(Photo by Harry How/Getty Images)

Earlier this year I looked at how passing stats correlated to teams winning over a 20 game sample size. I’ve now completed tracking all of the Devils and their opponents’ games from last season and can more comfortably illustrate which stats matter more than others.

As it was at the 20-game mark, more efficient teams at generating shots off of their passes won 67 of the 82 games, or 81.7 percent. That was significantly higher than any other stat tracked or used in this comparison. What plays out over these results is evidence that how efficient teams are matters more than the volume of offense they can generate. However, there are advantages in where teams generate their offense from.

Thanks to zone entry work spearheaded by Eric Tulsky, Geoffrey Fetweiler, Robert Spencer, and Corey Sznajder, we know entering the zone with possession improves the number of shots and goals a team will tally. When we look at the teams that are more accurate in the neutral zone, those teams win at a 57.3 percent clip, 6.1 percent better than being more accurate in the defensive zone and 11 percent better than being more accurate in the offensive zone. Certainly this a result of controlled entries via passes into the offensive zone.

Also, as an extension of this importance in the neutral zone, teams that generate more shot attempts with passes from the defensive or neutral zones than their opponents win 56.1 percent of their games. Now, this is not a particularly high number, but look at the disparity between it and the 37.8 percent winning percentage of teams that generate more shots from within the offensive zone. Like in Soccer, teams that transition effectively from defense to offense tend to be more effective at generating quality attempts at goal.

Let’s compare this data with more widely known stats such as Corsi and Fenwick. Between the Devils and their opponents, the team that won the Corsi battle won only 47.6 percent of their games, and the team that won the Fenwick battle won 54.9 percent of their games. There are two reasons why these numbers are lower:

  1. The Devils lost all thirteen shootouts they played, which impacts some of the numbers; and
  2. The Devils generally won the possession battle in most of their games last season.

So, let’s look at only those games ended prior to a shootout.

In these 69 games, we see an increase in most, if not all categories. Shot Attempt Generation Efficiency (SAGE) increases to correlating to a team winning 58 of 69 games, or 84.1 percent. It becomes even more accurate when removing shootouts. The importance of completion percentage is not as drastic, but neutral zone percentage remains more important than completing passes in the other two zones. Corsi and Fenwick both get slight bumps, as does SF/CF% (Shots For/Corsi For%). Changes all around, but pretty even throughout.

Now, Corsi and Fenwick both have illustrated that they do a good job of measuring the top teams in the NHL (look no further than the below chart on the previous three Stanley Cup Champions), but the Devils represent somewhat of an oddity and are an example of needing more data and isolating phases of the game to reveal “how” and “why” things happen, as opposed to just “what” happened as Corsi and Fenwick do.


This chart shows the last three Stanley Cup winners and they were one, three, and four respectively in these puck possession stats at Even Strength. Corsi and Fenwick are generally solid predictors of future success. But possession isn’t the only thing.

Some of the more notable stats at the other end of the spectrum were overall possession (measure by total pass attempts), offensive zone possession (pass attempts originating in the offensive zone, and controlling the share of offensive zone SAG (shot attempts generated). The team that won each of these three stats lost the game far more often than winning it. So, what does this tell us about what areas of the ice are more important than others? It tells us that while Corsi and Fenwick are solid stats now, as you peel back more layers of the game, you’ll find that measuring efficiency matters more than volume.

Corsi, Fenwick, and SF/CF% data was pulled from ExtraSkater.com.

Ryan Stimson is a contributor at InLouWeTrust on SB Nation and has been a lifelong New Jersey Devils fan. He believes that the future of hockey analytics is in analyzing phases of the game to reveal tendencies of winning teams. Follow him on Twitter @RK_Stimp