Playing with Percentages

For those who used to be a fan of Stats.HockeyAnalysis.com, for just under 2 years now, they’ve been moving their stats to a more user-friendly and constantly-improving site called Puckalytics.

A very brief bit of hockey analytics history:

The hockey analytics community is surely thankful for the work David Johnson did and continues to do with that site. The era of ExtraSkater came and went as founder, Daryl Metcalf was hired by the Maple Leafs. The heir apparent was War-on-Ice which was run by A.C. Thomas and Sam Ventura. In their wake, Corsica has emerged as the preeminent advanced stat choice. And it feels like long after they have come and gone, David Johnson will still be chugging on.

So now, for what the article is about — percentages. Percentages are great because they are something that everyone that passed pre-algebra can understand. Long before the advanced analytics era, the NFL used completion percentage, the MLB used batting average, the NBA used FG% and FT%, and the NHL used save and shooting percentages. Even as we moved into advanced analytics we stuck to that model as the oft-referenced Corsi (or SAT for you NHL.com people) is a percentage stat. But then you get into the “relative” stats. These are not complicated computationally, but they can be laborious to describe. That’s why the “%ofTeam” statistics are helpful.

These statistics show similar data to the “relative” statistics but are a little less abstract. The table below from the Puckalytics Glossary page says it all:

They are very straightforward statistics that tell you very important information about the players importance, achievement, and usage on the team. I’m going to point out a few notables here. The only possible misconception I could see is people thinking these are individual statistics rather than on-ice statistics. So remember, %ofTeamSF does not mean the player took that percent of the shots, it means that is the percent of shots taken while that player is on the ice.

Percent of Team Goals For

Michael Cammelleri: 56.14% (#1 in the NHL)

Again that means that in the 42 games he played, he was on the ice for 56.14% of the goals the Devils scored. Patrick Kane was 2nd among forwards in this statistic with 50%. This is hardly surprising to anyone who watched Devils games early last year, but it is still remarkable. I also wanted to note it because people may have forgotten just how important he was to this team before his injury. If he is healthy, he has every opportunity to lead this team in points even with Taylor Hall on the roster and Adam Henrique and Kyle Palmieri seeming to continue to improve. He was the offense before he was injured and it will be interesting to see if he can pick up where he left off.

Adam Henrique: 43.93% (#7 among forwards, #24 overall)

Henrique benefited from being part of the SCAR line with Cam and Lee Stempniak. But, he obviously did his share with a 30-goal season. whats just as impressive is that he was only on the ice for 29% of our goals against to Cammalleri’s 33% despite significantly more difficult zone starts.

Taylor Hall: 42.86% (#11 among forwards #30 overall)

Just in case you forgot, he’s good too. Although he did finish behind Connor McDavid (46.67%, 3rd in NHL).

Womp: Jacob Josefson 6.94% (5th worst in NHL), Jordin Tootoo 7.06% (6th worst)

Only goes to prove that JJ is a Special Teams Specialist, and Jordin Tootoo is just bad.

Others noteworthy: Stefan Matteau (11th worst among forwards), David Clarkson (18th worst among forwards), Jaromir Jagr (8th among forwards).

Percent of Team Ice Time

Larsson and Greene were outside the top 20 overall. So the only noteworthy ones are:

Michael Cammelleri: 30.82% (#14 among forwards)

First of all, that explains a little of why all those goals were scored with him on the ice. He was on the ice a lot. Maybe we cut back on that a bit this year? A lot of people think that the Hall, Henrique, Palmieri line will be the #1 and dominate ice time, but Cam will make sure you don’t forget him. That being said ...

Taylor Hall: 31.88% (#4 among forwards)

He’s a workhorse. He lead the Oilers fowards in ice time and will likely lead ours. He’s young, fit, and strong at both ends.

Ode to Larsson/Greene

Adam Larsson and Andy Greene. This is probably going to be my last chance to compliment them and I know I’ve said a lot about them but I’m gonna do it one more time because this pairing cannot be given enough praise for what they did. My first reaction after I heard about the Hall trade was not unqualified elation as it was for some. You’ll see why here.

Adam Larsson and Andy Greene played 55.46% and 54.96% of our defensive zone faceoffs respectively. That’s #1 and #2 in the NHL. No other defender had over 50%. But what good is that right? So they were put in tough situations a lot. The crazy part is that that’s not the crazy part. Despite that, they were only on the ice for 25.19% and 26.72% of the goals scored against us. Go this link right now and try to find the next defensive zone specialists that were that successful. It’s sorted by % of Team Defensive Zone Facoffs so the further you scroll, the less starts they get in their own end. Seriously, go. I’ll wait...

...

...

...

Find it? No not yet? How far did you scroll? Oh no you gotta go further than that.

...

...

Got it? Yeah you have to go to all the way to page 5 before you get to Anton Stralman. The thing is, that’s not remotely comparable because he was only in 35% of the defensive zone faceoffs. In fact, he was actually in more offensive zone faceoffs. No on in the NHL was anywhere remotely in the vicinity of this pairing when it came to getting their team out of a jam. There is no amount of hyperbole to describe how statistically ridiculous it is that Adam Larsson was 1st out of 240 qualified defenders in percent of defensive zone starts, but 205th out of 240 in percent of goals against.

Summary

Cammalleri’s really good. Hall’s really good. Larsson/Greene was really good. And David Johnson is really good. But he’s good at stats, not sports. I mean he might be good at sports, that’s just not what I’m saying here. Don’t everybody go telling David Johnson that I said he sucks at sports now okay, you know that’s not what I meant.

Feedback

What did you guys think of the percent stats? Are they easier or harder to understand? Do you think they are useful? What do you think about the players mentioned? Do you think Cammalleri is going to be our best player again or will Hall take the throne? How badly will we miss Larsson and Greene? Do you think I need to get over them already? Shut up and let me mourn!

Except don’t shut up. Instead leave your comments below.