zone start. Zone Starts are then measured by percentage.

Offensive Zone Start Percentage (oZS%) = Offensive Zone Starts / Total Zone Starts
Defensive Zone Start Percentage (dZS%) = Defensive Zone Starts / Total Zone Starts

Example time! Let’s use David Perron because he tweeted at us one time and it made us blush. Perron starts the game on the ice, the opening faceoff is taken (no Zone Start scored). The puck is iced and the Golden Knights get a faceoff in the offensive zone (+1 oZS). Shot goes into the netting, another offensive zone faceoff (+1 oZS). Play continues, the puck goes out of play in the Golden Knights zone, Perron stays on the ice for the faceoff in the defensive zone (+1 dZS).

Offensive Zone Starts (oZS) – 2
Defensive Zone Starts (dZS) – 1
Offensive Zone Start Percentage (oZS%) – 2/(2+1) = 2/3 = 66.6%
Defensive Zone Start Percentage (dZS%) – 1/(2+1) = 1/3 = 33.3%

Perron had an oZS% of 43.9% last season and a dZS% of 56.1. The two years he played in Pittsburgh though, his oZS% was 60% and his dZS% was 40%.

As you can see, you really only need to focus on one side or the other as oZS% and dZS% offset. (43.9+56.1=100 / 60+40=100)

Here’s a look at the Golden Knights top oZS% players. First forwards.

PlayerOffensive Zone Start Percentage (oZS%)
James Neal59.6%
Jonathan Marchessault53.1%
David Perron53.1%
Oscar Lindberg52.3%
William Carrier51.5%

Now defensemen.