“It is an eyeballs business. But there’s what you see, and being able to incorporate some checks and balances into it with data, and what you can track and pull out and reaffirm. If you can be even 1 or 2 per cent more certain in what you’re seeing and decisions you’re making, that can be a massive benefit in the long run. Things are always evolving. You have to kind of be on the edge.” – Kyle Dubas
Like Kyle Dubas, I’m a firm believer that hockey players should primarily be evaluated simply by watching them play. That said, I’m also aware that I see and remember only a small piece of the information that is presented to me over the course of watching a game (let alone a season), and that bias is an unavoidable reality. As a result, also like Dubas, I believe that the use of statistics is useful; to weigh the underlying numbers against what I see to help provide more informed judgments.
I’m not an expert in this arena, by any stretch of the imagination. Truthfully, I don’t ever really care to be. This will be the first time I’ve written on it, in fact (so go easy on me!)
There is a lot of debate within Leafs Nation right now around which players are adding positive and negative value. Start talking Nazem Kadri, Jake Gardiner, Roman Polak or Tyler Bozak, for example, and you will quickly see how divided we are about these guys. I isolated those 4 because they have something in common, the eye test and the numbers test don’t seem to reconcile. For many, Kadri and Gardiner fail the eye test, but the numbers seem to tell a different story. For Bozak and Polak, the numbers test in some respects doesn’t look as flattering, but the eye test seems tells us that they are positive contributors. What do we do with that?
With that in mind, I will focus on 3 questions, with the hope that we can get a clearer picture of just what these players (and the others) are contributing, accounting as much as possible for the context that surrounds their play:
1. Who is playing against the easiest/toughest competition?
2. Who is getting the easiest/toughest assignments?
And, with those two factors considered,
3. Who is contributing the most to shot attempts for and against?
The best way to measure that, as far as I can tell, is to isolate 5v5 “close” score situations, in order to compensate for score effects (teams that are down by 2 or more tend to generate more shots, and teams that are up by 2 or more tend to go into a defensive shell), and look at 3 stats: Corsi Rel%, Corsi Competition%, and Off ZS%.
So, before we go any further, let’s make sure we’re all on the same page with the terminology.
Refers to situations when the game is within 1 goal (1st and 2nd periods) or tied (3rd period or overtime)
The percentage of on-ice shot attempts (on goal, missed, or blocked) taken by the player’s team; also known as CF%
The player’s on-ice Corsi% minus the player’s off-ice Corsi%; off-ice Corsi% is the percent of shot attempts taken by the player’s team when the player is not on the ice (but in games where the player is in the lineup); also known as CF Rel%
The TOI-weighted Corsi% for a player’s competition
The percent of all non-neutral zone on-ice faceoffs taken in the offensive zone
Using those data sets, we’ll separate forwards from defensemen, take a look at what the numbers say, and see if the data confirms what our eyes have seen watching the first 21 games of the season. Well…let’s dive in.
The chart below basically just puts the info in the table above in graphical form. Players higher on the chart face tougher competition (Corsi Competition%), with left to right measuring the percentage of shots the team generates with them on the ice vs with them off the ice (Corsi Rel%). Red indicates a higher percentage of defensive zone starts, and blue indicates a higher percentage of offensive zone starts. The size of the circle correlates with the player’s average ice time.
What do the numbers tell us?
Six forwards have a positive Corsi Rel%. They are Lupul, Smith, Leivo, Kadri, Winnik and Panik. Of those six, Winnik, Kadri and Lupul have faced the toughest competition, while Winnik, Kadri and Panik all start in the defensive zone more often than the offensive zone. Smith, Leivo and Lupul all have much more favorable zone starts, with Leivo and Smith standing out in particular as being given very sheltered minutes and competition. Lupul has an unreal Corsi Rel% of 17.6, but the sample size is small and his Off ZS% is pretty high. This still supports what we saw from him when healthy; he was pushing the play and getting pucks to the net with authority.
Kadri’s numbers actually look pretty solid. With a Corsi Rel% of 7.55 (4th best) despite respectable competition and a high percentage of defensive zone starts, the numbers seem to show that Kadri, while not putting up the usual point totals, is in fact creating more positive events than negative ones on the ice. Will the points come for Naz if he continues to drive the play? Time will tell.
There’s a cluster of 7 forwards that all face similar competition, including, Bozak, Komarov, Winnik, Van Riemsdyk, Kessel, Santorelli and Clarkson. Of those 7 Winnik stands out with a Corsi Rel% of 5.72. Santorelli and Komarov stand out as well, starting in the offensive zone only 37.14% and 38.94% of the time, respectively. Carlyle trusts these two. Of that group, Bozak faces the toughest competition, but has a Corsi Rel% of -4.26. This too seems to confirm what our eyes see; that Carlyle trusts Bozak with tough minutes, but that his line often is on the wrong side of the shot differential.
What do the numbers tell us?
Phaneuf faces the toughest competition of the group, followed by Franson, Gardiner and Polak. Of those 4, Gardiner stands out with an extremely high Off ZS% of 53.47, and Polak with a Off ZS% of only 38.41 (plus the second highest TOI average among D, next to Phaneuf). The only player who starts in the dzone more than Polak is Stuart Percy, who has an Off ZS% of only 33.33%, though with Percy we’re only looking at 8 games played and his competition is a lot lighter. In other words, nobody has towed more weight on the blueline than Roman Polak. His Corsi Rel% of -2.1 is more than understandable, given the context. Gardiner is obviously being sheltered by Carlyle, who has been pressing Jake for improved defensive zone play.
Franson (4.89%), Rielly (2.91%), Phaneuf (2.2%), and Gardiner (.99%) all have a positive Corsi Rel%, which is most impressive for Phanuef and Franson, given the competition they face and the amount they start in the defensive zone.
As eluded to before, Percy has faced the easiest competition of the 7 defensemen, but his high defensive zone start percentage counterbalances that. Carlyle appears to trust Percy’s ability to transition the puck out of the zone quickly, but didn’t want to over-expose the young blueliner.
Robidas stands out in a negative way with a Corsi Rel% of -7.24, despite easier competition than Phanuef, Franson, Gardiner and Polak and only semi-difficult zone starts. These aren’t flattering numbers for a 37 year old locked up for 2 more years. Coming off an injury could explain a lot of that though, as his play does appear to be improving as the season moves along.
What do you think?
Do these numbers confirm or conflict with what you see while watching the games? Are the numbers reconcilable to what you see from guys like Kadri and Gardiner? Obviously this only shows a small piece of the puzzle. For example, they do not give us any sense of how players are producing, or if their play is leading to quality scoring chances against. This just gives us a basic idea of how the team is driving the play when these guys are on the ice. Let us know what you think in the comment section.
Thanks for reading.
Data source: war-on-ice.com