Let’s talk about pulling the goalie

When Anaheim’s Sam Carrick delivered a dagger to the heart of the Kraken’s comeback hopes with 2:39 to play on Thursday night in the way of an empty net goal, it raised some questions about game strategy. Pulling the goalie is a risky move after all, and when it doesn’t work it looks really bad. Many Kraken fans had some questions regarding head coach Dave Hakstol’s decision to yank Philipp Grubauer with nearly 3 full minutes left and a mere one-goal deficit to overcome.

Seeing as how the Kraken have found themselves trailing in the waning minutes of the game more often than not so far this season, Seattle fans have grown accustomed to watching the goaltender head to the bench in favor of a sixth attacker. The problem is we haven’t seen it work yet.

When pulling the goalie fails, it often fails spectacularly. Watching your opponent get a hold of the puck and whip it down the ice from their own defensive zone into a wide open cage straight up sucks. It’s the final nail in the coffin, stealing away your hopes of a comeback and cementing the idea that you’ll be going to bed grumpy about watching your favorite team lose. Well, I’m here to tell you two very important things:

Lose by 1, lose by 2, it’s really all the same

In some sports, like soccer, your team’s goal differential comes into play in standings tie breakers quite often. It’s also a big factor in sports betting. And in predicting future wins, you’re probably better off using a form of goal differential over simple record. But in the NHL standings? It doesn’t mean a damn thing.

In the standings, teams get 2 points for a win, 1 point for a loss in overtime or shootout, and 0 points for a regulation loss. In the event that two teams finish the year with the same number of points, the league uses the following criteria to break the tie:

More often than not, after 82 regular season games, it doesn’t get farther than ROW when it comes to tiebreakers. Goal differential almost never comes into play. So, in terms of your overall place in the league, losing by 2 or 3 goals hits the exact same in the standings as losing by 1.

You were probably going to lose anyway

This, for me, is the hard one. The thing about sports is that there’s always hope that your team will win, so long as there’s at least a little bit of time left on the clock. The hope is the fun part, and having it dashed earlier than expected with an empty net goal does not feel great. But the fact is, if a team is losing in the final minutes of a game, they’re probably going to lose that game no matter what.

Over the past 3 seasons, there were 1,434 games where a team trailed by exactly 1 goal with 3 minutes to play. The trailing team managed to come back and win or at least take it to overtime in just 152 of those contests.

That gives Seattle about a 10% chance of coming back to tie this one at the time they sent Grubauer to the bench, and a 90% chance of losing in regulation. That’s in line with the win probability model from Evolving Hockey as well.

Now, this doesn’t mean teams should just throw in the towel because they’re probably going to lose. Over an 82-game season, it’s more likely than not that we see the Kraken come back and tie a game late and win it in overtime. In fact, they already did manage to tie a game late once in Arizona with the goalie pulled, we just only got to enjoy it for 13 seconds. But in all likelihood, a game in which you are trailing with less than 3 minutes to go isn’t going to end well, so toss that extra attacker on the ice and try to make something happen.

Are they really more likely to score with an extra skater?

In short, yes. Pulling the goalie for an extra attacker increases both team’s odds of scoring a goal. There’s a great piece by Meghan Hall on hockey-graphs.com breaking this all down, and I highly recommend it. She found that teams score almost 3 times as often playing at 6-on-5 compared to 5-on-5 (6.39 goals per 60 vs 2.25 goals per 60). That’s about the same goal-scoring rate as the average power play.

Pulling the goalie used to be something coaches didn’t do until there was maybe a minute or so left in the game. But in recent years, the trend has been to start pulling the goalie earlier and earlier. Most public analytical models support the idea of pulling the netminder early and in fact say we are still quite far away from maximizing the potential standings points gained from pulling the goalie. One 2018 study even concluded that teams losing by just 1 goal should pull the goalie with as much as 6 minutes to play!

Losing by giving up an empty net goal stinks. It’s no fun and we’ve seen it far more often than we’d like to in the first month of the season. But leaving him in longer isn’t going to increase the odds the Kraken come back to tie a game late. It’s a high-risk, high-reward strategy, and the only way to really avoid employing it is by holding onto a lead into the third period. With any luck we’ll see more of those games soon instead.