Toronto Maple Leafs v Florida Panthers

I can almost pinpoint the day that I first dove headfirst into advanced hockey analytics. It was at the conclusion of the 2009-10 regular season, and I was set to appear on a radio program as part of a panel to discuss our picks for the NHL Awards. With a couple of days notice before my appearance, and knowing who else would be joining the panel, I did some background work to see who my peers would be picking for individual awards. When it came to the Norris Trophy, I discovered that the two other panelists were leaning toward Mike Green of the Washington Capitals. I was all-in on Duncan Keith, the eventual recipient.

Knowing full well that the Mike Green crowd would be pointing to his superior point and +/- totals as reasoning to go with the Capitals defenseman over Keith, I set out to learn more about the analytics that you can find on Armed with a few days of research and a slightly more than fundamental understanding of how stats like Corsi, quality of competition, and zone starts are calculated and weighted, I felt like I effectively argued how Keith was a superior defenseman.

Today, advanced stats have become a major part of my daily player analysis routine. Of course, a balance between watching games and analytics is optimal when formulating an opinion on a specific player or team. I wanted to do something important. Like, the most IMPORTANT thing ever. So, I took it upon myself to apply this same analytical approach to the most barbaric aspect of hockey: PUNCH FIGHTS. Anybody can look up a scrap on and cast a vote for a winner, but, much like goals don’t tell you the whole story, a knockdown punch or homerific voting system can’t tell you everything about a fight.

So, I’ve spent far too much time over the last three weeks watching hockey fights and counting punches exchanged in an effort to look at who among the NHL’s fighting major leaders controls the battles.

Enter: Punch Corsi


A Corsi number can be summarized as the difference in shots attempted by the player’s team and the shots attempted against the player’s team in standard 5v5 situations. The stat is named after its inventor, and Buffalo Sabres goaltending coach, Jim Corsi. Corsi essentially measures puck possession. Although it is particularly useful in measuring possession, Corsi is best understood in context with numerous other advanced analytics and their varied incarnations, such as quality of competition, Fenwick, PDO, zone starts, etc. When used properly, advanced analytics help paint the bigger picture. Both Cam Charron and Daniel Wagner have written some excellent pieces on advanced stats here, and use them effectively in their day-to-day work around the hockey blogosphere.

Punch Corsi, while certainly not telling the whole story, attempts to analyze who controls the fight. Throwing the most punches does not necessarily indicate control, as tossing wild haymakers is bound to leave a brawler vulnerable, but it gives us an indication of punching ability. For instance, the NHL’s fighting major leaders thus far in 2013 are B.J. Crombeen and Jared Boll with ten and nine, respectively. If you’ve watched either of those guys fight, you would know that they usually take more shots than they dish out. Guys like Mike Brown and Jordin Tootoo, though, usually appear to toss more than than they receive. Heading in to this project, I assumed that either Brown or Tootoo would likely boast the highest punch Corsi number.


Watch fights, count hands thrown, drink beer. That’s pretty much it. “Jersey jabs” were counted as punches thrown/attempted so as long as a distinct punching motion existed.

Punch Corsi is calculated as follows:

Punch Corsi Number = (Punches on Target For + Missed Punches For + Blocked Punches Against ) – (Punches on Target Against + Missed Punches Against + Blocked Punches For)

For the purpose of this study, only players with a minimum of six fighting majors (as of March 20, 2013) in this current season were included.


Player Fighting Majors Punch Corsi Number
Mike Brown 8 48
Colton Orr 8 36
Frazer McLaren 6 22
Dale Weise 7 12
Jordin Tootoo 7 4
Tom Sestito 6 1
George Parros 6 0
Mark Fraser 7 -4
Rich Clune 7 -7
Zenon Konopka 7 -7
Chris Thorburn 7 -19
Brandon Prust 7 -21
Matt Hendricks 6 -22
Jared Boll 9 -28
B.J. Crombeen 10 -41
Krys Barch 6 -51


As I had expected, Mike Brown posted the highest punch Corsi number. This is because Mike Brown’s fighting style is what I like to refer to as “cocaine fists”. Punch now, punch fast, think never. I mean, seriously, look at this animal throw punches:

No surprise here that punch Corsi dictates both Jared Boll and B.J. Crombeen are human punching bags. Krys Barch should probably find another line of work. Maybe it’s a case of naivety on my part, but I was rather surprised by how well Colton Orr rated. I was under the assumption that Orr was more of a power puncher, but his punches thrown/attempted are up over 20 per fight, and his opponents routinely looked like they were doing little more than struggling to hang on. Colton Orr is a bad, bad man. I can’t say for sure, but it would appear that the Toronto Maple Leafs place a lot of value in punching.

Also, body shots are back like it’s a Mexican themed freshman party.

It’s difficult to place a value on punch Corsi in relation to the outcome of a fight, although, in the majority of cases the player who directed the most fists toward his opponent was the victor (as per voting system). Perhaps there’s an intimidation factor that could be tied to punches thrown/attempted. I am not qualified to speak to the psychological impact, however, neither are most people.

In Conclusion

While it was certainly fun to watch and count punches in over hundred scraps, punch Corsi does not tell us everything we need to know about hockey fights. Hockey fights require more in the way of advanced analytics to be properly contextualized. Early on in the research process, it had dawned on me that punch Corsi is rather useless without some accompanying statistics. Actually, punch Corsi is rather useless altogether, but it was fun as fuck to do this. I’d like to develop and include some new hockey fight analytics in the future, such as fight quality of competition, and another metric minus the blocked punches for and against. I’ll probably name the latter after a commenter at, ‘braindog1017′ and ‘chad420′are current favourites for a namesake.

Punch Corsi, fight quality of competition, and several other fight metrics will all be featured on my yet to be opened website tentatively titled This website will come to fruition just as soon as I can convince my mother to lend me her credit card to purchase the URL.

When I present these findings to the New England Journal of Sports Information, I will require a name for punch Corsi that is unique. I hereby nominate ‘Kordic‘ as the proper name for punch Corsi. Now if you’ll excuse me, I’m going to spend the rest of my time on earth not watching hockey fights.