Explaining Advanced Stats: Zone Starts

Many of our readers may be aware of the “advanced stats” paradigm shift taking place in hockey analysis recently with statistics like Corsi and Fenwick becoming more mainstream for armchair analysts like us here at Stars and Sticks. Still, these statistics are relatively new and can take some getting used to. In part three of our series explaining some of the statistics we use to analyze the Caps and the rest of the league, we cover zone starts today and how it can be a useful tool in evaluating skaters along with some of the other stats we have covered this week.

When we’re talking about zone starts, we are generally evaluating individual players rather than team performance. They help give us context around how well a player is truly performing and can help illuminate his other numbers like Corsi and Fenwick stats.

By simple definition, zone starts simply tell us where a skater was on the ice when faceoffs occurred – offensive zone, neutral zone or defensive zone. Like Fenwick, it is typically expressed as a percentage. Here are the textbook definitions from stats.hockeyanalysis.com.

OZFO%OZFO% = Offensive Zone Face Off Percentage -percentage of faceoffs in offensive zone while on iceDZFO%DZFO% = Defensive Zone Face Off Percentage -percentage of faceoffs in defensive zone while on iceNZFO%NZFO% = Neutral Zone Face Off Percentage -percentage of faceoffs in neutral zone while on ice

How is it useful?

Let’s use a hypothetical example to illustrate how zone starts are a useful tool. Let’s suppose Defenseman A has an OZFO% of 80% during a single game, meaning 80% of the faceoffs that defenseman A was on the ice for during the game occurred in the offensive zone. Meanwhile, Defenseman B on the same team had a DZFO% OF 80%, meaning 80% of the faceoffs he was on the ice for were in the defensive zone.

Which player took on the toughest assignments? Obviously it was Defenseman B in this case, and that is likely to affect his stats such as Fenwick and Corsi since starting in your defensive zone makes it much more difficult to direct more shots at the opponent. Defenseman A in this scenario is far more likely to have better Corsi and Fenwick since his starts are mostly in the offensive zone.

More from Capitals News

When analysts talk about deployments and usage, a lot of that has to do with zone starts. In modern day hockey, coaches are well aware of optimal deployment strategies and will often give their best defensive players more difficult zone starts, while they might shelter players with better offensive skill who may be defensive liabilities by giving them favorable zone starts.

The stats community and analysts in general are pretty torn on how much zone starts affect a player’s possession stats. Tyler Dellow, the famous advanced stats blogger who was hired by the Edmonton Oilers this summer, evidently thinks zone starts matter quite a bit. Check out this post which breaks down Dellow’s piece (the original post was taken offline when he joined the Oilers).

Real Life Examples

Brooks Orpik is a perfect case study on why zone starts should be used in addition to other advanced stats to evaluate players. Brooks Orpik received a highly-controversial large contract from the Caps. Much maligned by the stats community, the contract seemed an overpay given Orpik’s weak Corsi and Fenwick stats over the last couple of seasons in Pittsburgh.

Those poor possession numbers have continued in D.C. this year, as Orpik is the worst defenseman in raw corsi percentage on the Caps so far and third worst overall skater on the team with 46.9%. He posted a similarly weak rating in Pittsburgh last year at 46.4%. And yet Orpik plays top-pair minutes for the Caps, perhaps the best defensive corps in the conference, and Barry Trotz trusts him enough to give him more ice time than any defenseman besides his partner, John Carlson.

GM Brian MacLellan thought highly enough of Orpik to give him a substantial contract in the offseason and even addressed the stats communities’ gripe with Orpik and mentioned that he’s aware of Orpik’s possession numbers but not too concerned. Why is that? MacLellan and Trotz have been around the game for a while and are pretty smart guys.

They and other hockey insiders are higher on Orpik than the analytics community in large part because he faces the other teams’ top lines and also because he takes the worst zone starts among his D partners. The Orpik-Carlson pair is starting less of their shifts in the offensive zone than any other D pair on the Caps. Conversely, Mike Green and Nate Schmidt no doubt make a wonderful third pair on defense, but start a lot more of their shifts in the offensive zone than Orpik or Carlson.

How Are the Caps Doing?

As previously mentioned, zone starts aren’t very useful as a team stat, so let’s look at how the individual skaters on the Caps are being deployed in terms of zone starts. The chart below from War-on-ice.com plots zone start percentage on the X axis (updated through 11/14/2014).

Mandatory Credit: War-On-Ice.com

The players who start less faceoffs in the offensive end are skewed to the left, and those with more offensive zone starts to the right. Jay Beagle has the least favorable zone starts of anyone on the Caps, meaning Trotz trusts him to take D-zone faceoffs. Jason Chimera is also given unfavorable zone starts as are the aforementioned Carlson and Orpik.

On the other hand, Green, Andre Burakovsky and Evgeny Kuznetsov are relatively sheltered and start a lot of their shifts in the offensive zone. Perhaps that is why Trotz scratched Kuznetsov last week against the Blackhawks. He may not have wanted to start both Burakovsky and Kuznetsov, two centers who aren’t trustworthy in their own end yet, against Chicago on the road.

Again, zone starts are just another tool we can use to evaluate players, one of many that should be used together to help us evaluate what happens on the ice.

Home / Capitals News