With Or Without You Analysis: Zach Parise's Corsi Impact on the 2009-10 New Jersey Devils

I've decided to take a small respite from looking at where the 2009-10 New Jersey Devils may have struggled, to engage in a bit of a statistical analysis.  Today, I was reading Tyler Dellow's site, mc79hockey, and came upon two analytical posts about Shawn Horcoff and Alex Ovechkin.  Utilizing the new statistical scripts from Vic Ferrari's Time on Ice site (explained here), Dellow determined the Corsi percentage for those players' teammates both when he was on the ice with them and when he wasn't.  

In the case of Ovechkin, Dellow expanded it further by determining Ovechkin's own Corsi numbers impacted when he isn't playing with a particular teammate.   Dellow referred to both analysis as WOWY - with or without you; ultimately showing by Corsi impact who is most hurt when not playing with a certain player as opposed to playing with him.

I was interested in the methodology to Dellow's studies and he graciously helped me figure it out so I understood how he came to those numbers.  After reading both posts thoroughly and trying out the new scripts at Time on Ice, I immediately wondered Zach Parise's Corsi impact on his teammates and likewise.  I'm not the only one, user BenHansa at Lighthouse Hockey has done two such WOWY analysis to show how great Frans Nielsen has been for the Islanders back in 2009-10 - one on Nielsen himself and one to determine who was the best linemate for Neilsen and Trent Hunter.

So I went ahead and compiled the numbers to determine Parise's Corsi impact on the 2009-10 New Jersey Devils using the numbers from Time on Ice and Dellow's WOWY methodology.  If nothing else, I believe the results further justify the notion that Parise was one of the best, if not the best, skaters on the New Jersey Devils roster.  Not to mention further justifying my selection for team MVP.  Keep this in mind going forward if/when contract extension talks heat up during the 2010-11 season (hopefully). 

Note: Before continuing after the jump, please select Wide on the top right box next to the headline. This way you can see the entire charts within this post.

A Quick Overview of Corsi and What It Entails

Corsi is named after Jim Corsi, a goaltending coach for the Buffalo Sabres who summed up the shots on net, blocked shots, and missed shots in a game to get a better sense of how much work the goaltenders were getting.  Corsi has been recently interviewed by David Staples of the Edmonton Journal and the whole article is incredibly insightful into the man, the stat, and how it's grown since it's formation (Behind the Net's Gabe Desjardins is featured). It also comes with the excellent definition of what Corsi is - something I fully didn't grasp as much as I thought until I read this from the Staples article:

It is desirable to have a high Corsi value as it indicates that you are present when a shot is attempted - regardless of whether it goes on net, gets blocked, or misses the net - and it decreases when present for an attempted shot against. The general idea is that a shot being attempted at all is indicative of your team having the puck.    Does it necessarily mean the player is fully responsible? No, but over thousands of instances, a high Corsi percentage (Corsi for/Total Corsi) would indicate that a player has done well - since it's highly, highly unlikely that they just happened to be present for thousands of situations for attempted shots without having anything to do with the resulting play.

The Following Measures Even Strength Impact

The numbers pulled from Vic Ferrari's Time on Ice are all from even strength, 5-on-5 situations in every game  in the 2009-10 regular season with a goaltenders in each net.  This isn't to say that special teams or other situations aren't important; but the majority of most hockey games is going to be played in at 5-on-5 with no empty nets.  Therefore, the numbers compiled consistent with the majority of ice time players get to take part in.   If a player really is important to their team, then it should show up in 5-on-5 hockey.

Zach Parise's WOWY with the New Jersey Devils 2009-10 Team

The following chart details the Corsi each teammate had when they were on the ice with and without Zach Parise.  It also shows Parise's Corsi when he isn't playing with the listed teammate. 

I should point out that the following excludes goaltenders since the goaltender is ever present on the ice and breaking down Corsi for them won't mean much.  I also excluded called-up players, since they played so few times with Parise it doesn't bear mentioning with two exceptions - Vladimir Zharkov and Matt Corrente, as they ended the season on New Jersey's roster. Neither will play much of a role in this analysis, however.

Here's the full team chart.  It's too big to really put inside as a picture but please check it out.

The first WOWY column indicates how much that teammate is hurt when they aren't playing with Zach Parise.  For a majority of the team, not playing with Parise hurts their own Corsi.  Only four Devils have experienced a higher Corsi not playing with Parise than playing with him: Paul Martin, David Clarkson, Rod Pelley, and Brian Rolston.  Incidentally, the only players where Parise's Corsi is hurt more by not playing with him than the other way around were Paul Martin and Vladimir Zharkov.

Before you jump to conclusions, I'd ignore most of the players I just mentioned except for Martin.  Clarkson, Pelley, Rolston, and Zharkov (among others) are grayed out in the team chart as they didn't really play with Zach Parise enough times to even have 100 Corsi events.  They were on different lines at even strength throughout the season and perhaps only played with Parise in passing, like the end of a power play or in the middle of a line change.  Like in Dellow's posts on Horcoff and Ovechkin, I really am only concerned with Parise's teammates with whom he's had enough Corsi events with to measure a proper impact. 

Forwards and Parise - 100+ Corsi Events Together

Let's take a closer look at them, first with the 6 7 Devil forwards.  Again, please click on the image for a bigger view.

The big conclusion here is that Travis Zajac, Jamie Langenbrunner, Dainius Zubrus, Rob Niedermayer, Patrik Elias, Niclas Bergfors and Dean McAmmond all suffered in terms of Corsi percentage when they aren't with Zach Parise. However, Parise's own Corsi percentage is only hurt when he's without Zajac, Langenbrunner, Zubrus, and Elias.

Zach Parise's own Corsi actually improves (drastically in Niedermayer's case) when he was not with Bergfors, McAmmond or Niedermayer on the ice. That may explain why their total number of events together has been so few for the latter two.   This may make intuitive sense since both are third and fourth line players, Parise's usually only with them in a 5-on-5 situation during a change or if he's being double shifted, and both McAmmond and Niedermayer aren't nearly as offensively talented as Parise.   As for Bergfors, it's hard to say - he was a developing rookie and played a little bit with Parise.  Parise's Corsi improved apart from Bergfors, whereas Bergfors' Corsi fell - but not quite so much.

Travis Zajac represents the biggest WOWY percentage both ways.    It would be quite beneficial for Parise to play with Zajac, and moreso for Zajac.  This is essentially proof of how these two have been effective in terms of positive territorial advantage.  If I'm the new head coach of the New Jersey Devils, then it would be to my benefit to keep these two players together. 

Not that Parise with Patrik Elias as a center (or a wing) has been bad; but Parise's Corsi isn't hurt nearly as much without Elias and it is for Elias to be without Parise.    Even so, Elias sported the highest Corsi among this group of Devil forwards at the end of the season without Parise.

What's really intriguing is how close the Without Parise impact is for both Langenbrunner and Zubrus. While Langenbrunner has played much more with Parise than Zubrus; their Corsi percentage was nearly the same when they are with Parise and without Parise.  From this view, one can argue for either player to be on Parise's opposite wing.  However, Parise's own Corsi reduced more without Langenbrunner, much more than playing away from Zubrus.  

It seems to me that while a Parise-Zajac-Zubrus line can certainly be effective in terms of Corsi; the best group for Parise remains ZZ Pops.   Hopefully, the new head coach will see it this way and hopefully all three players can reap the rewards.

Defensemen and Parise - 100+ Corsi Events Together

Unsurprisingly to me, more defensemen have been present for Corsi events with Parise.  There are fewer defensemen in a game and they sometimes rotate separately from forwards, leading to some interesting combinations.   Therefore, 8 defensemen have been on the ice for more than 100 Corsi events with Parise.  Let's take a closer look at them.

Surprisingly to me (and maybe you), Mike Mottau has been present for more of Zach Parise's actions than any other defenseman.  After a little digging, I shouldn't be so surprised; he did lead the Devils in even strength ice time in 2009-10 (and among all UFAs, actually).  Right behind Andy Greene, so the two of them being so present for when Parise was on the ice makes sense as a result. 

What does come as a surprise is that while the Devils' top 4 (Mottau, Greene, Colin White, Bryce Salvador) for much of 2009-10 saw significant drops in Corsi without Parise, Parise's own Corsi wasn't hurt all that much without . In the case of Mottau and Salvador, it actually improved and it's nearly a wash with White.   I suppose this speaks further to the defense's own offensive struggles - they may need to be "helped" by offensive forwards to make their own Corsi numbers look good.

The one defenseman who really stuck out here was Paul Martin.  At only 197 events, it's not a large population size, but Martin is the one Devil who had his Corsi improve without Zach Parise on the ice.   Likewise, Parise's own Corsi improves when Martin isn't available.    I'm at a loss as to how that could be; I know Martin isn't an ace when it comes to offensive abilities, but he's not lacking that department either.  His own Corsi wouldn't be so high if he was such a drain on offense.   Yet, it seems that the two players just don't perform as well as they do apart.  If I were Martin's agent, I'd use this as a selling point. (e.g. While most of NJ's defense had their offense seemingly rely on Parise's work, my client actually benefited on offense without having the Devils top forward on the ice, etc.)

Johnny Oduya represents an interesting result as well. Oduya's Corsi tanked with Parise but Parise also suffered greatly without him.  Together, the shot attempts flowed.  Apart, they weren't working as well.  Of course, the impact on Oduya was far larger than it was on Parise.  Did this result in many points? Well, 434 total events isn't a whole lot and Oduya's 2 goals and 2 assists in 40 Devils games suggest that it really didn't - and largely argue why he isn't exactly missed since the trade.  Maybe with more time together, there could have been more production, but that's all I can guess at from these numbers.

Main Conclusions

For all the talk of constant line changes, it's clear that Parise has been largely played with Zajac, Langenbrunner,  Elias, and Zubrus in.  Zajac and Langenbrunner remain as the most common forwards to play with Parise; and therefore, Parise suffers the most when away from them than any other forward as they suffer without Parise.  Basically, the new coach should consider keeping Parise with Zajac and Langenbrunner to start unless he sees something different with either player.

In terms of defense, Parise has provided a positive impact to all defensemen for New Jersey not named Paul Martin.  This isn't to say that Martin needs to go; just that constantly sticking Martin with Parise in most 5-on-5 even strength situations may not be as effective one may think.  Of course, those numbers for Martin aren't based on a ton of Corsi events - largely because Martin was injured for most of the season, leading to increased events for other defensemen.

Parise is an excellent player and he's skilled enough on offense to have a positive impact on most of his teammate's Corsi values.  Essentially, Parise is a net benefit to the team's shooting attempts, which suggest a benefit to territorial possession.  Since 21 of 22 14 out of 15 teammates that had 100 or more Corsi events with Parise suffered in Corsi when not playing with Parise in 2009-10, he truly carries an important value to the team.

Let me know what you think of all of this in the comments. If you think I made a mistake or need to clear something up, please feel free to ask.  This took me a lot longer than I expected, so don't expect to see me hammering these out day after day.   Big thanks go to Tyler Dellow of mc79hockey for his help and inspiration to do this work myself, Vic Ferrari of Time on Ice for coming up with the scripts to get the base numbers, and you, the reader, for taking all of this in.