Minnesota Wild fans: there are only two ways that stat nerds will stop trash-talking your team. Either they stop winning or they win everything. Just embrace it.

Last week I spent some time putting the NHL’s top scorers in context using a few advanced statistics. One of those statistics was zone starts, which has been shown to have a profound influence on an individual player’s results. Understanding zone starts is simple: you just need to know who is on the ice during a faceoff and whether that faceoff is in the offensive, neutral, or defensive zone.

I enjoy looking at zone starts because they make a certain amount of intuitive sense even to those not well-versed in advanced statistics: of course players who start in the offensive zone tend to get more offensive opportunities. Of course players who start the majority of their shifts in the defensive zone will face more shots from the opposition and struggle to make an impact offensively. These truths are, as it were, self-evident.

What I’m curious about is team zone starts, which doesn’t seem to have been investigated as much as individual zone starts. Which teams tend to force faceoffs in the offensive zone? Which teams end up having to take a lot of defensive zone draws? How does a team’s overall zone starts impact an individual player’s zone starts?

The team that most radically uses zone starts to their advantage is the Vancouver Canucks. The top three players in OZone% (offensive zone starts vs defensive zone starts) are Daniel Sedin, Henrik Sedin, and Alex Burrows, the Canucks’ top line. The bottom four players are also Canucks: Manny Malhotra, Dale Weise, Aaron Volpatti, and Maxim Lapierre. Alain Vigneault consistently sends his top offensive players out to start in the offensive zone, giving them a head start on creating offensive opportunities. In order to do this, he buries his fourth line, which is as defensively responsible as many team’s third lines.

The only faceoff Henrik Sedin took all night that wasn't in the offensive zone.

If you’re wondering how important all this really is, I point you to this article from a couple years ago by Gabriel Desjardins of BehindtheNet.ca, where he illustrates that “when you lose a faceoff in your own end, opponent shots on goal go up so quickly that it’s as though you gave the other team a 10-15 second power-play. For several seconds, the rate of shots allowed is as high as it is on a 5-on-3.”

The opposite is also true: a won faceoff in the offensive zone is equivalent to a brief powerplay. Given how deadly the Sedins are on the powerplay (1st and 5th in powerplay points last season), it should be clear that giving them significantly more offensive zone starts and the equivalent of 10-15 seconds of powerplay time on every won offensive zone faceoff would have a significant impact on their point totals.

I have seen it asked, however, whether certain players see more starts in the offensive zone than in the defensive zone because their team simply starts in the offensive zone more often. A team’s checking line, for instance, is frequently praised when they can start in the defensive zone and create an offensive zone faceoff for the team’s top line. Teams that are able to create more offensive zone faceoffs create more offensive opportunities. Of course, the opposite is also true: teams that create more offensive opportunities will create more offensive zone faceoffs by forcing a goaltender to make saves and freeze the puck.

I started wondering about team zone starts and whether some of these questions could be answered and, well, there’s no easy way to say this, but you’re about to be slapped in the face by a table:

Team Zone Starts at 5-on-5

DET 407 295 442 1144 48.6% 57.6% 58.0%
OTT 436.8 360 527 1323.8 51.5% 50.6% 54.8%
CHI 436.6 367 554.2 1357.8 57.4% 49.9% 54.3%
S.J 399.4 350.2 445.4 1195 53.9% 50.3% 53.3%
PIT 432.8 385.2 531.6 1349.6 51.2% 53.9% 52.9%
BUF 413.6 383 486.2 1282.8 49.4% 50.4% 51.9%
STL 403.4 377 471.4 1251.8 50.4% 49.9% 51.7%
TOR 407.8 382 501 1290.8 48.7% 49.7% 51.6%
PHX 412 386.2 487.8 1286 51.4% 50.1% 51.6%
EDM 379.8 359.2 486.2 1225.2 43.7% 47.1% 51.4%
MTL 413.2 392 535.8 1341 47.0% 53.4% 51.3%
PHI 354.4 337.2 464.4 1156 44.5% 46.8% 51.2%
L.A 367 354 446.4 1167.4 49.6% 55.6% 50.9%
VAN 378.6 367 450.2 1195.8 52.2% 50.5% 50.8%
COL 405.6 394 506.4 1306 52.0% 47.2% 50.7%
WSH 369.4 362 468.2 1199.6 46.9% 53.0% 50.5%
WPG 371 365 483.2 1219.2 41.8% 51.3% 50.4%
CGY 406.6 403.2 509.8 1319.6 48.1% 42.9% 50.2%
FLA 370.6 367.6 536.6 1274.8 46.5% 50.6% 50.2%
BOS 382.4 380 469.2 1231.6 56.6% 57.6% 50.2%
N.J 353 350.8 436.8 1140.6 47.0% 50.1% 50.2%
NYI 348.6 356 421.8 1126.4 51.5% 50.8% 49.5%
ANA 330 339.8 499 1168.8 46.7% 47.9% 49.3%
CBJ 399.4 430.6 485.2 1315.2 52.9% 51.1% 48.1%
T.B 358 392 528.2 1278.2 47.1% 48.2% 47.7%
DAL 340.2 405.6 457.4 1203.2 44.6% 51.2% 45.6%
CAR 371.8 448.8 523.4 1344 48.9% 50.7% 45.3%
NYR 311.4 378 424 1113.4 49.9% 53.2% 45.2%
NSH 376.2 466 493.6 1335.8 51.2% 45.9% 44.7%
MIN 321.6 480 457.2 1258.8 46.7% 52.9% 40.1%

All of this data was mined from Behindthenet.ca using the individual players to tabulate the team zone starts. This is why you see the oddity of fractions of faceoffs as the NHL’s data from which Gabe pulls these statistics can sometimes have minor errors. Fortunately, it’s not enough to make the results invalid. All of the data is at 5-on-5, so no powerplay, penalty kill, or 4-on-4 play is included.

Some quick observations can be made:

Has anyone else ever made the observation that the Red Wings are really good at hockey? No? I'm the first? Sweet.

  • The Detroit Red Wings, well known as a possession powerhouse, comes in at the top of the chart with an OZone% of 58%. Not only do they rarely have defensive zone faceoffs, they win more of them than any other team in the league other than Boston, as they both have a 57.6 faceoff percentage in the defensive zone at 5-on-5, preventing their opponents from creating offensive opportunities when they do manage to force a defensive zone faceoff.
  • Last week I was checking each team’s OZone% and noticed that the Red Wings were the only team in the league with no players below 50% in OZone%. That’s no longer true as Jiri Hudler, Valtteri Filppula, and Chris Conner now fall just below 50%, but their high team OZone% explains why the vast majority of their players start more often in the offensive zone than the defensive zone. The entire Red Wings team seems to create more offensive zone faceoffs than defensive zone faceoffs with their puck possession game. Either that or Jimmy Howard never covers up the puck.
  • This is yet another way to show that the Minnesota Wild are struggling when it comes to puck possession: their OZone% is worst in the NHL by over 4%. Even more stark is that they are the only team in the league who have started more shifts at 5-on-5 in the defensive zone than in the neutral zone. To their credit, they win 52.9% of their defensive zone faceoffs, but unless they are able to start shifting possession into the offensive zone more regularly, their league-high defensive zone starts will come back to bite them.
  • The Canucks are just about an even 50% in offensive zone to defensive zone starts, showing that the Sedins’ abnormally high OZone% is entirely dependent on how they are used by coach Vigneault and not because the Canucks force more offensive zone faceoffs.
  • A big reason for the Boston Bruins’ league-best even-strength goal differential would appear to be their faceoff percentage in the offensive and defensive zones at even-strength, which is higher than their overall faceoff percentage. Though their number of offensive and defensive zone starts is about even, they make the most of their offensive faceoffs with the second highest offensive zone faceoff percentage in the league and limit their opponents’ opportunities with the highest defensive zone faceoff percentage (tied with Detroit).
  • The team with the highest offensive zone faceoff percentage is the Chicago Blackhawks at 57.4%. They also have the third highest OZone%, so they frequently start shifts in the offensive zone with possession. It seems likely that this is part of the reason they lead the league in goals-for at even-strength, but it might be a little early to draw that conclusion.
  • One team that I did not expect to see so high in OZone% is the Ottawa Senators, who are currently last in the Northeast Division and 12th in the Eastern Conference. The issue hasn’t been scoring goals for the Senators, who are currently 5th in even-strength scoring, it’s been preventing them. The Senators have allowed the second-most goals at even-strength and it seems fair to place a little bit of the blame on their goaltending. Craig Anderson has one of the worst save percentages in the league at even-strength. Actually, that might explain why they have so few defensive zone starts: Anderson hasn’t been making enough saves.

I feel like this is just scratching the surface of the analysis that could be done with team zone starts. For instance, it would be interesting to see how well zone starts correlates with Corsi and Fenwick. Unfortunately, I don’t have those numbers handy for teams, just for individual players. It would also be interesting to see how score effects impact zone starts: when teams are down by a goal or two they tend to shoot more and by consequence have more faceoffs in the offensive zone. Simultaneously, when a team is ahead they tend to sit back and not press play into the offensive zone and thus end up with more faceoffs in the defensive zone.

As a preliminary look at team zone starts, however, this may still prove useful to some of you. Do any other interesting results jump out at you from the table of team zone starts?

Comments (24)

  1. Good stuff DW.

    I’ve been experimenting with examining usage %, ie: the total number of offensive (or defensive) zone starts versus the total number of available faceoffs.

    Tomas Holmstrom, as an example, has a Ozone rate of roughly 70%, but that isn’t all that much higher than the team’s average. He’s started on the ice for 24% of all of Detroit’s team faceoffs in the offensive zone.

    Hopkins’ “start” rate is lower than Holmstrom’s (68.5%), but he starts in the offensive zone for 32% of all of Edmonton’s offensive opportunities. Why is this? Because Edmonton gets fewer available offensive zone starts.

  2. Seriously Bourne, your writers are idiots.

    There is an r-squared correlation of 0.88% between offensive zone starts and win percentage. I took your table, I took win percentage, and there is literally no correlation. None. 0.88% is about as close to zero as it gets.

    I’m done reading this blog if you keep putting out trash like this and call it journalism. Get some stats guys that know what they’re doing rather than coming up with random theories and throwing it against the wall without even checking to see if their statistics are worth using.

    • It’s okay, you can call me an idiot directly if you like.

    • He’s making the point that some teams are not as good as their win records seem to make them out to be. The table shows that teams like MIN and NYR need to put more pressure on in the offensive zone in order to keep play there more often, or else they will begin to lose more often.

      Just because it’s a different way of looking at it doesn’t make him an idiot.

      Further, a 0.88% R^2 doesn’t make sense. R^2 values aren’t even given in percentages – the value goes from 0 to 1 generally. So if you mean that R^2= 0.88, that’s actually very high. Unless you meant that R^2=0.0088 = 0.88%, in which case, a) you’re the idiot for writing it that way, and b) you’ve just show that they don’t correlate FOR THIS SEASON. How about giving us some correlations for last year? Or the last ten years? Then we might know how closely they actually correlate.

    • R^2 is a coefficient, not a percentage. .88 is very high.

    • Even ignoring the fact that “r-squared correlation of 0.88%” makes absolutely no sense as a statement, the point of methods like these is to predict future winning percentage; to separate unsustainable, luck-driven elements of team success from those that are within a team’s control, such as moving the puck forward and generating possession which are measured reasonably well by zone starts.

  3. Nice article, and good analysis. I’m interested to see what else you can come up with using those stats. The link to the behindthenet.ca article was really interesting too.

    However, as a Rangers fan, I have to note that this (obviously) doesn’t take into consideration when play is pushed into the offensive zone without the use of a faceoff. The Rangers have a very strong forecheck, and very aggressive, which to me partially explains the season they’re having now. Just because they aren’t starting in the offensive zone doesn’t mean they aren’t pushing into it, and keeping it in there, even if it’s just boxing their opponents players inside. Are there stats for how long the puck is any one team’s offensive zone, regardless of who controls it? That might help show how strong a forecheck a team has.

    • They have the fourth worst Fenwick Tied ratio in the league at 0.458 (although they’re trending upward in that regard), so their zone start differential, at least at this point, is probably a pretty good proxy for their possession ability at 5-on-5.

  4. danielson

    i think your third last point paragraph uses “offensive” once when you mean “defensive.”

  5. This article reminds me of a scenario that played out in the youth hockey game I was reffing this past weekend. While Team A appeared to be carrying the majority of the play, Team B had scored 4 unanswered to overcome an early 1-0 deficit. A team A player, presumably trying to fire his guys up, loudly exclaimed the following: “C’mon, let’s go! These guys SUCK!”

    My internal filter- keep in mind these were just PeeWees- prevented me from saying the first thought that came to mind. “Yeah? Well SUCK is kicking your @$$ 4-1 right now.”

  6. Looking at the Wild’s numbers at behindthenet.ca, something does stand out to me. Regardless The top three lines all are around 50% Fin Ozone% regardless that they are starting in the ozone at 35-40 % of the time. To me that suggests that regardless of where they are starting, they are pushing the puck up ice and into the offensive zone fine which gives a little bit of credibility to the idea that the system of moving the puck and forechecking hard is working. But I have no idea if any studies have been done on this type of thing and what it signifies.


  7. Daniel,

    This is great stuff. But to make your job easier and to get rid of those fractions of zone starts, use timeonice instead. For example, Detroit’s numbers so far this season:


  8. this site needs more stat writers. you only have two.

  9. Honestly, as a Wild fan, the stat nerds can trash them all they like because, at the end of the day they keep finding a way to win and that’s the only stat that matters.

    Do the Wild play in their own zone more often than the offensive zone? Yeah. Do they start in their own zone more often than the offensive zone? Sure. You don’t need stats to see that. All you need to do is watch one of their games. The stat heads are pushing out these numbers at everybody and thinking that they’re revelations to Wild fans when, in reality, they’re not. We know that there’s no statistical explanation as to why the Wild keep winning.

    Stats aren’t the end all, be all of why teams are or aren’t good. Sometimes they explain it (like, for instance, the Kings — they’re not good because they can’t score), but other times it’s like this when you keep giving statistics that show that the Wild really aren’t a good team, but they keep winning in spite of that. (Which, really, doesn’t that mean that they are a good team because they keep finding ways to win, even when the stats show that they shouldn’t?)

    The answer to why teams do and don’t succeed aren’t always found in the stats. Sure, it can give you an insight into how the team plays, but it’s not going to always tell you whether or not a team is a good team.

    Like I said…You can keep pouring out all of these statistical reasons why the Wild shouldn’t be the top team in the league, but none of these stats are going to be revelations to Wild fans. We know that the Wild doesn’t play the prettiest style of hockey. We know that the Wild are overachieving right now and I don’t think that there’s any Wild fan out there that doesn’t expect them to regress at least a little at some point in time. But I can explain to you why the Wild are winning right now, and it has nothing to do with stats:

    They’re getting timely scoring, they’re getting great goaltending, they’re playing great team defense and they’re playing smart, team hockey. Just watch a couple games of the Wild’s and you’ll see that, even though they spend a lot of time in their own zone, they’re never panicking and they’re not making a whole lot of mistakes in their own zone. That’s why they’re winning.

    • We can debate all we want about why the Wild have been winning to this point. The issue is whether they’ll keep winning. Stuff like goaltending and timely scoring is very volatile. Possession numbers, not nearly as much.

      I think it’s fair to equate “good team” with “team that will win the most in the near future.”


    • That’s why I included the caption on the picture. There are two ways that stat nerds will shut up about the Minnesota Wild: if they regress to the mean and stop winning they/we will say “See, we told you so” then shut up about it. Or, if the Wild keep winning, going far into the playoffs and continuing their success in the following season.

      So, keep cheering for your team, hope that their style of play can overcome all the advanced statistics that say they shouldn’t be succeeding. Embrace that your team is somehow an underdog despite having one of the best records in the NHL. It’s actually a pretty fun place to be.

  10. 1) The Wild are not as good as their record indicates due to taking a multitude of in zone faceoffs, which lead to goals against and eventually more losses (the point of the article)

    2) Taking a high amount in zone faceoffs does not matter all that much to a team that is good at blocking shots and has solid goaltending. Look at the Wild and Rangers who are at the bottom of the list but near the top of the league, what do they have in common?

    Just two ways to interpret the data.

    • You’re right, the Wild and Rangers are both blocking a lot of shots and getting great goaltending. The issue is that it’s unlikely to last forever. If Josh Harding and Niklas Backstrom continue to play as well as they are playing right now, the Wild are in great shape. It just doesn’t seem all that likely. Even Lundqvist seems unlikely to continue his current .932 save percentage as his career-high is .923. The safer option for continued success is puck possession.

Leave a Reply

Your email address will not be published. Required fields are marked *