Tom Tango and I wrote this piece a while ago, but today seemed as good a day as any to post it.
START
1979 was a momentous year for professional hockey: safety finally trumped long-standing tradition, and National Hockey League President John Ziegler announced that helmets would be mandatory for all players entering the league. Since then, the best-of-seven Stanley Cup Finals have gone to a seventh and deciding game eight times. Notably absent from those eight games? Despite their status as another long-standing tradition in hockey, there were zero fights in these high-stakes matchups.
Nor have there been any fights in the final games of the Canada Cup, World Cup of Hockey or the Olympics. In international play, fighting leads to ejection, so it should come as no surprise that fights are almost non-existent.
But regardless of the punishment for fighting, high-quality games can occur with no fighting, nor threat of fighting. There’s no need to change the punishment system against fighting: we simply need to change the incentive system.
***
It should be clear that there are also great games that did feature fighting. Jarome Iginla fighting Vincent Lecavalier in the Stanley Cup finals did not detract from the excitement of that series. Indeed, as one of the indelible moments of that series, it may have contributed to it. But Iginla and Lecavalier however were not enforcers.
The common thread among great games is good players doing great things. We want to see the most talented players as much as possible, while limiting the playing time for the rest. And, when it comes to enforcers, the playing time shrinks to something non-existent in great games.
***
Fighting in the playoffs occurs far less frequently than it does in the regular season. Why is that the case? To begin with, enforcers play less frequently in the playoffs than in the regular season. Enforcers provide, theoretically, both a positive and negative effect. The most direct negative effect is that teams allow significantly more goals than they score when they ice an enforcer. On the plus side, a possible but indirect benefit is that enforcers spur on their teammates. While coaches can live with that kind of tradeoff in the regular season (82 games, 16 teams in the playoffs), they likely reason the tradeoff is too negative, with their limited margin of error (lose a series, eliminated
from playoffs, potentially lose job).
***
The NHL is able to limit fights without actually having to change any rules. The league simply creates a different atmosphere (the playoffs) that incentivizes the coaches and players to limit fighting on their own. We don’t even need to necessarily adopt the harsher system in international play.
***
It’s helpful to understand when fights actually happen in regular season NHL hockey games. Even though only one-sixth of the game is played with a 3-goal lead or more, 27% of all fights occur then, with fighting much more likely when the home team has a big lead.
|
Home Team Lead |
Down 6 or more |
Down 5 |
Down 4 |
Down 3 |
Down 2 |
Down 1 |
Tied |
Up 1 |
Up 2 |
Up 3 |
Up 4 |
Up 5 |
Up 6 or More |
|
Fights per 60 Minutes |
0.63 |
0.69 |
0.78 |
0.65 |
0.46 |
0.41 |
0.45 |
0.45 |
0.48 |
0.78 |
0.96 |
1.05 |
1.27 |
Despite claims from the proponents of fighting that it serves as a deterrent for dirty play – after all, a player who commits an egregious stick foul or hits another player from behind will have to fight his next time on the ice – fights seem much more likely to result from a road team’s coach sending a goon out on the ice to take out the team’s frustrations during a big loss.
***
It should be clear that when we talk about goons, we aren’t merely talking about players who fight or take a lot of penalties; our concern is the “pure” enforcer. Some players with high penalty minute or fight totals, are often skilled, physical players who can both fight and play hockey. The player we’re looking for is the one whose existence is so often defended by hockey insiders: he has high fight totals but so lacks other talents that he is never used on special teams. These players specifically train as fighters, before even entering the NHL.
Our goon list consists of 34 players: anyone who finished in the top 20 in the NHL in fights in any of the last four seasons while averaging less than 20 seconds of power-play time per game. Some of the names are famous – Donald Brashear, George Parros, Georges Laraque, Paul “BizNasty” Bissonnette – but everyone on the list has enjoyed a lengthy career as an NHL enforcer.
So what kind of on-ice work do those lengthy careers involve aside from fighting and not playing the power-play? Their usage is primarily at even-strength, where they average roughly eight minutes and twenty seconds per game, a figure that puts most of them dead last in ice time on their respective teams. Overall, they were on the ice 1.82% of the time in the first period, 1.72% in the second, 1.5% in the third, 1.37% in the second half of the third and 1.34% in the last two minutes of regulation.
This is hardly damning given how teams shorten the bench late in a game, but their case becomes more extreme when you consider the game state: in the last two minutes of the game, they were than three-and-a-half times more likely to be playing if one team had a lead of three goals or more than in a tie game and five times more likely than in a one-goal game. And that’s the story of the goon: as essential a part of a successful team as hockey observers claim he is, he is almost never on the ice when the game is really on the line.
If you slice the game state more narrowly, the case for the goon is even murkier. Not all even-strength time is created equal – most of the regulation game is played at 5-on-5, but coincidental penalties can result in 4-on-4 situations. With each team a man down, our goon squad is three times less likely to be on the ice than at 5-on-5. And in overtime, which is played 4-on-4 to increase offensive play, they are 13 times less likely to be on the ice than they are overall, and 23 times less likely to be on the ice than in the last two minutes of a blowout, their prime time for causing trouble.
***
Playoff hockey is again a different story. With a high stakes seven-game series, teams dress their best rosters, regardless of injury status. Goons account for almost 40% less even-strength playing time than they do during the regular season, primarily because they don’t even dress in the playoffs.
Perhaps unexpectedly, goons play more in overtime in the playoffs than they do during the regular season. This is unavoidable: unlike regular season overtime, which is limited to five minutes, playoff overtime can go on forever, lasting on average 11:39 compared to 3:55 during the regular season. In a 70- or 80- or even 100-minute game, a team can’t afford to have any of its players sitting on the bench for an entire period. In the first five minutes of 4-on-4 overtime play in the playoffs, goons don’t play any more than they do during the regular season. After the first five minutes of OT, goons play five times as often – the unfortunate outcome of dressing a goon in the playoffs is that a coach will actually need to use him in the highest-stakes situation possible.
***
Not every team buys into the need for a goon on its roster. The Detroit Red Wings, the 2007-08 Stanley Cup champions and the second-winningest team during our four-year study, did not employ a single player from our goon list in any of those seasons.
***
Imagine therefore that we change the atmosphere of a regular season NHL game so that the coaches are incentivized to bench their enforcers. What if the NHL adopted a four- on-four system while the game is tied? This will just about eliminate any ice time for enforcers during tied games.
Since they are already not playing much otherwise, teams may simply decide that they can’t afford to have a spot on the bench go unused for long stretches of a game. In addition, the four-on-four-while-tied scenario will give us a more open game of hockey. With goons almost completely eliminated from tie games, their average ice time would likely drop from eight minutes to below six minutes per game. Over the last four years, only 20 players have recorded so little ice time while playing more than 40 games in a single season and their full-time playing careers have generally ended immediately after. Even this subtle a reduction in enforcer playing time will likely result in significantly fewer enforcers employed league-wide.
END
Footnote: Stanley Cup Final Game 7s (winner penalty minutes, loser penalty minutes) in brackets.
’86-87: Edmonton 3 – Philadelphia 1 – no fights (14, 14)
’93-94: NYRangers 3 – Vancouver 2 – no fights (8, 10)
’00-01: Colorado 3 – New Jersey 1 – no fights (8, 12)
’02-03: New Jersey 3 – Anaheim 0 – no fights (2, 4)
’03-04: Tampa 2 – Calgary 1 – no fights (6, 10)
’05-06: Carolina 3 – Edmonton 1 – no fights (10, 12)
’08-09: Pittsburgh 2 – Detroit 1 – no fights (4, 6)
’10-11: Boston 4 – Vancouver 0 – no fights (2, 4)
Footnote: the complete list of NHL enforcers:
Krys Barch, Paul Bissonnette, Jared Boll, Eric Boulton, Donald Brashear, Wade Brookbank, Mike Brown, Adam Burish, Matt Carkner, Kyle Clifford, Riley Cote, B.J. Crombeen, Derek Dorsett, Deryk Engelland, Tanner Glass, Eric Godard, Matt Hendricks, Cam Janssen, D.J. King, Zenon Konopka, Ian Laperriere, Georges Laraque, Jamal Mayers, Cody McCormick, Cody McLeod, Colton Orr, George Parros, Brandon Prust, Rick Rypien, Jody Shelley, Brad Staubitz, Zack Stortini, Shawn Thornton, Kevin Westgarth
Footnote: Regular Season Enforcer Playing Time by Game State, Relative to Average:
|
5v5 |
All |
1st |
2nd |
3rd |
Last 10 |
Last 2 |
OT |
|
Score 3+ |
1.351 |
1.272 |
1.224 |
1.424 |
1.517 |
1.853 |
|
|
2 |
1.021 |
1.177 |
1.129 |
0.900 |
0.762 |
0.614 |
|
|
1 |
1.058 |
1.215 |
1.137 |
0.829 |
0.651 |
0.399 |
|
|
0 |
1.121 |
1.186 |
1.176 |
0.880 |
0.725 |
0.519 |
|
|
4v4 |
All |
1st |
2nd |
3rd |
Last 10 |
Last 2 |
OT |
|
Score 3+ |
0.461 |
0.468 |
0.310 |
0.544 |
0.638 |
1.110 |
|
|
2 |
0.319 |
0.295 |
0.325 |
0.321 |
0.297 |
0.324 |
|
|
1 |
0.325 |
0.359 |
0.307 |
0.316 |
0.285 |
0.192 |
|
|
0 |
0.347 |
0.321 |
0.396 |
0.337 |
0.363 |
0.350 |
0.080 |
Expressed as an odds-ratio – e.g. a goon is 1.853x times as likely to be on the ice at 5-on-5 in the last two minutes of a blowout (3+ goal lead) than he is overall. He is 0.08x as likely to be on the ice during overtime.
Footnote: Playoff Enforcer Playing Time by Game State, Relative to Average:
|
5v5 |
All |
1st |
2nd |
3rd |
Last 10 |
Last 2 |
OT |
|
Score 3+ |
1.631 |
2.160 |
1.569 |
1.622 |
1.743 |
2.320 |
|
|
2 |
0.986 |
1.261 |
1.367 |
0.786 |
0.481 |
0.548 |
|
|
1 |
0.952 |
1.334 |
0.986 |
0.617 |
0.420 |
0.150 |
|
|
0 |
1.154 |
1.272 |
1.121 |
0.911 |
0.875 |
0.339 |
0.907 |
|
4v4 |
|
|
|
|
|
|
|
|
Score 3+ |
0.398 |
0.000 |
0.554 |
0.347 |
0.447 |
0.811 |
|
|
2 |
0.340 |
0.182 |
0.504 |
0.224 |
0.135 |
0.000 |
|
|
1 |
0.329 |
0.403 |
0.211 |
0.420 |
0.369 |
0.000 |
|
|
0 |
0.497 |
0.547 |
0.452 |
0.418 |
0.567 |
0.000 |
0.651 |
Some game states occur very rarely in the playoffs, so 4-on-4 data in particular is less reliable.
