I MIGHT Have a Solution to 3-on-3 OT Retreating

I make no promises, however

3-on-3 OT is on everyone’s mind right now, following the 2026 Milano-Cortina Olympics, where both the Bronze- and Gold-medal games ended in OT on the Women’s side and 3-of-4 Quarterfinals and the Gold-medal games ending in OT on the Men’s side. Much has been made of the format, with many fans I saw critiquing it and then Canadian Men’s coach Jon Cooper also decrying the choice of format following his team’s loss — which was not a good look for him. It’s true that this is a tough way to decide a single-elimination game, but the Olympics — and any international tournament — runs on time constraints and cannot risk dedicating indefinite time to a single game. The fact that they allow indefinite OT of any kind is a bit of a blessing rather than resorting to a shootout after such-and-such an amount of time. That’s the reality.

But 3-on-3 is also becoming an issue in the NHL. When it was introduced, 3-on-3 was considered a brilliant way to limit the number of shootouts it better reflected 5v5 hockey than a shootout and produced more decisions than 4-on-4. It also highlighted explosive speed and enabled a rapid exchange of chances. But now he metagame has been established and optimized and it’s clear that puck-possession over pushing forward is the way to win in 3-on-3. And that’s a bit tough because it means teams play the same way: maintaining possession, backtracking out of the offensive zone while their teammates change, and avoiding sub-optimal confrontations.

Over-and-back: My Proposal

This has lead to talk of a rule preventing moving “over-and-back” over centre or the offensive blueline in the NHL, with Russian junior leagues already implementing such a rule. As this article from NHL Trade Rumors from August 2024 explains,

With the new rule in the KHL junior league, the first violation will result in a faceoff in the offending team's zone, and every repeat violation will wind up with a two-minute minor being called.

A two-minute minor is a little draconian in my opinion (even for a second violation), but it makes sense as its the only escalating recourse after a faceoff in the offending team’s own zone has already been awarded. It’s also worth noting that 4-on-3 Powerplays are regarded as the deadliest type of powerplay. So, even though this may in the short term increase the effectiveness of 3-on-3 OTs at settling games, this might not be the way to go about doing it as inevitably there will be instances of defenders knocking pucks away and the attacking team going back to retrieve it anyway. Furthermore, we know that referees are generally impact-adverse, leading to non-action their wherever possible (WES MCCAULEY! YOU SPIT THAT WHISTLE OUT RIGHT NOW, YOUNG MAN!), so we can assume that if we imagine the set of true infractions under this rule, a lower percentage will be called after the first warning.

My main concern with this rule is that this creates a dead zone where play can break down between the offensive blueline and the redline. As soon as you enter the offensive zone, you’re trapped. That is the entire point, yes, but it’s not necessarily conducive to regenerating offence for the team with the puck. At a junior level, you may be able to get away with this as skill imbalances can work well enough to create breakdowns in coverage or force turnovers, but this might not translate to senior levels where players are better coached.

Rather than create this dead-zone, my proposal would be to the implement this ‘over-and-back’ rule, but further alter the way the lines on the ice affect the game. Specifically, the redline becomes a mutual offside line and each team’s d-zone blueline becomes their respective icing line. The goal of this is to maximize the accessible half of the zone in exchange for restricting voluntary access outside it. If there is more 2125 square feet more of space for players to go without committing an over-and-back infraction, it will organically compensate for the restriction that rule imposes. Why would you ever need to return to your own half if you have 2125 more square feet of offside-free ice?

You might be wondering, though, why players would need more space if there are already 4 fewer players than usual. The answer is that the trade-off to having fewer players and more skating space is that there is less occupancy in the intermediary spaces, so spreading the ice to create space for the puck-carrier is only beneficial so long as the puck-carrier has speed and position to exploit it. From my eyes, that doesn’t always appear to happen. We’ve seen many star players retreat to bad ice in 5-on-5 to generate speed in-zone, but that luxury doesn’t exactly exist in 3-on-3 because there’s fewer players for a puck-carrier to defer to.

Lengthening the offensive zone poses a conundrum to the defence: neutral zone ice is bad shooting ice, especially when there is less traffic, so do you back off if the puck is carried out that far and allow the opposition to attack downhill with speed, or do you close the gap, apply pressure, and risk creating a scenario where you’re even further out of position.

The goal of this is to create the initial enamourment with 3-on-3 was speed: it was so fast. But the methodical, puck-possession meta-game has completely replaced it. Now, I love puck-possession as much (or more) than the next person, but the willingness to retreat over surrendering puck-possession is telling us something that zone offence isn’t currently effective in the 3-on-3 paradigm, so they’d rather regroup and manufacture rushes. But these regroups are often initiated when a puck-possession team realizes they don’t have enough energy to attack. With some who are concerned about the game-ending effectiveness of 3-on-3 suggesting extending the format to 7 minutes instead of 5, I think we need to realize that such an adjustment would worsen the regroup problem, especially until coaches became more aggressive in lengthening their benches. And speaking of benches, in the NHL, the benches during 3-on-3 are situated between bluelines, so this allows teams to change in-zone and immediately be onside for the play.

And this is the goal: to create more usable space in-zone and fewer movement constraints to allow teams time and space to not only regroup, but also create more opportunities and use the space for attacking purposes more intentionally.

Caveats: The General Unpredictability of Rule Changes

Now, you might have noticed my non-committal title. That’s for a reason. All of these are my personal best guesses. I have no idea how this will actually affect the way 3-on-3 is played. I could definitely make the problem worse.

Despite its removal being one of the single-best decisions the NHL has ever made, Denis Potvin, in 2003, believed the removal of the 2-line pass would slow the game down even further, claiming that it would encourage teams to just retreat further than they already were in the dead puck era, and stack-up along the blueline.

This ended up not being the case as, after the rule’s removal, teams were more easily caught in transition, and that, coupled with the tighter officiating, led to the Trap ceding dominance in the NHL metagame and becoming a more strategy rather than a strict default.

The point of this is hockey is such an ecologically interactive sport, that changing one thing does not usually create predictable results. Hell, the league cracked down on obstruction, and pretty soon that lead to players resorting to slashing by the mid-2010s.¹

But here are some issues I do see with my proposal (and some of my theoretical rebuttals):

1. The defensive blueline becoming the icing line simply shifts the issue of a “dead-zone”/chokepoint sooner.
   I don’t know exactly how changing the icing line affects things. Icing at 3-on-3 is basically anathema at this point, so it may be moot, but I think there is slight possibility of “punting” occurring as a set play, similar to this series winner by Pastrňák.

   NHL players are skilled enough to get good dumps usually and will have enough speed to overrun D. Not only that, but theoretically they should be able to get more separation after a dump-in due to the acceleration the forwards.

2. Limits potential for stretch passes —

   This is arguably the big one. A stretch pass is typically received between the redline and offensive blueline: the very area I’m proposing to make inaccessible by moving back the offside line. This could pose an issue and restrict offensive options for teams in transition, and would be the number one thing I’d be observing in a hypothetical AHL trial scenario. The way to combat this would be to adjust the rule such that the over-back rule only comes into effect once the puck crosses the offensive blueline, at which time the offside line would be moved back to the redline. That might be a confusing solution, so we’d have to make wholistic considerations about the impact of these rules. After all, the tradeoff may be that this forces a bit more of a forecheck because the opposition is trying to sneak behind the defence as much, forcing more confrontation, forcing more acts of putting pucks to space and skating into it, which in fact accomplishes the goal of this rule change, which is to reintroduce an aggressive, speed component alongside the more methodical way of playing. Speaking of confusing…

3. This might cause general confusion.
   It is a major alteration from normal rules that breaks with the spirit of keeping the game resembling 5-on-5 whenever possible. But, consider that it’s players jobs to know the rules; broadcasts can inform/remind the fans (something they’re usually very good at); and the entire point is to prevent the 3-on-3 mode from deviating in spirit from 5-on-5 mode in terms of overall strategy and pace.

But I think these are minor concerns — at least in theory. At the same time, I cannot guarantee anything, but it makes sense in principle given our current understanding of the game.

I think, however, we need to be careful about overreacting to optimizations found within the game because sometimes they self-correct. Recall that following the 2014-15 season there were some suggesting that the nets needed to be made bigger in response to the size and skill of goaltending at the time. Good thing they didn’t end up doing that, because the very players who were just being drafted into the league at that time and the time since are now in their primes, and it turns out, they are collectively so much better at efficiently generating powerful shots from composite sticks, among other offensive skills than the late-80s/early-90s cohort. This has lead to league-average SV% to decline fairly consistently following despite the fact that goaltending itself hasn’t exactly stagnated.

So we do need to be sensitive to the needs of the game, but also consider that overcorrection is certainly possible — and even likely.

Does the retreat-happy 3-on-3 fall into that category? No. Because unlike the scoring environment, which has fluctuated for all of NHL history, the idea of retreating with the puck and ceding the offensive zone is almost unheard of at 5-on-5. The fact that it is now the norm at 3-on-3 is thus problematic.

And I don’t know I have the solution to it, but I think it’s worth a shot.

1. This is purely anecdotal, but I do remember the implied standard being stick-based (did you knock it out or break it) leading to a lot of hand and wrist injuries. The league famously changed course after Marc Methot nearly lost an entire phalange to a Crosby slash.