In part 1, we looked at some of the intrigue at the NHL level surrounding the idea of advanced stats including Corsi and discussed the nature of the Corsi statistic and how it is calculated.

Even back when Corsi was developed, we already had lots of statistics in Hockey.  Goals, Assists, Shots, Hits, etc. are tracked and have been tracked for most of a century.  What is the point of adding more stats?  We don’t need to look past goals and assists to know that Wayne Gretzky was a great player.  But how great was he?  What was his contribution in all three zones?

Well, standard plus-minus can tell us a lot about a forward’s two-way play. We have had that stat for years.  Isn’t that enough?

There are two problems with plus-minus.

1.  You can give up a goal when you are individually doing everything right and your team can score when you are individually doing very little right.  This means you can be rewarded for poor play or punished for good play.  These circumstances are rare, but they make the data confusing especially if we don’t have a lot of data points to work with.

2.  There are only a handful of goals per game and, while they are most often built on a sequence of good plays by one team, many sequences of good plays by a team don’t result in goals.

Essentially #2 above tells us that using goals leads to problems of a “small sample size”.  Further, #1 above tells us we would need a large sample to even out the “bad luck” or “good luck” that can influence goal scoring.

How can we incorporate a larger sample?  

We can do this by using shot attempts.

To a lesser degree than with goals, shot attempts also typically result from a series of good plays by a team.  They also happen WAY more often than goals.  So, we can start to get at the same question (how effective is this player in a two-way sense at contributing to his team’s success?) but with far more events to factor into the stat and “smooth out” the luck.  And so... we arrive at the concept that is Corsi.  Keep in mind that Corsi for an individual player only factors in what happened when that player was on the ice.

What can all of this tell us about Youth Hockey?

At all levels of the game, a goal scorer is celebrated.  The same, to a lesser degree, can be said about assist makers.  What about the guy who made the pass that was three passes before the goal (key play but no assist).  What about the guy who consistently keeps defensive inside position on the player he is supposed to cover (the other team won’t score as often, but it is hard to visually “see” the good plays that went in to this)?  Or how about the D that uses elusive skating and an accurate breakout pass to neutralize the forecheck pressure the other team had created (turns defensive into offense, but again, without an obvious stat to show us that).

All of those things are somewhat obvious to expert hockey coaches, but not to all hockey parents.  All of these things have the potential to contribute to a win just as much as or more so than scoring a goal.

How many times have we heard “my kid has been stuck at D all year”?  Your kid’s contribution at D may be more impactful on team winning than the flashy goal scorer who tends to get the glory.  When I coach at a youth level the players who play at defense are the ones I can trust to keep the play in front of them through a combination of mental quickness and savvy plus agile forward, backward, and transitional skating.  This “trust level” that I have for those players is a badge of honor in my view.

Corsi doesn’t care who scored the goal.  Corsi doesn’t care who got the assist.  Corsi does care about a player’s total contribution in all zones and how much that impacted his team’s ability to generate shot attempts vs. the other team’s.  Make sound hockey plays all over the ice that make things harder for the other team while making things easier for your teammates and Corsi will reward you.  Play selfishly looking to maximize your personal goals and assists stats and Corsi will punish you.

NHL Coaches have built up very good “eye tests” with which they try to answer this same question of two way play in all areas of the ice.  However, when we watch hockey (and this is true even for experts) we mostly remember impact plays around the puck.  Corsi doesn’t care about proximity to the puck.  It is built to understand all of the little plays and how positive or negative they are for your team.  It does this indirectly by measuring the results from sequences of those plays… those results being shot attempts.

And NHL teams are using stats like these because the total contribution to team success is what matters.  In truth, a player’s goal total does tell you part of this picture.  However, as you move up levels of the game, a higher and higher percentage of goal scoring happens as a result of sound play by all 5 skaters.  At high levels, this sequence of good plays leading to a goal most often starts when the other team has the puck, continues through the process of transitioning to offense and finally concludes around the net.  The goal stat tells you only happened to be lucky beneficiary of this play (of course you have to keep in mind that a player who can be this lucky beneficiary often is still very valuable!).

So, the influence of stats like Corsi at the highest level of the game should tell youth hockey parents to look a little deeper and consider the contribution of players in all phases of the game and parts of the ice.  The fact that it uses shot attempts as a replacement for goals to measure players demonstrates the need for more common events to smooth out the luck-factor that goes into goal scoring.  This should also tell parents that a really good forward who will score a lot over the course of a season is going to go several games without scoring and this doesn’t mean they are playing poorly… it is just the puck luck not falling in their favor.

There are many reasons to go beyond the narrow focus of goal scoring when considering player value and to wait longer than a few games before drawing conclusions about production on the more random stats like goals and assists.  Instead one should think about how hard a player makes it for the other team to win.  The player who makes it hardest on the other team is the player that contributes to their team winning the most, regardless what the more well known stats may say.