One area of hockey that is growing is the development of more in-depth statistical analysis and how widely these stats are being used by bloggers, media members and certain team's front offices. Stats like goals, assists, points, hits and plus/minus are what usually carry the most weight when it comes to judging a player, but factors such as luck and variance show that these stats shouldn't be the only things that are looked at.
With more information being available to the public through game play-by-play sheets on NHL.com and web sites like Behind The Net, everyone can look into the many different factors that play into the game of hockey and how much we can find out about a certain player's performance through them. Through these stats, we can find out things like how many shots a player attempts in a game, what kind of situation he is playing in, how much he is doing to generate offense and whether or not a bad season could have been due to luck.
This part of the game is still a work in progress and there are still a lot of kinks to be worked out, but there is definitely a lot you can discover about a team or a player when you look at some of the more in-depth metrics that are out there. Names like "Corsi," "Fenwick" and "QualComp" might be off-putting at first, but they are pretty easy to understand once you look into them.
These stats are something I specialize in on my blog Shutdown Line, and I will be sharing a lot of this information here on Canes Country. After the jump, I am going to explain some of these stats and apply them to the Carolina Hurricanes most recent season.
Before we dive into some of the raw data, let's take a look at how the Hurricanes players were deployed this season and what situations they were playing in. Not all ice time is created equal as some players are used strictly in defensive situations and may have the ice tilted against them more than a player who gets easier, offensive minutes. The type of situation a player is used in has an effect on his underlying numbers, which is why it will be helpful to look at this before going into the raw data.
To do this, I am going to use a method drawn up by Rob Vollman of Hockey Prospectus called an "OZ/QoC Graph." This is a basic scatter plot that compares a player's offensive zone start percentage (OZ%) with his corsi relative to quality of competition (CorsiRelQoC) to show the situations that he is being played in. OZ% is the percentage of shifts that a player starts in the offensive zone and CorsiRelQoC tells you what kind of opponent's a player is being used against with 0 being the average.
Offensive zone start percentage is on the x-axis and CorsiRelQoC is on the y-axis, which means the players on the top left are playing the toughest situations and the players on the bottom right are playing the softest minutes. Data is taken from Behind The Net Hockey.
Brandon Sutter, Patrick Dwyer, Tim Gleason and Bryan Allen are the guys who played the toughest minutes on the Hurricanes last year. Most who watch the team should know that as these four are regularly sent out against opposing team's top lines, especially at home when the Canes have the last change. A lot of people in the stats community like to call these kinds of players "heavy-lifters." Andreas Nodl and Chad LaRose also played some tough minutes last season and so did Jiri Tlusty.
Eric Staal, Joni Pitkanen, Jay Harrison, Justin Faulk, Drayson Bowman and Jerome Samson regularly played against tough opponents but were deployed more in the offensive zone than the heavy-lifters were. We usually say that these kinds of players are used against "secondary" competition because they are more known for their offense but the coaches trust their defensive game enough to not protect them. That's certainly the case with Staal as he has been used in this kind of role for the last few years. Jay Harrison is a new addition, though. He was used a lot more in the defensive zone during the 2010-11 season, but mostly against third and fourth liners. Faulk being trusted against secondary competition at only 19 years of age speaks volumes about his development and it's also nice to see that Bowman didn't need to be protected.
That provides a nice transition into the next category, which would be players who slightly protected at even strength. They are used against average competition and are deployed regularly in the offensive zone, usually a sign that the coach doesn't trust their defensive game as some of the other players. Jeff Skinner, Tuomo Ruutu, Jamie McBain and Jussi Jokinen fall under this category.
Finally, we have the guys who need to be protected from tough competition to succeed. This category usually consists of fourth liners, powerplay specialists and bottom-pairing defensemen who are more known for their offensive upside. The Carolina players used in these situations were Anthony Stewart, Zac Dalpe, Jaroslav Spacek, Derek Joslin, Zach Boychuk, Tim Brent and Brett Sutter. Brent and Sutter42 didn't need to be protected territorially, though.
Now that we have some context, we can start looking at the raw numbers, the first of which being "Corsi." To put it simply, corsi is the number of shot attempts at player is on ice for at even strength, which includes missed and blocked shots. This shows how well a team is performing territorially when a certain player is on the ice. A player's corsi rating is expressed through a differential rating similar to the plus/minus system used for goals or a ratio which compares the number of "corsi events" that occurred in favor of his team when he was on the ice with the number of overall "corsi events."
The main reason why stats like corsi are looked at is because they show how well a certain player could be driving the possession on his team. The general goal behind hockey is to get the puck into the opponents' end and create shots, which will hopefully lead to goals. A 20+ goal scorer is obviously valuable to a lot of teams but so are players who can play tough minutes and be capable of driving play into the opponent's end even if they do not score much.
Another stat we are going to look at are "scoring chances" which are shots directed at the net from what is defined as a "dangerous scoring area." Missed shots are generally counted for scoring chances, but blocked shots are not because the defense was able to deny the chance. Corsi is good for figuring out who is driving possession, but scoring chances give us a better idea of which players are creating more offense rather than just sending harmless 50 ft. wristers toward the net. I have been tracking scoring chances all season at Shutdown Line and 14 other bloggers have been doing the same for their respective teams.
Now that all of the explanations are out of the way, let's start breaking down the Hurricanes' numbers. Stats are courtesy of Vic Ferrari's Time On Ice site and the idea comes from Rob Vollman's weekly "Black Box" feature on FlamesNation.
ESP/60 = even strength points per 60 minutes, CEF = corsi events for, CEA = corsi events against, Corsi% = corsi ratio, SCF = scoring chances for, SCA = scoring chances against, SC% = scoring chances against, GF = goals for, GA = goals against, Goal% = goal percentage
When it came to scoring and creating chances at even strength, Jeff Skinner was the Hurricanes' best and most consistent player. His possession rates weren't particularly impressive for a guy who plays somewhat soft minutes, but he was creating scoring chances at a higher rate than every other forward. The team's regular forward with the next highest scoring chance percentage was Jussi Jokinen, who didn't score as much as he did in previous years but the team was creating a lot of chances when he was on the ice.
Speaking of guys who were creating offense but not scoring much, Chad LaRose also fell into that category last year. LaRose has never been a great finisher (i.e. he doesn't score a lot) but the guy played tough minutes, drove possession in the right direction and the team was creating more scoring chances with him on the ice. He's not an ideal top-six forward, but he proved to be one of the bright spots of the Hurricanes last season.
An interesting observation here is that Eric Staal and Tuomo Ruutu had similar underlying possession numbers but their scoring stats are where they differ. Staal was on ice for a TON of goals against (which I'll get into later) but scored at a slightly higher rate than Ruutu did. Staal actually had the second highest scoring rate at even strength on the team but the abnormally high number of goals against he was on ice for brought his goal percentage and plus/minus rating down.
The player with the second highest scoring rate at even strength was Anthony Stewart, which is interesting when you look at his underlying numbers but not as strange when you consider that his shooting percentage was the second highest on the team, bested only by Tim Brent. Both Stewart and Brent were getting pinned in their own zone and surrendering chances whenever they were on the ice but they struck gold whenever they shot the puck as over 15% of their shots ended up being goals. They were also getting matched up against weak competition so the fact that they couldn't keep the puck out of the defensive zone then isn't exactly a good sign for long-term success.
Brandon Sutter and Patrick Dwyer also have some pretty ugly looking possession numbers but remember, these two start almost 70% of their shifts in the defensive zone against other team's top lines, which means it's basically an uphill battle for them all the time. This is where raw data can be deceiving because players like Sutter & Dwyer are going to have low numbers based on their situations no matter what. To fix this, we can adjust their scoring chance ratings for zone starts, which I will address in a later post. For now, let's marvel over how good Dwyer was defensively this season. He was on ice for fewer scoring chances against than almost any other top-nine forward despite playing in tough situations. If he had a larger offensive upside, he would be one hell of a hockey player but he appears to be fine as a defensive forward on the Canes, as well.
A couple final notes on the forwards:
- Jiri Tlusty's shot metrics aren't terribly impressive, but he did a much better job at creating and preventing quality chances.
- While he was with the Hurricanes, Drayson Bowman's play was impressive. As a team, the Canes controlled about 47.5% of the scoring chances and having a player like him who could create chances and drive possession was a nice addition. He always seemed to play well even if it wasn't showing up on the scoresheet.
Let's move onto the defensemen now.
Only one Carolina defenseman had over 50% of the even strength scoring chances go in the team's favor and unsurprisingly, it was the most protected one in Jaroslav Spacek. When healthy, Spacek played his role as a third pairing/offensive defenseman perfectly. Jamie McBain nearly hit the 50% mark and he really seemed to excel when he was playing with Spacek on the third pairing. His role changed throughout the season but he managed to have a respectable year when all is said and done.
The other Carolina blue-liner who is also known for his offense, Joni Pitkanen, also seemed to play his role when it came to producing scoring chances and points but he gave up of chances in his own zone, as well. This is partially because he was starting in the defensive zone more than usual at the start of the year. With Tomas Kaberle needing to be protected, Pitkanen had to taken on some extra responsibility and gave up a hefty amount of scoring chances as a result.
Jay Harrison and Justin Faulk were one of the team's more active defense pairings at producing scoring chances, but they also gave up a lot against second line competition. That is expected when you have a 19-year-old and someone who never played top-four minutes until this season play in those kinds of situations. Faulk is still young and his play in his own zone is a work in progress but I have my doubts on whether or not Harrison is a top-four defender even if he had a career season.
Bryan Allen seemed to perform a little better than Tim Gleason as the heavy-lifter of the defense corps as Gleason was on ice for a ton of scoring chances and more shots against than any other player on the team. These two fall under the same category as Sutter & Dwyer since they play in really tough situations and adjusting their raw numbers for zone starts will give us a better idea of their performance.
I don't really have anything to say about Derek Joslin's numbers other than they aren't good. At all.
For goaltenders, we are going to evaluate their performance through even strength save percentage and the amount of "quality starts" they had. A quality start is another metric from Hockey Prospectus and is awarded to a goalie who gives his team a good chance to win, which is when they have a save percentage of about .912 or allow less than three goals depending on the amount of shots they faced. Even strength save percentage is usually looked at before anything else because most of the game is played there and it gives us the best idea of a goalie's overall talent. PK & PP save percentage are also looked at but they tend to be very random from year-to-year.
GS = games started, QS = quality start, ESSV% = even strength save percentage, PKSV% = penalty kill save percentage, PPSV% = powerplay save percentage
For the second year in a row, it was all Cam Ward all the time for the Hurricanes and he performed at about an average rate. He gave the Hurricanes a chance to win in over 60% of his starts, had an average even strength save percentage and seemed to get lit up on the PK a bit. It was definitely an up-and-down year for him if you look at his monthly totals but I'll get into that another time.
Brian Boucher was not good at all but he hardly played at all so it's tough to say that this is his true talent level. Still, it's hard to defend a goalie who played at replacement level and wasn't very lucky either. Justin Peters, on the other hand, was rock solid at even strength in all six of his starts but got torched on the PK. I guess that's expected with a somewhat young goalie, though. Like Boucher, it's hard to draw any conclusions about Peters because he played in only six games. That is way too small of a sample size to determine anything. I do find it interesting that the Canes saw great play out of one of their back-ups and replacement level goaltending from the other.
I think that should do it for starters. My analysis-based posts usually aren't this long but I wanted to post something to serve as an introduction to advanced stats for those who aren't familiar with it. I'm going to be exploring many different things about the Hurricanes over the next couple of weeks and will have an analysis of their performance on special teams up sometimes in the year future.