Comments / New

Panther Problems: A look at some troubling advanced statistics

We speak here at LBC of advanced statistics sometimes with reverence, and sometimes with criticism. The numbers-whether you agree with them or not- were worrisome prior to the Florida Panthers game on October 30th against the Boston Bruins, when this current losing streak kicked off. At that time the Panthers had one of the highest PDO’s in the league, (second to Montreal) which typically means that there is a significant amount of puck luck going a team’s way. That number usually regresses to an average, and the Cats have certainly experienced a shift to “below average” in that respect since that time. But why the sudden debacles? Why the sudden plethora of shots against, along with a total lack of scoring? What turned on this team, aside from an injury to one first line center?

Some player metrics look very good even now. But are those numbers telling the whole story, or even an important part of it? A recent correspondence between former staffer Shane O’Donnell and myself on Twitter in regards to Aaron Ekblad and his numbers highlighted that stats can be looked at and used in a multitude of ways. I will address more on that conversation with Shane in the comments section to this article, as it’s not altogether relevant to what I am looking at in this piece. But that look at Ekblad also brought to my attention some things that are not readily apparent, but need to be understood as the Panthers try to dig their way out of the hole they have dug. The things I started looking at were zone starts, zone finishes, and quality of competition metrics and the results are eye-raising.

The first thing we need to take a look at is zone usage. What is zone usage? Very simply: where is a coach deploying his players when given the opportunity to do so after a stoppage of play? That usage has an important impact on the player’s possession metrics. To quote from hockey analysis.com:

if you are doing a corsi/fenwick/shot/shooting percentage based analysis accounting for zone starts is really important because it can have significant impacts on these stats (less so for ratios though).

Why is this? It’s as simple as it sounds- players starting shifts in the offensive zone have an immediate opportunity (if a face-off is won) to get shots off- which possession metrics use as key measurements. Further, starting in the offensive zone means the opponent must complete a breakout, a neutral zone transition, and an offensive zone entry before a negative Corsi event– a shot against- can be recorded. Obviously, players starting in the defensive zone do not have that luxury and are more likely to suffer shots against, which equate to negative Corsi events. Players who start more often in the defensive zone will have more damaged Corsi results.

We could look to a metric called: D-Corsi. I don’t typically use that metric. D-Corsi is a rather complicated metric to me, that attempts to roll a lot of different factors into one stat. My personal opinion is that D-Corsi is overly complicated, and still lacks certain critical information. The most critical element that is lacking is what 5-man unit a player is playing with. For instance, Mitchell and Gudbranson are often being used as a 5-man unit with the first line. We know that the first line is not defensively gifted and their lack of defensive ability is putting extra pressure on that defensive pairing. This is insufficiently accounted for in most metrics. If you are looking for what D-Corsi is, here is the explanation in all its glory. Keeping things relatively simple by using zone starts provides a lot of very good data. Consider this, from NHLnumbers.com:

We are farthest along at quantifying the impact of where a player starts his shifts. Five years ago, Vic Ferrari applied a quantitative correction to possession numbers based on faceoff locations, and since then a lot ofdifferent approaches have all come up with similar quantitative reults. The most widely-used approach assumes that each offensive zone start is worth an extra 0.8 shots on the player’s Corsi shot differential (which is like +/- but using shots rather than goals, and correlates strongly to puck possession and offensive zone time).

Using that approach, an average player who started in the offensive zone as often as Daniel Sedin did last year would have been expected to see his team get 56.1% of the shots, while an average player put in Manny Malhotra’s skates would see his team get just 39.2% of the shots. With differences that large, we can understand why it would be important to correct for a player’s zone starts.

Looking at the Panthers offensive zone starts, on defense, Aaron Ekblad and Brian Campbell are getting even greater offensive zone use than last season, and Willie Mitchell and Erik Gudbranson are still being used heavily on the defensive side of the puck. Here is a breakdown- after 14-games this season- of the Panthers offensive zone starts for defense:

Name Offensive Zone Start Rate %
Aaron Ekblad 64.3
Brian Campbell 62.7
Alex Petrovic 47.1
Dmitry Kulikov 44.9
Willie Mitchell 33.1
Erik Gudbranson

35.0

With zone starts heavily weighted to defense- you should expect that Gudbranson and Mitchell are going to have the worst individual Corsi rates of the defensemen, and they do. But lets take a look at how well they are doing at job #1- getting pucks out of the defensive zone and/or keeping pucks in the offensive zone. We can do this by looking at the offensive zone finish %, which tells us the percentage of times the player is finishing a shift in the offensive zone.

Name Offensive Zone Finish %
Aaron Ekblad 57.3
Brian Campbell 56.2
Alex Petrovic 46.7
Dmitry Kulikov 45.8
Erik Gudbranson 45.1
Willie Mitchell 45.0

What does this tell us? Ekblad and Campbell are finishing slightly fewer shifts in the offensive zone than they are starting there (a drop off of roughly 7% each). This is not awful, nor entirely up to them, as the forwards will have a greater impact on that statistic than the defense will in most respects, but it is also not what you’d prefer to see. It would also be nice to see Mitchell and Gudbranson finishing more than 50% of their shifts at whistles in the offensive zone. Nonetheless, these last two D-men are seeing a 10-percent or better jump from their starts on defense, to finishes on offense.

Moving to forwards, we are going to start seeing problems with the line up. Here is the breakdown:

Name Offensive Zone Start %
Garrett Wilson 71.4
Connor Brickley 60.8
Shawn Thornton 59.2
Quinton Howden 57.5
Brandon Pirri 57.3
Logan Shaw 57.1
Rocco Grimaldi 55.9
Derek MacKenzie 54.6
Vincent Trocheck 53.6
Jussi Jokinen 48.4
Dave Bolland 44.4
Reilly Smith 44.3
Jaromir Jagr 37.8
Nick Bjugstad 37.0
Aleksander Barkov 35.8
Jonathan Huberdeau 34.9

First and foremost- wow. The majority of the team’s top offensive talent is starting shifts after whistles in the defensive zone. Based on visual observation (“the eye test”) the breakout is not Jagr or Huberdeau’s strong point, yet, they are getting the vast majority of these assignments. Before going further lets see how they are finishing.

Name Offensive Zone Finish %
Garrett Wilson 65.0
Rocco Grimaldi 57.9
Jussi Jokinen 51.9
Reilly Smith 51.8
Brandon Pirri 51.1
Vincent Trocheck 50.0
Quinton Howden 49.5
Derek MacKenzie 49.5
Jaromir Jagr 49.1
Dave Bolland 48.8
Jonathan Huberdeau 48.2
Aleksander Barkov 46.8
Connor Brickley 46.1
Nick Bjugstad 45.9
Shawn Thornton 45.9
Logan Shaw 33.3

What we can see here is that some of the team’s 2nd-3rd line forwards are finishing a majority of their shifts in the offensive zone. This would include Trocheck, Pirri, Smith, and Jokinen. Of course- the problem is that Pirri and Trocheck are starting more shifts in the offensive zone than they are finishing- not a great sign for possession. On the other hand, Jokinen and Smith are going from sub 50% O-zone starts to above 50% finishes- solid indications of good transition play. Compared with how often they are starting in the offensive zone, the team’s first line forwards are doing a solid job finishing shifts in the offensive zone, while the 4th line and parts of the 3rd are losing possession and finishing in the defensive zone far too often.

Lets take a look at one more thing before we move into some conclusions here. The other “hidden” statistic that has a rather significant impact on possession metrics is how good the competition is that a player is typically matched up against. Lets take a look at the Corsi Relative to Quality of Competition numbers to see how tough the competition is on each of our players:

Name Corsi Rel. Qual. of Comp.
Rocco Grimaldi 5.807
Jaromir Jagr 3.353
Nick Bjugstad 3.063
Quinton Howden 3.013
Reilly Smith 2.780
Jonathan Huberdeau 2.523
Jussi Jokinen 2.476
Garrett Wilson 2.458
Brandon Pirri 2.394
Vincent Trocheck 2.204
Dave Bolland 2.019
Connor Brickley 1.936
Derek MacKenzie 1.850
Willie Mitchell 1.753
Erik Gudbranson 1.638
Aleksander Barkov 1.624
Shawn Thornton 1.072
Alex Petrovic .274
Dmitry Kulikov .251
Aaron Ekblad -.332
Brian Campbell -.678
Logan Shaw -.058

I included Logan Shaw- but he only played one game.

What becomes abundantly clear is that the Panthers 4th line is being given a majority of offensive zone starts, as well as the easiest quality of competition. It’s been repeated again and again: in today’s NHL a team must have 4 roll-able lines to be competitive. I am sure I am starting to sound like an annoying broken record, but this is precisely why the Panthers cannot afford to play Shawn Thornton right now. They simply must have 4th line players capable of contributing offensively against lesser quality of competition and with positive zone starts. To a great extent, this is why Trocheck, Jokinen, and Smith are among the team leaders in points- they are playing against lesser competition than Bjugstad, Jagr, and Huberdeau are facing, and are getting more beneficial zone starts.

This is equally true in the defensive zone. More capable 4th line players will enable Gerard Gallant to employ his top line in more offensive zone situations. Patrick Kane, with the Blackhawks, has a 61.2% offensive zone start rate, and a Corsi Relative to his quality of competition of -0.092. Our top scorer from last season- Nick Bjugstad has an offensive zone start rate of 37% and is facing a Corsi relative to quality of competition of 3.063. We may be starting to see some of the root causes for our lack of offensive production.

I also must agree with the comments by Barkovthe18yearold on prior articles, Dave Bolland has been an important defensive zone contributor. Bolland is finishing more shifts in the offensive zone than he starts, and is facing solid- if not overwhelming quality of competition. The injuries to Jagr, Bolland, and Barkov have cost the team in terms of line reshuffling. In theory, a two-way line that can handle tough competition, and has good transition numbers should be taking as many defensive zone draws as possible. This looks like a line of Smith, Jokinen, and Bolland. The presumptive 1st, and parts of the 2nd line, are doing a solid job of starting shifts in the D-zone, and finishing in the O-zone, but doing so is likely taking away some offensive prowess.

What exactly are we talking about here? The Panthers are using their top offensive talent in defensive roles, and their least offensively talented players in offensive situations. This is not how a league leader is doing things. To see how a champion runs assignments, here is a look at the Chicago Blackhawks top forwards:

Name Offensive Zone Start Rate %
Artemi Panarin 66.1
Artem Anisimov 65.7
Marian Hossa 63.4
Patrick Kane 61.2
Teuvo Teravainen 59.3
Jonathan Toews 57.4

Toews is widely considered one of the best two-way centers in the NHL, and he is still seeing an offensive zone start rate 20.4% higher than Bjugstad. All the Chicago forwards noted above have Corsi Relative to Quality of Competition percentages of less than one. In short- Chicago uses offensively gifted players primarily for offense, and players like Andrew Shaw, Andrew Desjardins and Marcus Kruger are being used for defensive assignments.

This brings some of the Panthers present woes into focus. Our top offensive players are being used in two-way roles that badly reduce their offensive effectiveness. It would appear that there is a lack of respect by the coaching staff for the 4th line on defense, and with Thornton in the lineup, their brief experiment in 4th line offensive contributions (Quinton Howden- 4 points in 12 games, Derek MacKenzie- 5 points in 14 games, and Connor Brickley- 3 points in 10 games) has ended. The Panthers have reversed Chicago’s strategy by giving the 4th line more O-zone starts than D-zone starts, and their 1st line the exact opposite.

First and foremost the Panthers need to get healthy. The team is thin on top-6 talent, making the injury to Barkov far more devastating. Starting at the bottom, the Cats must have a 4th line that can play sufficient defense to warrant being rolled out for defensive zone face-offs. If that line is going to be used offensively against lesser quality opponents then they must have players who can contribute offensively in the lineup. This will likely mean a line consisting of Howden, MacKenzie, Brickley, and/or Wilson. A 3rd line more gifted in defensive zone starts that can turn into offensive zone finishes via solid transition play needs to be crafted. That likely means Jokinen, Bolland, and Smith. The team can than have the luxury of letting its most offensively gifted players play more offense.

It bears mentioning something I will call the “Serge factor.” I have named this for a friend who has decided that under Gerard Gallant, the Panthers are using a Montreal style system of play. Looking at the metrics we studied above, he may be correct. Last season, Montreal finished 21st in the league, giving up 30.1 shots per game, and 20th in the league in goals for per game with 2.61. The Canadians relied upon stellar goaltending and strong defense to get into the playoffs and advance. They played a very conservative game plan that was notably short on offensive output. Look at the Canadians player usage:

Name Offensive Zone Start %
Alex Galchenyuk 63.5
Alexander Semin 63.1
Lars Eller 62.4
David Desharnais 55.1
Dale Weise 51.6
Tomas Fleischmann 48.5
Brendan Gallagher 46.3

Notice you have not yet seen any of Montreal’s top-line forwards, but we are seeing their middle 6 as well as young and sheltered Mr. Galchenyuk. So lets go ahead and look at the rest:

Name Offensive Zone Start %
Tomas Plekanec 42.7
Max Pacioretty 42.2
Devante Smith-Pelly 35.8
Torrey Mitchell 29.9
Brian Flynn 27.6

Pacioretty and company are seeing better zone starts than the Panthers 1st line, but not by much. They are also seeing less quality of competition than the Panthers top-line forwards are, although Montreal does give its first line mid-level quality of competition. But to some degree, the Panthers are also running a similar forecheck to the Canadians, with a conservative 1-2-2 for their top lines. The Panthers also run this conservative forecheck with their 4th line, although the 1-2-2 is far more aggressive, and often turns into a 2-1-2 (whether this is intentional or not I cannot answer). The 4th line is creating more turnovers than the other lines in their limited ice time utilizing this forecheck. We know that Gallant came from the Montreal bench, so it is entirely possible that the plan is to let Luongo and Montoya handle more shots and forego offensive opportunity by playing heavy defense. The problem- of course- is that the first line-without Barkov- is not a defensively responsible lineup.

No matter whether I am correct in this supposition or not, the numbers are certainly eye-opening and go far in explaining the offensive woes surrounding the team of late. Our player usage suggests that we are using more of a Montreal system than a Chicago style system. It makes perfect sense if you plug it all in- the team is not being set up to succeed. If your giving your 4th line heavy offensive zone time against poor competition, why would you be using Shawn Thornton? And why give the 1st line such heavy defensive zone responsibility when even the casual fan can see Jagr and Huberdeau struggling in the extreme on breakouts? I suppose this works in Montreal, and maybe that’s the hope here as well, but one can’t help but scratch their head. Having said this, I am also not willing to use this to provide an excuse for the players. Put bluntly, the team underperformed in the extreme over the last five games, and the lack of performance has gotten worse in each game as we have progressed. As we watch the next two games- both of which the team should win on paper, we will be able to keep an eye on usage. Unfortunately, that may frustrate us even more than we have already been.

Talking Points