Drawing Calls & Puck Possession: Two Problems for Devils, But No Strong Signs of Correlation

After Saturday's 1-3 loss at Carolina, I realized the New Jersey Devils have not been on the power play all that often in recent games.  Prior to the game against Anaheim, the Devils have had four straight games of only earning one power play opportunity. Further, prior to Sunday's games, the Devils are 29th in the NHL in total power play opportunities. While the Adam Oates-driven combination of dump-ins and the 1-3-1 formation has varied between feast or famine in terms of offense, it is better to have the time than not.  It's time where the other team is forced to not attack, keep certain players on the ice, and hopefully give the Devils' forwards an opportunity to use the extra space to create a good opportunity to score, much less score.  Not having that time can hurt the overall cause.  Especially since the Devils aren't that good at even strength this season.

Per War on Ice, we know the New Jersey Devils are near the bottom five in Corsi For%.  They're at 46.9%, which is definitely their worst mark since Corsi could be calculated in 2005-06 (NHL didn't count blocks and misses before then).  While the team has only been out-scored in 5-on-5 play by four goals, that just speaks to how amazing Cory Schneider and Keith Kinkaid have been in the larger picture of this season.  That larger picture features the Devils not having the puck, not taking it forward, and not doing much more than playing to the wills of their opposition.  This must change for the Devils to become more competitive in future seasons.

I began to wonder, are the two related? Typically, a player draws a penalty because they or their team has the puck. Most fouls have to do with restraining players or more violent actions, and that's usually by the team defending the play.  Maybe they got beat, maybe they're frustrated with getting beat, maybe they cross a line in the act of something else.  But those penalties taken mean penalties are drawn.  Is there a correlation between the two?  Since I'm a bad blogger and unlike luminous reporter Rich Chere, my headline gives you the answer: no.  There are no strong signs that the two are correlated.  This post will provide the evidence that led to my conclusion and what that means.

I started by going to War on Ice and pulling the Corsi For% for all thirty NHL teams prior to Sunday's games and how many penalties they've drawn.  I made a scatter plot and added a trend line to get a coefficient of determination (R^2), which points out how well a trend line fits the data.  Stats that correlate tend to have high coefficients of determination.

Within this season so far, there's very little correlation between the two stats.  A coefficient of 0.147 with a power trend line strongly suggests no relationship.  While the Devils are near the bottom in both penalties drawn and CF%, that relationship doesn't hold across the league. For example, Our Hated Rivals have drawn fewer penalties than the Devils but they have a superior-to-the-Devils CF% of 50.1.

However, this is just one season.  If the idea is to see if there's any relationship, then we'd need to look at more data. So I expanded the scope of seasons to start from 2007-08 through to this season so far.  For some reason, the penalties taken and drawn numbers at War on Ice from 2005-06 and 2006-07 are really small.  I highly doubt the Devils only took sixteen penalties and drew nineteen in 2005-06, for example.  I'm not sure why the counts shot up so much in 2007-08; I think the NHL may have re-defined how that was recorded.  In either case, I started from there.   And I got even more noise.

A coefficient of determination ranges between 0 and 1.  If the best-fit trend line has a coefficient that requires four decimal places, then it's safe to say there's no relationship between CF% and penalties drawn.  Even if I were to remove the lockout-shortened 2013 season, the logarithmic trend line (the best fit I could do with all seasons) actually fits worse with a lower coefficient of determination.   Basically, there's no correlation between a good possession team and drawing penalties.

It was at this point I realized that penalties drawn is not necessarily the same as drawing power plays.  Teams do get power plays for victimless penalties.  These include, but are not limited to, delay of game calls, hand on puck, unsportsmanlike conduct with referees, and bench minors.  I don't think they are so frequent, but there could be enough of them over the course of a season to make some difference between the number power play opportunities and the number of penalties drawn.

Therefore, I went to NHL.com and pulled the number of power play opportunities for each team from 2007-08 through this season so far.  After charting those values with the CF% of each team in each season, I'm further convinced of a lack of a relationship between the two.

The coefficient of determination is better but it's still really, really low.  Far too low to conclude of any correlation between the two stats.  It appears that drawing power play opportunities involves other factors than how much a team is out-attempting (or getting out-attempted) their opposition.

I recognize that I'm limited to only four trending lines: linear, exponential, power, and logarithmic functions.  It is possible there's a better model that would be a better fit for this data that would suggest a relationship between the either two stats (CF% vs. power play opportunities, CF% vs. penalties drawn).  However, given the sheer amount of data clumped together over the two charts over seven seasons, I can't really see what model would really fit.

I also recognize that I'm comparing raw data from various seasons.  A correction factor may be appropriate both for penalties drawn and power play opportunities.  It is very true that teams are getting called for fewer penalties than they were back in 2007-08 and 2008-09.  Teams were still adjusting to changes in how referees called certain fouls - namely, hooking and slashing; touching the glove with the stick equaled a call - after the 2005 lockout.  Therefore, many teams easily exceeded 300 power play opportunities.   While there are still two weeks in this season, no team will break 300 power play opportunities unless a team decides to constantly hack Detroit and Dallas, currently tied for the league lead at 269.  That said, even with a correction factor for season to season, I doubt that's going to lead to a strong relationship between either CF% and penalties drawn or CF% and power play opportunities because there's a weak relationship within seasons.  (Aside: I did do CF% vs. power play opportunities for 2014-15; my coefficient of determination is approximately 0.1095, which is still really low.)

There also appear to be contributing factors that would lead to the lack of a relationship between these stats.  One that comes to my mind are the players themselves.  There are definitely players who draw more attention to others.  Players like Alex Ovechkin, Vladmir Tarasenko, Nazem Kadri, and Nathan MacKinnon are among league leaders in penalties drawn this season per War on Ice. They're scoring talents, they're quite quick, they can beat defenders, and so they'll draw fouls if only to deny them opportunities to score.  There are also players who are seemingly more adept than drawing calls than others.  Players like Dustin Brown, Tom Wilson, Rob Klinkhammer, and Brock Nelson aren't big scorers or even players who necessarily play in scoring roles are also up there in penalties drawn.  It is possible to add pests who just get under an opposing player's skin to get an extra call just as much as it's possible to get someone so dangerously talented that opposing player's are forced to foul.  Not every team has such players, and so that can skew matters. The Devils don't appear to have either, or if they do, not enough of them.

For the Devils, not drawing a lot of power plays or penalty calls is one problem.  While the team's power play has serious issues with their breakouts or build-up to gaining the zone and it's questionable whether the 1-3-1 formation is the right one for the players that they have, it is an issue of getting an opportunity to be out there. Again, even an unsuccessful power play takes two minutes off the clock, forces the opposition to defend, and it can provide a means to generating some offense even when the team has struggled to do so at even strength.  A successful one can obviously make the game easier or more attainable, which is always a plus for a low-scoring team like the Devils.  I think figuring out how the team can generate more opportunities for power plays, which would come from drawn calls as the Devils can't force bench minors or delay of game calls, is worth pursuing.

For the Devils, getting out-attempted on a regular basis is another problem and arguably a bigger one. Even when the referee is whistle-happy, the majority of a hockey game will be played at even strength.  A 46.9% CF is reflective of what has been seen throughout the season from the Devils.  Bad passes, dump-ins, and turnovers on offense; bad passes, dump-ins and turnovers in the neutral zone; bad passes, turnovers, and chipping pucks away all lead to the same thing: the other team getting the puck and doing something positive with it.  While all teams suffer from inaccurate passes and turnovers, and every team does run into situations where dumping the puck away is all they can do, the Devils have done it frequently enough to be an issue in most games.  And the result is an opposing team that makes the Devils look and play foolish hockey, rely on their goaltenders to be great just to keep the score from getting out of hand, and end up losing more than winning.   There are deeper issues with how efficient the Devils are at moving the puck, where I think improvement in that regard will improve possession (CF%).  It's still important that the Devils figure out what adjustments they can make in terms of talent, tactics, and execution to have the puck more often.  Not only will CF% rise, but their performances on the ice will be more competitive and that can lead to more winnable games.  Not necessarily wins, but games where it's possible the Devils can get one.  That's why it's critical the Devils try to improve in this aspect as they re-build.

Ultimately, what we have here are two problems that aren't necessarily going to be solved by only one of them being solved.  They might appear to improve at the same time, and ideally they will over the next few seasons.  But as league data has shown from this season and the previous seven, there does not appear to be any correlation between the two stats.  It would be more of a coincidence if that were the case.

Perhaps there's something else I'm missing. Maybe I misinterpreted the results of the charts.  Maybe there's some other approach to determine a correlation between CF% and drawing penalties or obtaining power play opportunities.  Maybe establishing the relationship isn't as important as is recognizing the reality that the Devils will finish poorly relative to the rest of the league in both stats in 2014-15 and so they need to make improvements.  Please leave your thoughts, suggestions, corrections, and other thoughts about these two issues and the math and logic attempted within this post in the comments.  Thank you for reading.