The Low-Event State of the Devils Possession: 2007-2016

Alex wrote about the New Jersey Devils’ possession rate - it’s Corsi For percentage (CF%) - on Saturday. Namely that it stunk last season and it’ll need to improve for the team to have a better chance of making the postseason. As he wrote, most playoff teams are a positive (50 or better percent CF%) and that is not a coincidence. (It’s also not a guarantee - something Alex also noted.) Expecting a big turnaround would not be right, but some improvement would indicate that things are heading in the right direction. This led to the following comment by AlienDev:

Hey Alex – any ideas why our team dropped off so precipitously? Is it correlated to a change in strategy or formations? I would prefer to think of the data to support a change in something we want the guys to do, versus just asking them to make this stat go up. For example – do we want shorter passes, or less dump & chase, or more pass completions, or less passes and more skate with the puck? Likewise, on the D side, do we want to pressure more, push up more, swamp the puck carrier? Just looking for ideas… maybe a follow up post?

While I’m not Alex, I do have an interest in this sort of thing. I do not have the answer as to why. At least, not yet. The first step is to look at the bigger picture. With War on Ice offline, Corsica is a good site as any for this sort of data. They have it from the 2007-08 season through to 2015-16 for all thirty teams, or 270 instances. From that, we can see how Devils have compared with the rest of the league in terms of the number of Corsi For (all shooting attempts for), Corsi Against (all shooting attempts against), rate stats, and CF% in 5-on-5 play. Within the Devils themselves, we can see if there’s been any changes over the seasons. Last season, the Devils were great in terms of limiting attempts against and abysmal in generating them, such that it led to a really low CF%. Was it always this way? If not, then perhaps we look a little closer at that season and compare it to this past season to get closer to giving AlienDev, Alex, myself, and others the answers as to these questions.

Data by Corsica.Hockey

Immediately, I’m confused how Corsica has 83 games played for the Devils in 2011-12. That’s odd. The 5-on-5 time on ice appears OK, perhaps it’s a glitch in their database. Anyway, we can see that last season really stood out in a bad way. While there were two seasons where the Devils’ CF% was really just above the breakeven point, it fell down for the last two seasons. However, CF% alone doesn’t get to the Devils’ issues. Looking over to the left a bit at the CF/60 (Corsi For per 60 minutes) and CA/60 (Corsi Against per 60 minutes) highlights what the Devils have and have not done well. Namely, they have not done well at generating shots and they’ve done really well at causing their opponents to generate less. 2014-15 was the first time in eight seasons where they allowed more attempts than they took. It got worse last season. Clearly, they cratered from the status quo of prior seasons.

However, that 2008-09 season is notable for having the only CF/60 rate above 55. Why does that matter? Let’s see how it ranks among the other teams per Corsica’s site:

Data from Corsica.hockey

For some strange reason, Corsica does not have 5-on-5 time on ice for Dallas’ and Nashville’s 2007-08 seasons. So the first two ranges are out of 268 instances and the CF% range is out of 270 instances. From this, we can see that most teams are somewhere between 50 and 60 CF/60 and CA/60. There have been seasons where teams were exceptionally good at CF/60 or exceptionally awful at CA/60; there have been even fewer that have been exceptionally awful at CF/60 and exceptionally good at CA/60. That makes for the CF% range to be mostly between 45-55%. A few teams have had a great or terrible differential - and that one team that finished below 40% CF% was the “glorious” 2014-15 Buffalo Sabres. This all points to where most teams are and provides perspective of what’s good and what’s bad beyond being relative to a breakeven point. So let’s see how the Devils ranked in each of those categories from their past nine seasons.

Data by Corsica.Hockey

The cells highlighted in yellow mean that the Devils were better than the league median in that stat for that season. That’s 134 or better for CF/60 and CA/60 and 135 or better for the other three stats here. For CA and CA/60, its better to have a lower rank. The Devils have certainly have been great in that regard. Despite the changes in coaches, personnel, and general team philosophies, the Devils’ CA/60 has always been on the right side of the median. In fact, six of these nine seasons are among the fifteen-lowest CA/60 values of all NHL teams since 2007. Two of the three are in that really rare range of being between 40 and 45 CA/60 (2012-13 and 2013-14). The CA/60 has bumped up over the last two seasons. Relative of the rest of the league, though, it does not look like this is the issue.

The issue, again, is on the offensive side. While there were a few seasons of where raw CF - total number of Corsi For, or shooting attempts - wasn’t so far away from the league median, they’ve only topped it once in CF and CF/60. The 2008-09 season. While the Devils have had better CF% seasons, those were driven by some amazing defensive seasons. The Devils’ generation of shooting attempts was quite poor, they just made their opponents much poorer than them in 5-on-5 play. The 2008-09 season is the one season that appears to be the best from both worlds: a respectable offense from an attempts standpoint as well as a strong defensive performance.

Therefore, I think a closer look into that season will find some answers as far as what the Devils should do tactically. It’s the one season where they weren’t a low event team on both sides of the puck. It could yield some ideas on how the team can be - let’s all say it together now - more “fast, attacking, and supportive.” It may even spur some thinking for the re-build beyond get some new guys. We’ll see.