The True Impact of a Hockey Fight

Fighting continues to be one of the defining traits of hockey. When a player puts his body on the line in a hockey fight, there is often an expectation that the fight will make a difference in the score of the game. If a player destroys an opposing player in a fight, it is largely accepted in the hockey world that that team will use the physical victory to motivate itself, and raise its level of play as a result.

This assumption is so common and ingrained in hockey culture that it can even be seen in NHL video games, where winning a fight provides energy to the winning team after the fight and boosts the in-game ratings for the players on that team. Of course, as with any assumption in professional sports, it is important to determine whether or not it holds up in practice. Do losing teams tend to win more fights by sending out goons to turn the tide of the game? Does winning a fight actually result in a change in the game’s “momentum” and a better result for the more successful combatant?

There has been some research done on the impact of fighting, but rarely has it examined actually winning the fight as a potential catalyst for victory. Research done and data collected by Mark Appleby has found that fights can have a positive impact on goals scored (by both teams) slightly in the time directly after a fight.  Phil Birnbaum, also looked into this concept and found that the fight data was not statistically significant in causing an increase in goals and that the effect of fighting was negligible in terms of actually impacting the game. There is still, however, significant room to look at how games play out following fights. Fans often make a big deal about having fighters who can actually lay a beat down. Are they right to do so?

It is difficult to draw conclusions about the impact of winning a fight in hockey because there is no official procedure for determining a winner. Luckily, there is a website that allows for large quantities of hockey fans to decide who they feel is the winner of every in-game fight that occurs.

Spend five minutes on the website hockeyfights.com and you will find a community in which the the spirit of fighting in the NHL is alive and well. Every user of hockeyfights is able to watch each fight and directly vote for a winner and give the fight a match rating. With anywhere from 100 to 300 votes per fight, the sample size is large enough for us to use the site as a credible means to test whether winning a fight has a positive impact on a hockey game for the victor’s team.

We compiled the data from all 342 fights from the 2012-2013 NHL season and the score at the time of the fight and at the end of the period. By recording this, we were able to grasp whether winning the fight caused a boost to a team’s “momentum”, and allowed the team that won the fight to have a better goal differential after the fight than before the fight. We made four groupings of the 342 fights. The first is composed of every hockey fight in this database. The “good fights” group is composed of the fights that were voted by the community as having a rating of higher than 5 — vicious and drawn out fights that one might hypothesize as having a larger impact on the score. “Clear winners” is the group of fights in which the community has very obviously picked a winner of the fight. Here, one player and one team would clearly be the victors of the fight, which might also contribute the most “momentum”. The “clear and good winners” group combines these two restrictions.

Every Fight

	Total Two Teams Combined Goals Scored (Goal Sample Size)	Percentage of Goals Scored By Team Winning Fight	Standard Deviation
From Start of Game to Start of Fight	680	54.6%	1.9%
From End of Fight To End of Period	341	50.7%	2.7%
From Start of Game to End of Fight Period	1021	53.3%	1.6%

Good Fights (Ft. Rating >5)

	Total Two Teams Combined Goals Scored (Goal Sample Size)	Percentage of Goals Scored By Team Winning Fight	Standard Deviation
From Start of Game to Start of Fight	290	57.9%	2.9%
From End of Fight To End of Period	191	50.8%	3.6%
Total from Start of Game to End of Fight Period	491	55.1%	2.3%

Clear Winner (Wins by at least 25% and with >40% of the vote)

	Total Two Teams Combined Goals Scored (Goal Sample Size)	Percentage of Goals Scored By Team Winning Fight	Standard Deviation
From Start of Game to Start of Fight	505	55.3%	2.2%
From End of Fight To End of Period	279	50.4%	3.0%
From Start of Game to End of Fight Period	784	53.7%	1.8%

Clear Good Winner (Ft. Rating >5 and Wins by at least 25% and with >40% of the vote)

	Total Two Teams Combined Goals Scored (Goal Sample Size)	Percentage of Goals Scored By Team Winning Fight	Standard Deviation
From Start of Game to Start of Fight	243	58.9%	3.2%
From End of Fight To End of Period	142	49.3%	4.2%
From Start of Game to End of Fight Period	385	55.3%	2.5%

We can see in all four of these data sets that there is no evidence that winning a fight leads to better results in the immediate aftermath of the fight. In fact, it appears that the team winning the fight will score slightly less goals in the game than they did previously. In all four groups, the percentage of goals scored by the winning team is within 1.96 standard deviations of the percentage of its total goals scored, indicating that the results are negligible. Thus, we cannot reject the null hypothesis that winning a fight has no impact on a team’s momentum and goal differential.

Surprisingly, what we can best draw from this particular data set is the idea that teams that have a better goal differential tend to win more fights. In fact, in all of these data sets, the percentage of goals scored by the fight winner’s team by the end of the period is slightly outside the 95% confidence interval from the null hypothesis that fighting is independent of goal scoring (that the fight winner would score 50% of all goals, and the fight loser would score 50% of all goals).

	Percent of Goals Scored By Fight Winning Team Total	Standard Deviation	95% Confidence Interval
Every Fight	53.3%	1.6%	(50.1, 56.5) – Reject Null of 50%
Good Fight	55.1%	2.3%	(50.5, 59.7) – Reject Null of 50%
Clear Winner	53.7%	1.8%	(50.1, 57.5) – Reject Null of 50%
Clear Good Winner	55.3%	2.5%	(50.3, 60.3) – Reject Null of 50%

An additional conclusion that can be drawn from this data set is an idea that also seems to buck conventional hockey wisdom. The cause of fights has long been described as the losing team sending out a player with the intention of winning a fight to turn the tide in a game. In all four groupings, however, the share of goals scored before the fight by the team winning the fight is significantly larger than the team losing the fight. Either the idea of losing teams instigating fights isn’t as accurate as it is believed to be, or the gap in physicality between teams that are winning and teams that are losing is actually larger than expected.

In the 2012-2013 season, there was no tangible evidence that winning fights affected the outcome of hockey games, and the idea that losing send more goons onto the ice in order to claim momentum needs to be questioned. There is a slight correlation between winning fights and scoring more goals than the opponent, but any possible causation would require further examination.

Data from http://www.hockeyfights.com/

Image courtesy of Jeff Gross / Getty Images

Xavier Weisenreder

Georgetown University Class of 2016

Arik Parnass (2015) and Tyler Deloach (2016) also contributed to this article

Follow Xavier on Twitter: @BeMoreChillNext
Follow Arik on Twitter: @ArikParnass
Follow Tyler on Twitter: @TyDeloach
Follow GSABR on Twitter: @GtownSports
Like GSABR on Facebook