# Crunching numbers to gauge value of Bruins Julien’s defense-first system

There has been so much clamoring for head coach Claude Julien to either loosen the reins on his defense-first system or lose his job, it got me wondering if playing defensive-minded hockey pays off in the long run.

We know that in the Bruins’ first three seasons under Julien, they have lost in the first, second and second round of the playoffs, respectively. But it seems that watching the playoffs, teams like Pittsburgh, Detroit and Philadelphia have fared pretty well beyond the second round when their defense and goaltending has been tight and they scored just enough to get by.

Luckily, engineer and part-time sabermetrician Nicholas Amendolare came up with the idea to look back at the first five post-lockout seasons and see if defense-first hockey is a better policy than playing a system more geared toward offense.

Here are Nicholas’ findings:

The key to winning games is scoring more goals than your opponent, which we can all agree.

But, in today’s NHL there’s some disagreement on how best to accomplish winning. Is the Bruins’ defense-first philosophy justified, or would they be better served by a run-and-gun approach like that of the Colorado Avalanche? To answer this question, I took a look at some key statistics in the post-lockout NHL and compared them to teams’ performance.

For each of the five seasons from 2005 to 2010, I ranked a team’s performance. The Stanley Cup winner was given a rank of 1, the Cup finals loser was ranked 2, and the two teams who lost in the conference finals were ranked in a tie for third, and so on. Teams who didn’t make the playoffs were ranked based on regular-season point totals. For example, last year’s season had Chicago ranked 1, Philadelphia ranked 2, and Toronto ranked 29.

I then took four key statistics to compare to these rankings: goals for, goals allowed, 5-on-5 ratio (goals for/goals against during 5-on-5 play only), and goal differential. The idea was to calculate the correlation of each of these statistics to a team’s ranking to determine which of the four statistics was a better predictor of a team’s success. Correlation is simply a measure of how related two quantities are, and it’s measured on a scale of -1.0 to 1.0.

For example, I expected that goal differential would have a very high correlation with team performance (the key to winning is scoring more goals than you give up, right?). On the other hand, one would probably expect something arbitrary, like a team’s jersey colors, to have a very low correlation with team performance (close to 0.0).

The correlation of each of our four statistics to team performance was as follows:

Goals For: -0.55

Goals Allowed: 0.71

5-on-5 Ratio: -0.70

Goal Differential: -0.84

So what does all this mean? For someone without a Math or Science background, here are the bare essentials:

**1. You can ignore the negative signs above. **The negative sign simply indicates the direction of the correlation, not the strength. More goals for resulted in a lower ranking (a negative correlation) while more goals allowed resulted in a higher ranking (positive correlation), just as common sense would have us believe.

**2. Goal differential was the best predictor of success. **This makes sense. Good teams score a lot more goals than they give up. Five-on-5 ratio was also a good predictor of success, but this was nothing groundbreaking given that it’s closely related to goal differential and given that most hockey is played 5-on-5.

**3. Minimizing goals allowed showed a stronger correlation than maximizing goals for. **The difference was significant, but not overwhelming. In the post-lockout NHL, defensive teams have fared better than offensive teams.

The third finding is the most important one. If your goal is to win hockey games and succeed in the playoffs, a defense-first approach seems to be justified by the evidence (sorry, Colorado fans). If your goal is fan entertainment, then the answer might be different. Does this mean Claude Julien’s job is safe? Of course not. But it does shed some light on the potential reasoning behind his approach. It could also lend credence to the adage, “Defense wins championships.”

Page 1 of 2 | Next page