An Introduction to Advanced Stats
By Mike Majeski
I had a moment of clarity last night regarding advanced stats in hockey. I have never really been a stats person, as I tend to view the game of hockey more in the emotional realm. I could tell you that Bobrovsky had an amazing season in net and that Vinny Prospal was a force on offense as well as the glue that kept the team together. However, what I have never been able to do is understand the stats and data that could either back up or disprove the above statements.
Mandatory Credit: Anthony Gruppuso-USA TODAY Sports
Corsi is an advanced stat that measures puck possession for 5 on 5 play. Seems simple enough, but every description I have read before ended up becoming too complex and made my head spin.
After a few Google searches, it was two blog posts on a Sens and Flames blog that broke through and helped me gain a basic understanding of Corsi, Relative Corsi, CorsiRelQoT, and CorsiRelQoC.
In one simple sentence, the Silver Seven post summed up Corsi as:
"The difference in shots attempted by the player’s team and the shots attempted against the player’s team."
The post went on to break it down into a simple to follow formula. (All scaled to per 60 minutes) :
Corsi Number = (Shots on Target For + Missed Shots For + Blocked Shots Against ) – (Shots on Target Against + Missed Shots Against + Blocked Shots For)
[ Courtesy of Silver Seven Blog ]
To actually use the stat to measure a player’s value to his team, you need to compare the players individual Corsi to that of the entire team. This stat is defined as Relative Corsi. The formula for Relative Corsi is:
Relative Corsi Number = Corsi Number of player – Corsi Number of Team when player not on ice [ Courtesy of Silver Seven Blog ]
If you do not want to sit there with your pencil and paper and calculate these stats yourself, Behind The Net ‘s Player Breakdown is a great tool to view and compare these stats for every NHL roster. They do all of the calculations, all you have to do is understand what the data means.
For this post’s example we will use Cam Atkinson from the 2012-13 season. When viewing the Blue Jackets for the 2012-13 season, I would recommend setting the limit to at least 20 games played, as some AHL call ups will cause inaccuracies due to small sample size.
Mandatory Credit: Charles LeClaire-USA TODAY Sports
Last season, Cam Atkinson had a Corsi number of 7.48 and led the Blue Jackets with a Relative Corsi of 20.7. So for every 60 minutes that Atkinson was on the ice, the Jackets were better off by a net 20 shots attempted. There are more stats that can help us understand Atkinson’s high 20.7 Relative Corsi.
CorsiRelQoT and CorsiRelQoC measure the Relative Corsi of the opposition players, as well as the Relative Corsi of Atkinson’s team mates that are on the ice with him. The Flames blog Match Sticks and Gasoline defines a high CRQoT and CRQoC as values greater than .75.
A high CRQoT would mean Atkinson was helped out by high talent line mates, and a low CRQoC would mean he faced lower talent opponents.
For 2012-13, he had a CRQoT of 1.681 and a CRQoC of 0.578, meaning very high quality of talent for his team mates and a lower quality of talent for the opposition. From this we can determine that while Atkinson has a high Relative Corsi, he was often surrounded with higher talent line mates and faced the opposing team’s lower lines. Atkinson’s high Relative Corsi means he was a valuable asset to the Blue Jackets, but not the elite NHL scoring threat that that number would suggest.
This was just an introduction to advanced stats, but throughout the 2013-14 season I will use Corsi to analyze the Blue Jackets performance, as well as examine the other advanced stats that exist. I am far from completely understanding how to use these statistics to evaluate NHL talent, but I feel some basic knowledge has been attained. I hope this post, as well as the articles referenced helped you gain some understanding as well.