“Fancy Stats” and Why They Are Important

With the expanding popularity of corsi and scoring chance statistics in the hockey community, there’s been a bit of a backlash by those who prefer to judge what they see by observations instead. There’s no problem with that but when they throw out accusations like “you people with your fancy stats probably don’t even watch the games” or “the numbers don’t show what happened on the ice accurately” when you present data that goes against one of their beliefs is where I start to draw some problems. What irritates me the most is there’s a lot of people who go against stats like corsi without knowing what it is or what it represents. I mentioned in my last post about Erik Cole that a poor corsi rating does not mean a player is bad, it means that he may not be as good as the 26 goals he scored say he is. To get a better understanding of this, let’s discuss corsi and scoring chances and why hockey statisticians like Gabe Desjardins, Derek Zona and Scott Reynolds put as much value into them as they do.

Corsi is a system that is similar to plus/minus system that uses shots (including missed and blocked shots) instead of goals. Corsi is a good way to determine how effective a player is at controlling the pace of play and which players are doing most of the work on their lines while others might be slacking. It basically gagues how effective a player (or a line) may be in terms of possession and keeping the puck out of out their own zone. This very useful for judging players who are on a main scoring line, like Erik Cole for instance. Basic logic should tell you that a player’s ability to get shots on goal and play well territorially benefits the team, which is why corsi is valued so highly by bloggers me like me. It is also important to look at corsi at even strength with the score tied, which you can do on time on ice’s site, because a team who is leading is more likely to play a more laid back style and give up more shots while a team trailing will likely be playing more aggressive.

Scoring chances are also valued very highly but they give you a better idea of big of a contribution a certain player is making to the team. A scoring chance is a shot on goal from a dangerous scoring area, which is defined loosely as the area in front of the net to the top of the two face-off circles inside the dots. Some scorers have different definitions on what the exact area is, though. Neil Greenberg from Russian Machine Never Breaks did a sample from 100 random NHL games last season and showed that more goals came from this area than anywhere else and the teams who were able to produce more shots from this area won more games. Like Corsi, basic logic should tell you that players who produced more chances end up scoring more goals for their team and thus give them a better chance at winning, which is why scoring chances are important. Scoring chances are also very useful for judging defenseman’s ability to prevent chances from occurring in their own end. Unfortunately, there are only a few bloggers who log scoring chances for their respective teams but I plan to add to that list this year by tracking scoring chances for the Hurricanes with the help of Vic Ferrari’s awesome Time On Ice site.

The main criticism that has used against these stats is that they fail to bring context into the equation, which is true if you go by raw data. A player who gets a ton of defensive zone starts like the Canucks Manny Malhotra or Carolina’s Brandon Sutter is likely to have a low corsi rating because there’s a good chance they give up a lot of shots due to starting so many shifts in their own end against tough competition. That’s where you have to consider things like zone starts, quality of competition, quality of teammates and what their role is on the team when using these stats. Anyone who doesn’t is either being lazy or simply a bad statistician. There have been some formulas drawn up like balanced corsi and corsi relative to zone starts to fix this problem, though but that is still a work in progress. A good example within the Canes is rookie Jeff Skinner, who had a very good rookie season scoring 31 goals, 62 points and won the Calder Trophy. His Corsi/60 rating was also positive and that’s gotten a lot of people excited about how he can improve on this next season and possibly move up to a line with Eric Staal. However, one thing we have to consider is that Skinner took 53% of his draws in the offensive zone, played against below average competition, was very poor at face-offs when he played center (36.9% success rate) and didn’t kill penalties at all. He is a terrific and creative offensive player with loads of talent and dominated the competition he was assigned (he also had one of the highest shooting percentages on the team) but I would wait until he has another season like this against tougher assignments before I declare him as the second best player on the Canes.

On the other hand, you have a player like Chad LaRose who is a player that provides secondary scoring for the Canes, takes a lot of defensive zone starts, kills penalties, faces somewhat difficult competition (not as much as Staal or Sutter did) and still managed to be a positive corsi player despite that. That’s the kind of player most coaches would like to have on their team and I was very happy the Canes decided to re-sign him to a new contract for a pretty low cost. Do I think he is better than Skinner? No. Comparing them is silly because they have different roles but I think a player like LaRose has a lot of value to the Canes despite never being someone who receives a lot of praise.

ArikJames from the Calgary Flames blog, Matchsticks and Gasoline illustrates the point I’m trying to make more clearly. When you watch a game, you are more likely to remember the guy who scored two goals despite playing poorly the rest of the game instead of the player who consistently got shots on net and looked like he was dominating territorially despite not scoring that game. This is where things like scoring chances and corsi come in handy. They help you remember these things. I am not making this post to say “my way of doing things is right and yours is wrong” but rather to shed some light on the world of advanced statistics in hockey and why they may not be as intimidating as some think they are. The numbers don’t lie but the way some people interpret them can definitely steer you in the wrong direction.

Quantcast