# "Zone-adjusted Corsi" and the Magnificent John-Michael Liles

Cam Charron
July 19 2012 02:47PM

I don't necessarily intend to use this space to write savage things about Jake Gardiner, but due to the limited number of young Toronto Maple Leafs' defencemen who can potentially crack a major league roster this season, expectations may be too high for the kid.

Grading offence isn't too hard to do. Usually, you can look at a player's goal totals and determine how good they are in the offensive zone. Defence is a little more suspect, however. How can you grade how good a player is defensively? You could use many things, to how many minutes a player is used on defence, or in what role a player is used by who his coach matches him up against.

It doesn't give you an exact answer, though. How good is somebody like, say, Jake Gardiner at playing defence? How successful does he, or will, he, make the Leafs?

Something statisticians, and now a growing number of online hockey writers, have been using is what's called a "Corsi" number. Corsi acts like plus/minus, with some differences. Here's an old post which goes through the basics:

Let's start with a simple definition. "Corsi" is the difference between all shots directed at net for and against at even strength. That is (shots+blocked shots+goals+missed shots FOR - SH+BLK+G+MS AGAINST). The purpose of the stat is to determine possession. It is, in fact, a proxy for "zone time". A positive corsi rate = more offensive zone time. Negative = more defensive zone time.

Here's an analogy that might help. Let's say a hockey game is a tug of war. Corsi is the how far right or left of center the rope is. On an individual level, it's an expression of which players are really pulling the rope. Therefore, if your team has a positive corsi rate, it means they are spending more time in the offensive zone at even stregth. It means they are pulling the rope harder than the opposition.

Corsi is tracked at a website called BehindTheNet.ca, but unfortunately, you can't just look at a player's Corsi number to determine his importance. It's an elephant gun, however. It's powerful, but imprecise. The reason shots are used is because there's far less variability year to year between shots than goals, so Corsi has a better predictive value than simple plus/minus.

Behind The Net calculates other important aspects of the game, such as which zone a player is more likely to start his shift. Obviously, players who start regularly in the defensive end are more likely to have a worse Corsi number.

Another thing it does is average the Corsi rate of opposing skaters to determine a player's quality of competition mark.

Using these three concepts, we can answer certain questions about Toronto's defence, and how successful they'll be next season.

 NAME GP TOI/60 ZAC Rel QoC GF GA John-Michael Liles 66 17.0 -0.8 -0.19 2.03 2.84 Mike Komisarek 45 14.7 -1.7 0.05 2.00 3.18 Dion Phaneuf 82 18.6 -2.4 1.40 2.21 2.80 Cody Franson 57 14.3 -3.2 -0.70 2.36 2.58 Carl Gunnarsson 76 17.2 -4.5 1.15 2.16 2.48 Jake Gardiner 75 17.1 -8.7 -0.33 2.95 2.99 Luke Schenn 79 14.3 -9.7 -0.15 2.71 3.18 Jay Rosehill 31 5.9 -10.1 -2.31 0.33 1.64

By accounting for the number of times a player started a shift in the defensive or offensive zone, I've adjusted the Corsi rates for the Leafs defence in an effort to answer questions about Gardiner, Mike Komisarek, and John-Michael Liles, the players that will mostly make up the Leafs' second pairing if nothing else changes. What is the best option for the Leafs to use if they want to continuously push play to the other end of the ice?

Well, this is what I found. "ZAC" is an acronym I use for "Zone-adjusted Corsi". It's a per 60 minutes rate. "Rel QoC" is the quality of competition faced by each player. Alongside the two, I looked at on-ice goals for and goals against per 60 minutes of each player to give us some indication of whether the player's plus/minus total synced up with Corsi.

Obviously, there are some surprises here. The Leafs gave up more goals with Jake Gardiner on the ice than anybody but Komisarek or the new Philadelphia Flyer Luke Schenn. Despite facing the second easiest competition on the Leafs, the team was hemmed in their own zone with Gardiner on the ice.

At the top, it's also surprising. Liles and Komisarek didn't play the tough minutes than Dion Phaneuf and Carl Gunnarsson played, but they were effective at moving the puck given their role. I think the Leafs need another guy in the top four who can play a plus Corsi game against decent competition, but I think it's clear that, if the season were to start tomorrow, Liles and Komisarek would be the second pairing.

As for Gardiner, there's still room to grow. He's an impressive skater with tuned offensive instincts. He needs to improve on his neutral zone play before the Leafs can be expected him to carry much of a load minutes-wise. The other thing is that Jay Rosehill should probably not play another game in the NHL, ever.

Cam Charron is a BC hockey fan that writes about hockey on many different websites including this one.

## Subscribe to our mailing list

* indicates required
#1 nhlcheapshot
July 19 2012, 03:16PM

Good to see some J.M. Liles love. Feel like I'm his last defender sometimes!

That being said if he plays like pre-concussion Liles instead of post-concussion Liles again that'd be great.

#2 Danny Gray
July 20 2012, 07:40AM

@nhlcheapshot

Ditto. I also wonder if Carlyle thinks he can "fix" Komisarek.

Obvs small sample but I wonder if he saw more ice under Carlyle. A Komi-Lile pairing could be not terrible.

#3 Not Norm Ullman
July 20 2012, 10:10AM

Nice piece, Cam. I think a couple of adjustments can help to focus things even further, bringing Liles squarely into the mix, but also, showing Gardiner much more clearly.

1. The chart shows Gardiner's Rel QofCompetition is -0.33.... so he's playing against weaker competition, absolutely.

But. He also had really weak team-mates, with a Relative QofTeam-mate even worse at -0.516. This was not only the weakest set of team-mates amongst Leaf defenders, and the 15th weakest set of team-mates of all defenders (over 70 games) in the entire League.

Gardiner -.0.516 Dion -0.043 Liles +0.212 Schenn +0.257 Franson +0.389 Komisarek +0.682 Gunnar +0.730

Besides helping to explain Gardiner's poor defensive results, this also shows an interesting positive fact. Namely, that when Gardiner was on the ice, the Leafs scored far more goals than when any other defender was on. And he did so, even with weak-ass team-mates.

No other NHL defender scored as well as Gardiner with such weak team-mates. Also, none scored higher while having team-mates who were weaker than the competition, as did Gardiner.

http://www.behindthenet.ca/nhl_statistics.php?ds=41&s=23&f1=2011_s&f2=5v5&f4=D&f7=70-&c=0+1+3+5+2+4+6+7+10+8+11+12+13+14+15+16+17+18+19+20+21+22+23+24+25+26+27+28+29+30+31+32+33+34+63+67+57+58+59+60+61+62+64+65+66+41+42+43+44+45+46

You can see how weak Gardiner's team-mates were (-0.516) as compared to the QofT of some other high-scoring defencemen:

Karlsson +5.780 White +11.424 Keith +9.968 Lidstrom +11.463

[These measures also show you how strong bloody Shea Weber was, scoring at a 2.74 rate, against strong +2.015 competition, and with weaker team-mates -6.090.]

So I think perhaps if your zone-adjusted comparison is to include QofC, it might usefully also include QofT. The net result would be to also improve Dion's numbers, which makes sense to me - and Liles would gain a bit, relative to the remaining 4.

2. It might also be useful to take a look at not just 5v5, but at 5v5 Close. Quite simply, over time, it became obvious how Wilson was playing Gardiner. As the year went along, it became clear that - offensively - Gardiner was 30% higher-scoring than the first pairing, Dion and Gunnar, in terms of offence 5v5. And so Wilson was gonna play him when they fell behind. In fact, in games where they fell behind, Wilson would ride him, double-shifting him again and again, trying to play catch up... but with the inevitable results in terms of Gardiner's goals against.

Sounds like just a narrative, and I couldn't find the Corsi numbers for 5v5 Close, but I found the goals scored per minute at 5v5 Close, and they're zone adjusted. They actually show an interesting story (or two) in terms of how Gardiner played when the game was close, versus when he was being double-shifted to try and jump-start a comeback.

http://stats.hockeyanalysis.com/ratings.php?disp=1&db=201112&sit=5v5close_f10&pos=defense&minutes=500&type=goal&sort=F20&sortdir=DESC

1st, they show that Gardiner was almost TWICE as good on offence in these situations as Dion, Gunnar or Liles.

Gardiner 1.165 ---> the 2nd best in NHL for defenders >500 minutes Schenn 1.03 Gunnar .694 Dion .693 Liles .559

And to compare with the rest of the league:

Keith 1.107 Karlsson 1.105 Lidstrom .966 Chara .900 Weber .858 Doughty .661

So, against what some might perceive, Gardiner wasn't just helping the Leafs score because he was up against weak compeition, and had positive zone starts. Though those things are true. But so too are the facts that he had weak team-mates, and that he helped them score when the games mattered, when they were close.

But then this stat also adds another fact to the picture. In terms of his DEFENCE, in these 5v5 Close, zone adjusted situations, Gardiner was better than Gunnar, and not that far from Dion (the best defender.) And in a league-wide comparison, he also did quite well - better than many big names.

So what we're seeing here is that MOST of the seeming failure of Gardiner as a defenceman came in situations where the Leafs were trailing, and he was being called on to try and salvage the situation - i.e. play run and gun.

Here are the figures, goals against per 20, 5v5, Close, zone adjusted:

Dion .773 Liles .804 Gardiner .858 Gunnar .875 Schenn 1.163 (3rd worst of NHL defenders playing >500 such minutes)

Keith 1.107 Karlsson .903 Keith .992 Chara .852 Lidstrom .773 Doughty .635 Weber .572

So, again, on this sort of measure, Dion gains a bit, as does Liles, while Gardiner gains a lot, Gunnar weakens a notch, and Schenn free-falls.

To conclude, absolutely, Gardiner has to improve on defence. And absolutely, he got a boost from positive zone starts, and weak competition.

But. He did so with weak team-mates. And when the game was close, he scored like crazy, while playing not bad defnce. It was when games entered their late stages, and the Leafs were well behind, and it fell to him to try and run and gun, that more goals against got scored.

#4 Danny Gray
July 20 2012, 10:41AM

@Not Norm Ullman

Holy Hell Not Norm. What a comment

#5 Not Norm Ullman
July 20 2012, 01:16PM

@ Danny Gray

GF's away... got more time to waste playing with hockey stats. Woo hoo?

#6 David Johnson
July 20 2012, 03:41PM

"Sounds like just a narrative, and I couldn't find the Corsi numbers for 5v5 Close, but I found the goals scored per minute at 5v5 Close, and they're zone adjusted."

The corsi (and and fenwick and shot) numbers are at stats.hockeyanalysis.com. You can either do the search or if you are at the goal link just change goal in the url to corsi. So

http://stats.hockeyanalysis.com/ratings.php?disp=1&db=201112&sit=5v5close_f10&pos=defense&minutes=500&type=goal&sort=F20&sortdir=DESC

becomes

http://stats.hockeyanalysis.com/ratings.php?disp=1&db=201112&sit=5v5close_f10&pos=defense&minutes=500&type=corsi&sort=F20&sortdir=DESC

or you can use "shots" or "fenwick" depending on what you need.

#7 Not Norm Ullman
July 20 2012, 04:07PM

@ David Johnston

Thanks ! From the Corsi tables (not sure I'm catching it all) it looks like he faced moderate competition, had weak team-mates, did quite well on offense, and relatively poorly on defence (though no Leaf defence did really well.) I guess we'll see whether his stronger results in terms of Goals versus Corsi was luck next season!

#8 David Johnson
July 20 2012, 05:23PM

@Not Norm Ullman

If you want to get an idea of his QoC sort by OppCorF%.

http://stats.hockeyanalysis.com/ratings.php?disp=1&db=201112&sit=5v5close_f10&pos=defense&minutes=500&type=corsi&sort=OPPPCT&sortdir=DESC

His QoC with that ranking is 103 of 140 so bottom third competition. His OppGF% ranks him 117th of 140 as well so definitely not tough QoC. That said, I feel QoC is significantly over rated. More important, IMO, is his QoT. His TMGF% ranks 138th of 140 and his TMCorF% is 106th of 140. So he hasn't played against the toughest competition, but he hasn't played with very good teammates either.

#9 Top Shelf
July 21 2012, 11:15PM

@Not Norm Ullman

Awesome, so....

2nd pairing Defender's Gardiner and Franson? Liles and Komisarek (Holzer?) 3rd pairing?