An NHL referee has an impossible job, making split-second decisions on things he may not be able to see. This was the case for Gord Dwyer in Saturday's Dallas Stars-Colorado Avalanche game. He had no angle to tell what actually happened when Nathan MacKinnon appeared to score with less than 10 seconds left in the second period.
Which is why it puzzles me when the NHL's war room fails to help the referees out when they absolutely can. The official word from Toronto was that replays were inconclusive as to whether MacKinnon's shot completely crossed the goal line. But when the overhead replay is examined closely, it sure looks pretty conclusive. The puck never appears to cross the goal line.
I'm a general defender of referees and league-office reviews. I think they get things right the vast majority of the time even when it makes teams unhappy. That's why it's so frustrating for me when bad positioning and a curious definition of "inconclusive" play such a large role in a game the Stars didn't deserve to win in the first place.
Before we get to the replay, the obvious place to start is the referee himself - how does Dwyer make such an emphatic goal call on the ice with no angle?
The answer is he guesses, but to be fair, he would have been guessing at a no-goal call as well. Look at his position here:
There is no way he knows whether the puck did or did not cross the line from there as he's screened out by Kari Lehtonen's pads and MacKinnon's body. He continues to drift to the back after this, not signaling until he's up against the boards. This is lazy skating - he needs to stop hard on the play and go to the net, trying to arrive before MacKinnon.
Still, since it would be a guess either way, I don't have a problem with the call of goal on the ice. He has to call something with authority, and guessing no-goal is just as tough for the Avalanche to deal with as guessing goal was for the Stars.
The more concerning point is that the poor positioning caused him to miss an obvious goalie interference that could have played a part in overturning a goal. MacKinnon's stick pushed Lehtonen's pad deeper into the net well after contact with the puck was lost. This is classic, rulebook-defined goalie interference that would lead to an automatic overturn of any goal should the contact have happened before the puck crossed the goal line.
But goalie interference, as always, isn't reviewable. Whether the puck crossed the line was.
From the review standpoint, the first thing we need to do is take the picture/video that made the rounds on Twitter, the low front angle, and discuss why you cannot take that as evidence for or against a goal.
A puck is an asymmetric object that, when on edge, has borders extending well beyond where it is in contact with the ground. This is, coincidentally, one of the main reasons that the multiple camera system used in tennis and soccer wouldn't work in hockey.
When a puck is up on edge, any angle that is not from directly above is useless for close calls. I took two pictures of a puck on edge near a "line" to show what I mean here. Pardon the bluriness of my terrible cell phone camera:
The above picture shows a relatively low angle where you can clearly see space between the puck and the line. In fact, it looks well beyond the line from this angle. But that's a trick of angles and the puck shape as shown in the image below:
The lesson from these shots is that if a puck goes up on edge, any low-angle shot is useless in determining whether or not the puck crosses the goal line unless the puck is over by a matter of several inches or more.
What the low angle shot can be used for is to determine what moments are in question, but that's all. You cannot defensibly say "Well, it looked like it wasn't over on one shot, but it looked really over on the low-angle, so it's inconclusive."
Because of this, we have to turn our attention to the overhead shot, shown here in GIF form here from the awesome @myregularface. The best part about the GIF is you can pause it to figure out where the puck is at any given moment, which I have used to track the puck through the course of the play below. If you pause it, the frame number becomes visible in the web address.
For convenience, I circled the puck in red, though I tried to leave enough of an outline to not obscure the goal line. I also circled MacKinnon's stick and how it forced Lehtonen's leg deeper into the net, causing whatever sort of confusion there was about the play.
In frame 70 above, you can see the back end of the puck creeping just over the goal line - it's the dark mark against Lehtonen's pads. At this point, it is conclusively not across the line.
In frame 75, the puck has fallen flat and is the dark circle against Lehtonen's pads. Here's where it comes the closest to going over the goal line, but again, there is no white space visible - conclusive evidence that is has not yet crossed. At this time, MacKinnon's stick begins making contact with Lehtonen's pad, pushing it back further into the net.
From this point on, this is goalie interference (a player is not allowed to shove the goalie into the net to make the puck cross the line). Goalie interference is not reviewable, as we've discussed ad nauseum, but this is just to point out how terribly out of position the referee was, as he missed something this blatant.
Below is the unedited version of frame 78 above, since the circles slighly obscure important parts. Again, MacKinnon is clearly pushing Lehtonen's leg into the net here - this is goalie interference regardless of where the puck is. You wonder how Patrick Roy would have taken this back in the day:
Again, no white space between the dark shadow of the puck and the goal line. This puck is conclusively not across the line with no obstruction of the camera at this point, and it has reversed momentum, moving back out of the dangerous area.
Lehtonen's arm has started to come into play here in frame 80, but you can see the puck in the middle of the red circle. That black bit is not a part of Lehotnen as the black stripe on his sweater is closer to the shoulder, his pads are too far away and the edge of his glove is clearly visible. That circled object is the puck, conclusively not across the goal line.
Anything that happens after this point clearly does not involve the puck crossing the line, as it has already reversed course and started to move out of the net.
Given that, the entirity of the time the puck was in the vicinity of the goal line, the goal line and puck were visible on camera. There was never any white space visible between the two; ergo, there is, at least to me, conclusive evidence that the puck never crossed the goal line.
Want further support that the black object above the puck? Here's the final screenshot:
Given the overhead angle and what we can see in detail from the screenshots, I contend that you can absolutely say the puck never crossed the goal line. The strip of white space that would have been visible between the two never appears.
Here's another exercise. Take the GIF linked above and start at frame 61, when the puck is on Lehtonen's pad and is clearly not over the goal line. Go frame-by-frame from there. You can track the puck fairly easily and never see any white space from it being over the goal line.
What the low-angle shot tells you is the puck kicked off the toe of Lehtonen's pad and had dropped to nearly flat at the time. This happens at frames 71-76, and you can see the puck conclusively not completely over the line as it reverses direction at that point.
With no obstruction, the only thing to contend with on the replay is shadow. Perhaps the NHL's war room got an SD feed and couldn't differentiate between shadow and puck, even though HD cameras and feeds are now standard. Perhaps they lack the capability to go frame-by-frame, though if that's the case, someone in the IT department needs to be fired.
The only thing I can think is that the NHL saw this angle clearly (because I refuse to believe either of the above is true) but decided it was not conclusive enough in conjunction with the (optical illusion) low-angle shot.
What they should have done is used the low-angle shot to tell them about the path of the puck, that it was up on edge then dropped flat, and the deepest it appeared to be in was when it hit the toe of Lehtonen's pad. Then, when going to the overhead replay, they would know exactly when to look for the tell-tale space between the puck and goal line. The low-angle shot is an aid to know when at what to look for, but the angle optics problems mean it is not telling evidence in and of itself.
The referee also deserves a talking to about his skating and positioning. This play is going to the net from the minute it is turned over, and he has no reason to fade back to the wall. He needed to stop hard at the goal line and come in around the back of the net. A good skating play by the referee likely would have caught the uncalled goalie interference.
The Stars have only themselves to blame for losing this one - Lehtonen gave up two weak goals, the defensive pair of Alex Goligoski and John Klingberg had a rough night and Kliingberg himself made the terrible turnover that led to the shot in question.
But the NHL also needs to do better, from making sure its referees don't put themselves in the position where they're forced to guess at a call to understanding which angles are useful and which are only going to give red herrings.