You know the scene: two teams slowly skate over to the face-off circle to the left of one of their goalies. The centres hunch over, hands sliding lower on their sticks for leverage, muscles tensing. The wingers engage in a heated battle for space at the edge of the face-off circle, trying to gain any tiny advantage they can. The official readies the puck above the face-off dot as the centres stare at it like a hungry dog spotting a piece of fresh meat. The puck dangles, dangles, dangles . . . and then a linesman says that one of the wingers is leaning in too close and everyone has to start again.

So the players all get back in place, and one of the centres seems to be cheating, so he gets thrown out, and they do it all again. And then the wingers are edging in too much again, so the linesman tells them to reset yet again. Eventually, some time long after the cows have come home, the puck finally hits the ice and play gets under way.

This is not my idea of entertainment. But what if there was a better way? What if we just dispensed with all that nonsense and got rid of offensive and defensive zone face-offs altogether? We'd speed up the game and create some much needed offence in the process.

The NHL loves face-off tweaks, but they don't matter
Over the past few years the NHL has made a couple of minor changes to the ways that face-offs work. In 2012 the Board of Governors approved a rule change that made it a minor penalty for either centre to make a hand-pass during a face-off. Then this past season another change was adopted, this time mandating that the defensive zone centre had to put their stick on the ice first. GMs believed the rule would create more offence.

One problem with tweaking face-off rules is that face-offs aren't actually that important. Back in 2011, Gabriel Desjardins calculated that it takes about 41 extra face-off victories in the offensive zone at 5v5 to create an extra goal. That might sound like it should add up, but because face-offs are so even across the league, it doesn't really create many goals.

The team with the highest winning percentage on o-zone face-offs this season was the Arizona Coyotes. They won 81 more offensive zone face-offs than they lost, about one per game. Their league-leading face-off prowess generated an extra goal every 41 games. The effect is very nearly zero.

Arik Parnass calculated the effect of the NHL's face-off rule change this season to be about 20 goals across the entire league over the whole season. That's fewer than one goal per team for the entire year. Face-offs just don't matter enough for small changes to produce much extra offence.

Get rid of o-zone face-offs

Offensive zone face-offs slow the game down, but being good at winning them doesn't lead to much additional offence. This seems like a lose-lose situation to me. Fans sit around waiting for play to resume while teams haggle over advantages that don't really matter. The solution to this is pretty straight-forward: the NHL should eliminate offensive zone face-offs (which obviously also eliminates defensive zone face-offs).

The idea may sound strange, but it's not really. Basketball games begin with a tip-off, but after that, teams simply inbound the ball to resume play. Soccer has thrown-ins to get the ball back onto the field, and play resumes after some other stoppages with a free kick. In both sports, the ball is simply given to one team, and that team gets the game going again.

The NHL should do something along those lines for offensive zone draws. I'm going to suggest one particular option they could pursue, but I'm not necessarily wedded to this idea; I'm open to alternate methods for giving the puck directly to one team instead of having a face-off.

My suggestion goes like this: any time there would be an offensive zone face-off, the puck will be given directly to the attacking team. I'd suggest that the defensive players must all start play below the top of the face-off circles, but offensive players can be anywhere on the ice. The referee will drop the puck on the ice directly in front of one attacking player just inside the blue line, blow their whistle, and play will immediately resume as normal.

The advantages to this system are two-fold. One, you decrease the amount of time fans spend waiting after the whistle. Two, you're creating extra offensive opportunities, which is what we want (at least, it's what I want, and I think many fans agree). If we want to create more offence, the kinds of rule changes to pursue are ones that reward teams for attacking and make defending more difficult. This achieves both of those, while speeding things up too.

How much would this actually increase scoring?

It's difficult to measure the precise number of extra goals this change would create, but I'll make an estimate, and then return with some caveats about why I believe it under-estimates how big the effect would be.

Let's assume that giving teams direct possession of the puck will have the same effect as winning an offensive zone face-off. If we want to know how many extra goals this change would create, we want to know what would happen if every defensive zone face-off victory became an o-zone win for the other team.

In 2015-16 there were 17,947 defensive zone face-off wins league-wide at 5v5 (source). Using Desjardins's estimate of 41 extra o-zone wins per goal, that works out to 438 extra goals over the course of a season, about 15 per team.

There were also 5036 d-zone face-offs won by teams killing penalties. Returning to Desjardins's post, we get an estimate of about 27 face-off wins per extra goal on the powerplay (because PP S% is higher than 5v5 S%). That gives us 187 additional powerplay goals, about 6 per team.

Add those two results together, and my suggestion for eliminating offensive-zone face-offs would seem to add 21 goals per team each season. Since there are two teams playing each game, that works out to exactly one extra goal every other game. Or, put another way, it would create an additional 0.5 goals per game across the league. That would bring the number of goals-per-game league-wide to roughly the same level as we saw in the 2005-06 season when the NHL's obstruction crack-down created a major boost to the number of powerplays.

(If you add in the roughly half-a-goal per-game that my other suggested rule change would add, you'd get quite a bit of extra offence league-wide.)

But I think that under-estimates the effect. Many face-offs that are won in the offensive zone are tightly contested, and do not result in one team clearly gaining complete possession of the puck. Even draws that go straight back to a defender leave the defensive team in relatively good position, with all offensive players on the same side of the ice, meaning shots are easier for the goalie to stop.

If attacking teams started with clear possession of the puck, with players in any formation they'd like, the opportunity to quickly get off high-quality shots will be much better. Additionally, teams would create set plays to quickly generate scoring chances. So while I've estimated that the change would result in an additional 0.5 goals per game league-wide, I think there's good reason to believe the effect would actually be larger. Exactly how much larger, I can't say.

One other effect that you might see is goalies trying to play pucks rather than freezing them. If you know that freezing the puck is going to result in possession being given back to the attacking team, you're more likely to try to find a way to move it to a defending player instead. This could create additional offensive opportunities, especially for teams that forecheck aggressively.

I know that a lot of people like face-offs, and I'm sure my suggestion will be met with skepticism, but I think it's pretty sensible and would be easy to enact. It would speed up games, increase scoring, reward attacking teams, and make defending harder. Those are all, I think, worthy goals, and eliminating offensive-zone face-offs would help accomplish them.

Analysis