What's the Best Use of a Fourth Line?

What's the Best Use of a Fourth Line?

Last time, we talked about different options for how to deploy your forwards. Most of those strategies leave the team with a fourth line- a place where the team stashes cheaper players who don't play as much. How these players are used belies a coach's overarching philosophy.

We're going to break down all of the possibilities, do some math to figure out the optimum impact of each, and figure out what's best. It will involve a lot of talk of expected goals, or xG- basically, given how many shots a team gets and where they come from, how often you would expect them to score. The key thing to look at is expected goal differential, because this gives an idea of the kind of impact a player or line has on their team. If we imagine every team as starting at .500 and 0 expected goal differential, 6 expected goals up is worth a win above .500, and 6 expected goals up is equivalent to losing a game.

Goon Line

We're going to get this one out of the way first. It's what your uncle wishes your favorite team would do. Just get the three meanest, nastiest bastards who can skate and have them pummel the other team with their fists.

It's been litigated to death at this point, but you simply cannot punch the puck into the goal. Fights, inherently, cannot win hockey games. The best it can do, based on academic research, is lead to a slight boost in shot rates for both teams for a shift or two after the fight. It comes out to about 0.16 additional shot attempts per minute for both teams in the 5 minutes following the fight (my source for that has gone away, you'll have to take my word for it). If our team has better goaltending, that could be a marginal gain.

The average shooting percentage for all shot attempts last season was 4.28% last season. If our goalie is 8th in the league, halfway between average and best, the other team would be shooting just 3.96% on us, giving us a slight advantage. That means that for a fight, it would be worth about 0.00256 of a goal based on goaltending alone.

But we know it does more than that. Fighting is dangerous, not just in the long term, but for immediate health. Concussions are the injury that comes to mind, with broken hands, noses, and jaws being common injuries as well. Last season, there were 757 instances of players missing a game from injury (we're not counting illnesses or cardiovascular problems as injuries), and it's been suggested that 10% of injuries, or 75, come from fights. I don't know that it's possible to know for sure since the reporting on injuries is pretty suspect, so let's run with that. There were 311 fights last season, which means there is a 24.1% chance a player is injured in a fight.

Let's be cynical. Machiavellian. Frankly unsportsmanlike. We want to see the maximum value of a goon line, so let's see how much value that injury impact can have. If we have good fighters, we can assume that they will dish out 2 injuries for every 1 they take. Naturally, a player who can get into fights with opposing star players, or at least impact players, is going to be more valuable than one who only fights other fourth-liners. The average expected goal value for a top-9 forward is 0.75 individual xG per 60 minutes, and for a top-4 defenseman is 0.18. That's an imperfect way to measure how much we're taking away from the opposition, but it's straightforward and we're already doing a lot of mathematical gymnastics to get here.

If our goons randomly pick out one of those impact players, there is a 9/13 chance it'll be a forward and a 4/13 chance it'll be a defenseman. We can weight the average individual xG by position, giving us 0.46 xG per 60 minutes. If those players are not injured, they'll be taken out of the game for 5 minutes, leading to a 0.038 expected goal value. If they are injured, we need to figure out an average for how long they'll be out of the game. Fights generally do not happen in the last 10 minutes of the game. Assuming a normal distribution across the remaining 50 minutes, the average fight would occur at about 25 minutes, causing our victims to be injured for 35 minutes, leading to a 0.26 xG lost value.

Mushing together the numbers, we can figure out that having one of our fighters take out an opposing impact player will result in them being taken out of the game for 9.42 minutes on average, assuming they can hurt their opponent twice as often as they get hurt and they have the average rate of injury. Using their individual xG value, that means we'll expect to score 0.072 more goals than the opposition from injury and penalty time.

Our goons are probably not going to be the best hockey players, or else they wouldn't be cheap enough to fit on our imaginary fourth line. The median NHL replacement level player, the sort we could expect to get plucking someone from the AHL, has -0.38 expected goal differential per 60 minutes. To be reasonable, let's say each of our goons is worth -0.4. They're not being recruited for their hockey ability, after all. That gives our goon line -9.84 expected goal value for the whole season if they play 6 even strength minutes per night- you don't want these guys out there longer than that. That's a lot of ground to make up for with punching.

If we're smart, we'd only have a line like this if it was completely beneficial to us, so we have a better goalie and we're only looking to fight opposing skilled players. That gives us 0.075 expected goals per fight. If each of our goons can get into 5 fights a year, we'd get 1.15 goals. If they can go off and get 8 fights per year, that's worth 1.8 goals. If they are all Dave "The Hammer" Shultz and get in 26 fights each, they would get the team about 5.85 goals. It's not the 70s anymore, so that's not even possible, barring some wild changes to both the sport and our understanding of CTE.

Taking into account that we're looking at sub-replacement level players, with an achievable number of fights that's a net loss of 8.72 goals per year, or one win in the standings.

Our analysis ignores something vital though- fighting doesn't take much time off the game clock. If your plan is to brawl, your 4th liners can get it done pretty quickly. If they do, they don't even need to play 6 minutes- in their heyday, it wasn't uncommon to see enforcers play as few as 3 minutes a game. Because of the stoppage in play from the fight, your other players are still getting their rest, and your star forwards are going to get the extra ice time that was going to go to your 4th line if you played it 10 minutes a game like everyone else.

At just 3 minutes a game, the impact of our goon line goes down to -4.92 expected goals from their hockey. Let's be generous and say our guys can find 10 fights each against more valuable opponents, which gives us a 4th line xG of -2.67. Pretty negligible. But with 7 extra minutes going to our skilled forwards, that adds some value too. We already determined a skill forward would average 0.75 xG per 60. We're playing 3 of them for 7 extra minutes, giving our team a bonus 0.26 xG. If we want to give the credit for that to our pugilists sitting on the bench, that makes our fourth line have a net value of -2.41 xG.

Not optimal, but teams have certainly iced worse lines in history. There is definitely room to incorporate tactical fighting into a gameplan- it has marginal value. But if those players aren't bringing anything else, you're hurting your team, especially if you're trying to make an ice-boxer play hockey too.

Checking Line

These days, this is the traditional use of a fourth line. Have some guys who can cycle the puck around, slow down the other team's scoring, and use some of that classic 4th line grittiness to annoy the other team and get you some powerplays. If scoring is what costs you salary cap dollars, then this is the perfect way to maximize value at the bottom of your roster.

We're going to get out the time machine to build this line- I'm sorry, these are just the players I think of when I think of this kind of unit. Think of it as a kind of Ur-example of a checking line.

Meet J.T. Brown. Currently, he's a broadcaster for the Seattle Kraken, but he was formerly an excellent defensive forward who put his team on the powerplay a lot. In 2017-18, he drew 1.07 more penalties than he took per 60 minutes, all minors. He also managed 0.11 expected goal differential per 60 from his hockeying, so his net value for that season would be about 4.45 goals.

Of the same vein, nobody embodies the ethos of "can't score, but don't let the other team score" better than Dimitrij Jaskin. He wasn't much of a shooter, but he was a great defender and almost impossible to knock off the puck. In his prime, he could create 0.29 expected goal differential per 60 minutes, across a whole season on our fourth line, we'd be looking at 3.96 goal differential.

And finally, let's add Lance Bouma. He was known for his shot blocking, and while many consider that to not be a valuable skill, he was good enough to generate 0.23 expected goal differential per 60 minutes. That's 3.14 goals per season.

This unit would generate 11.55 expected goals per season, almost enough to create 2 wins in the standings.

There's not much to say about this kind of setup. It's reliable, it's cost-efficient, and it's probably what your team does.

Short Yardage Line

What if we did the exact opposite? Let's bet on our own team's success and try to thrive off of it. If we have dominant offensive players, they are going to be finishing a lot of their shifts in the offensive zone. That gives our team an opportunity.

Winning an offensive zone face-off gives you a pretty big boost to your shot metrics for the next few seconds. For an individual face-off, that's not a huge deal, but over the course of a season it can be pretty big. In an average game, a team can expect to get about 13 offensive zone face-offs at even strength. Let's throw our fourth line over the boards 6 times out of those.

If we can put some kind of face-off mutant on our team, we can expect to win 59 percent- that's about how much historical aces like Samuel Pahlsson, Manny Malhotra, Zenon Konopka, and Paul Gaustad would win. Over the course of the season, that should give us 44.28 more face-off wins than a 50% face-off taker. According to Arctic Ice Hockey's Gabriel Desjardins' math, 100 offensive zone wins is worth 2.45 goals, so our face-off ace has gotten us 1 more goal than our opponents in overall value.

If we have such a good face-off taker, they might as well take some neutral zone face-offs too. They're less likely to lead to goals, but more numerous. In a game, a team could expect to have 1,140 per season, or 14 per game. Let's give our guys another 4 face-offs here, giving them about 10 minutes of ice time per game. Our puck-winner can expect to get us an additional 29.52 face-off wins, translating into 0.26 expected goal differential.

And you know what? A defensive zone face-off is just an offensive zone face-off for the other team. If we can win those, we're taking away their post-face-off shot boost. If we give our 4th line 4 of our d-zone face-offs, then we get an additional 29.52 face-off wins per season, worth 0.72 more goals each season.

But we can do better than just winning the face-off, so let's figure out the wingers. We need wingers who are going to generate offense at an above average rate, but who might get relegated to the edge of the lineup for whatever reason. Sometimes teams are looking for checking line players, or they have a bias for tall players, or they just make mistakes in talent evaluation. The reason doesn't matter.

One would think that in the age of analytics departments that there wouldn't be too many of these guys kicking around. But you'd be surprised. Just in 2023-24, any team could pick one up easily.

Two obvious examples Daniel Sprong or Sonny Milano. Both have an excellent track record of winning their team possession and of capitalizing on scoring chances, and yet find themselves on the outside looking in much of the time. Sprong has bounced back and forth between the third and fourth lines on every team he's played on and Milano has been a healthy scratch much of this season with Washington. 

Sprong literally got traded for nothing this year- talk about bargain value: he scored 1.47 goals per 60 last season and it was a down year for him. In terms of net impact, he generates 0.15 expected goal differential per 60. With the Capitals' salary cap situation, they might even pay you to take Milano and his $1,900,000 cap hit, despite helping his team with 0.21 expected goals per 60 minutes. But there's even more of these guys kicking around in the minors. Martin Frk is known o have a pretty good slap shot, but his real value is that he has scored 0.91 goals per 60 minutes in the NHL. Josh Ho-Sang is another example- for whatever reason, teams don't want to have him, but when he was in the NHL, he was putting up 1.41 primary points per 60 and 51% possession.

Let's imagine our face-off specialist is a serviceable but not impactful player, not helping or hurting the team outside their face-off talent. If we put Sprong and Milano on the wings, the total line would be generating 5.64 goals more than our opponents. That's almost worth a win in the standings.

Closing Thoughts

In terms of immediate efficiency, I think a checking 4th line is the easiest to assemble. If you're up against a salary cap crunch, the kind of grinders you want in that position won't chew up too much cap space, and there are lots of players who drive possession, slow down play, or draw penalties who don't make flashy highlight reels.

That being said, there's some versatility in a scoring 4th line. Your face-off specialist, if you can find one, can have some utility on special teams (notably less impact than even strength though). Your scoring wingers are likely to be useful higher up in the lineup when injuries inevitably strike, and it could be a useful way to integrate rookies into the lineup.

Individual goons are rare these days, so your uncle with the riding lawnmower can't have his wish. There is potentially some tactical value in fighting, but it mostly comes from taking better players on the other team off the ice. Ideally, you want to just have better players than the other team, but if there's ever a situation where Connor McDavid wants to fight one of your players, it would be foolish to decline.

Most teams can't or won't commit all the way to one of these philosophies. Naturally, more attention and energy is going to get spent optimizing the top of a team's lineup, and you can end up with a bit of a hodgepodge down there. That's fine most of the time- if something must be neglected, it probably ought to be the part of the team that plays the least. But for those teams close to their ceiling, having and building around a plan is vital.